Wars in space were supposed to begin with lasers and explosions. China’s 2007 anti-satellite test — a missile climbing to 850 kilometers and reducing a weather satellite to roughly 3,000 pieces of trackable debris — conditioned a generation of defense planners to think in terms of missiles striking targets in orbit. Shrapnel would then spread silently through space, tearing into everything in its way, creating even more shrapnel endlessly into the stars. Ad Infinitum Astra.

That model obscures what’s actually happening right now: a quieter, harder to detect, less sexy electronic disruption of space signals governments, militaries, and companies depend on.

Here’s an example of what war in space could look like on the ground. One minute you’re leading a unit of soldiers to an enemy hideout; their location is based on intercepted communications; on your remote controller screen, you follow your drone to the target. The drone also acts as a scout, looking for ambushes, tripwires, camera and radio signals etc. Suddenly, the controller screen goes dark. You stop. Your unit stops behind you. Which way was it headed? How much further to the target? Where is my effing drone?? Your soldiers aren’t in motion, they’re static behind enemy lines, that’s very, very bad. They’re waiting for your decision. “Everything OK?” HQ asks in your earphone on an encrypted channel. Suddenly the drone screen comes back to life. Same picture as before. Must have been a software glitch or reset. “ALL OK HQ. PROCEEDING AS PLANNED”. The drone continues to lead the way to the target, now just 2 blocks away.

Meanwhile at the enemy hideout, phones start buzzing with SMS notifications, waking up sleeping fighters. ALERT: ENEMY ADVANCING TO YOUR LOCATION. 2 MINUTES. 12 MEN. IN 3 SQUADS. 1 DOG. HACKED DRONE LEADING THEM RIGHT TO YOU.

That kind of scenario —where the signals soldiers rely on for directions, intelligence, and communications are silently disrupted without anyone knowing—points to the emerging asymmetry between space-powered modern armies like Israel and the US, and countries like Iran.

The Secure World Foundation’s 2026 Global Counterspace Capabilities report — an open-source inventory of who is building what in “counterspace” warfare—opens a window into how satellite battles are actually shaping up. Nobody is firing missiles at enemy satellites (as of this writing). Instead, there’s increased jamming, spoofing, and electronic interference with the links that connect space to Earth. The war above the war is a war over signals, and it’s unclear how much longer the US and Israel will continue to dominate this ‘space’.

According to to the report, the US, Israel, and Iran are fighting in space through entirely different instruments.

The United States fields the world’s most advanced network for tracking objects in orbit.

Israel has launched a Space Directorate Unit within the IDF, demonstrated the ability to intercept ballistic missiles above the atmosphere, and shown confirmed use of signal-jamming and electronic interference. Israel is not building the architecture like the US, but it is stress-testing it under live fire.

The report says Iran is unlikely to have weapons capable of reaching and physically destroying a satellite. But it doesn’t need to. In the ability to jam and interfere with signals, Iran “has demonstrated an electronic warfare capability to persistently interfere with the broadcast of commercial satellite signals and Starlink ground terminals.” If it cannot reach the satellite, it can cut the connection.

The temptation is to read this as a confirmation of Iranian weakness. That reading is almost certainly wrong. Iran has found the approach that extracts maximum leverage from minimum capability, and the strong side is structurally more exposed to it than the weak side.

The asymmetry runs as follows: The United States and Israel have built their military advantage on precision — real-time targeting, satellite-assisted navigation, networked command, continuous communication between sensors and weapons. The whole architecture of modern high-tech warfare depends on the signal working correctly, every time, without interruption, under pressure. The more thoroughly a military has integrated space-dependent systems into its operations, the more any degradation of those links matters. Iran does not need to destroy the system to affect it. It needs to make commanders doubt the system at critical moments.

In the worst case, a sufficiently severe act of interference during a high-stakes engagement crosses a threshold. A spoofed navigation signal causes a missile battery to fail. A communications blackout during a critical commando mission costs lives. Nothing actually explodes in orbit. The war above the war simply becomes the ambient condition of all conflict in the region — a permanent contest over whether the signal can be trusted.

Iran’s strategy is, in its perverse way, the most rational military doctrine available to a mid-tier power facing a technically superior adversary. It has chosen to compete on decision certainty: to inject doubt into the reliability of precision systems at the exact moment they are required—degrading targeting feeds by seconds, corrupting navigation signals by meters, introducing latency into command links just long enough to force hesitation—so that weapons, sensors, and decision-makers remain intact but no longer fully trusted under combat conditions.

The Islamic Republic may be losing most of the conventional engagements. It has found a way to keep contesting the one thing the strong side cannot afford to be uncertain about. That is a more sophisticated strategy than its public image suggests, and a more durable one than its adversaries (AKA us) have fully reckoned with.

Is Scale‑Up Nation a Mirage? Why Israel Struggles to Build Very Large Tech Companies

For more than a decade, Israel has been celebrated as the “Start‑Up Nation,” a global outlier in entrepreneurial density, venture capital activity, and technological innovation. In recent years, however, policymakers, investors, and founders have increasingly argued that the next stage of evolution must be a transition from start‑ups to scale‑ups: fewer early exits, more independent growth, and the creation of very large, enduring technology companies headquartered in Israel.

The ambition is understandable. Yet the persistent difficulty Israeli companies face in scaling beyond a certain size suggests that this challenge is not merely a failure of will or imagination. Rather, Israel’s struggle to produce very large tech companies is the predictable outcome of structural incentives, capital‑market dynamics, and geographic realities that consistently favor early exits over long‑term scale.

Subscribe now

The Ceiling Problem

Israel has produced many successful technology companies with market capitalizations of $1 billion or more. Far fewer, however, have managed to grow beyond that point. Only a small number have crossed the $10 billion threshold for any sustained period, and none have meaningfully broken through a ceiling that appears to sit around $30 billion. This pattern has remained remarkably consistent over decades, despite repeated calls for change.

The debate itself is not new. As early as the late 1990s, critics argued that Israeli founders were selling too early and should instead pursue long‑term growth and independence. Decades later, the same concerns persist. Why do relatively few Israeli companies pursue IPOs or late‑stage growth financing? Why do promising firms accept acquisition offers rather than attempt to become global category leaders? These questions are often framed as matters of ambition, but the answer lies deeper.

Capital Markets and the Incentive to Exit

One of the most important forces shaping Israeli outcomes is the structure of its venture capital ecosystem. Since the 1990s, Israel has been one of the world’s largest venture capital markets in absolute terms. This was not driven solely by local talent. International investors were drawn by the unusually high volume of exits, particularly IPOs on Nasdaq followed by a wave of acquisitions during the same period.

For venture capitalists, Israel offered an attractive proposition: a steady pipeline of technically strong companies capable of delivering relatively fast and predictable exits. Over time, this expectation became self‑reinforcing. Capital flowed in with the assumption that companies would be sold or listed within a reasonable timeframe, and founders built businesses accordingly.

The passage of the Sarbanes‑Oxley Act in 2002 fundamentally altered this dynamic. Going public became more expensive and complex, pushing IPOs further out of reach for smaller companies. As a result, acquisitions increasingly replaced IPOs as the dominant exit route. Israeli companies that accepted takeover offers did not undermine the ecosystem; they sustained it. The continued willingness to sell ensured that venture capital kept flowing, but it also entrenched a model in which long‑term independence was the exception rather than the norm.

Scaling Is Hard Everywhere—But Harder Outside the US

Turning a start‑up into a unicorn is difficult. Turning a mid‑sized company into a truly large one is far harder. It requires profound changes in management structure, corporate culture, governance, and global footprint. These challenges are not unique to Israel. Outside the United States—and, more recently, China—very few countries have managed to produce multiple technology companies with market capitalizations above $50 billion.

As former European Central Bank president Mario Draghi once observed, Europe has failed to create any new €100 billion companies over the last five decades, and has produced only a handful of tech firms exceeding €50 billion over that period. Israel’s experience, in this sense, reflects a broader global pattern rather than an isolated failure.

Geography and the Center of Gravity Problem

For Israeli companies, geography adds an additional layer of difficulty. Building a very large technology company almost always requires deep, sustained proximity to major customers. While Israeli firms excel at research and development, their markets are overwhelmingly international from day one. Sales, marketing, customer success, and partnerships must therefore be built close to customers, typically in the United States or other major hubs.

This dynamic gradually shifts a company’s center of gravity away from Israel. While R&D may remain local, non‑technical functions grow faster as the company scales. Over time, the organization becomes more American—or global—than Israeli. This process has repeated itself across generations of companies and often culminates in a leadership relocation abroad, as occurred when Mercury Interactive’s CEO moved to the United States in the mid‑1990s as the company expanded.

Share

The Limits of Independence: Amdocs and Check Point

Two of Israel’s most iconic technology companies, Amdocs and Check Point, demonstrate both what is possible and what is difficult. Both went public in the 1990s, experienced dramatic growth, and briefly reached extraordinary market capitalizations—Amdocs exceeding $17 billion and Check Point surpassing $45 billion at their peaks.

Yet neither sustained those heights. Today, both remain significant global players, but far below their historic highs with share prices that have gone sideways for some time. Their trajectories underscore how challenging it is not only to reach scale, but to continue growing over long periods, particularly for companies headquartered outside the primary centers of global tech power.

Talent Scales Better Than Companies

While Israeli companies have struggled to scale, individual Israelis have not. Perhaps the clearest example is Nir Zuk. After leaving CheckPoint in 1999, Zuk set out to build a dominant, multi‑product cybersecurity company. To do so, he relocated to Silicon Valley, where he founded Palo Alto Networks in 2005. Through aggressive expansion and acquisition, the company eventually surpassed Check Point and became a $100 billion enterprise in 2023.

The contrast is telling. Israeli talent can clearly build global giants—but often only by doing so within ecosystems that are structurally optimized for scale.

Rethinking What “Scale‑Up Nation” Means

The aspiration to turn Israel into a Scale‑Up Nation is appealing, but it rests on assumptions that deserve reconsideration. Some of the obstacles Israeli companies face are universal, while others are deeply local. Incentives favor early exits, markets are far away, and the organizational demands of scaling pull companies abroad.

At the same time, Israel hosts over 400 multinational corporations running high‑level R&D operations in the country. These centers are not peripheral; they are deeply integrated into global innovation pipelines and have played a critical role in developing local talent and expertise. No other tech hub outside the United States hosts multinational R&D activity at this scale.

Perhaps, then, the question is not why Israel has failed to produce more very large independent companies, but whether this is the right metric of success. If global technology leadership increasingly takes the form of distributed organizations, Israel may already be far more central to the world’s largest companies than the “Scale‑Up Nation” debate admits.

Check out Meir’s upcoming book: The Network Effect: The Origins of Israeli High Tech, out July 7.

The scale of what has been built is not in dispute. MAFAT for Startups, the Defense Ministry’s startup integration engine, has doubled its portfolio to over 300 companies in two years. More than 100 of those companies have had technologies adopted directly into IDF operations during the current conflict. In Michigan this summer, twenty battle-tested Israeli startups will tour the National All-Domain Warfighting Center, twice the size of Israel, to meet the innovation planning team for Northern Strike, the largest multi-service military exercise on U.S. soil. In Miami, CENTCOM, the Defense Innovation Unit, Lockheed Martin, Anduril, and Palantir will convene with Israeli founders. The iteration loop underneath all of it: deployment, combat testing, feedback, update, now runs in days. A drone detection company gets IDF feedback that its acoustic sensors can’t distinguish quadcopter motors from diesel generators; forward-deployed engineers push an update within the week and validate against live targets.

The dominant narrative holds that October 7 permanently broke the procurement bureaucracy — that Israel has installed a fast-iteration industrial base that will outlast the conflict. This is mostly wishful thinking. Defense establishments are slow by design, not by accident. They are slow because the consequences of deploying untested systems at scale are catastrophic, and because institutional incentives — political, budgetary, contractual — reward caution in peacetime with the same consistency that combat pressure rewards speed. The argument that wartime has permanently dissolved those incentives requires the war to keep going in order to stay true.

The pessimists, though, are wrong about what this means. They see a wartime anomaly that will inevitably revert when the conflict ends. What they underestimate is how much the physical architecture of the ecosystem has already changed: the people, the processes, the institutional relationships between founders and operational commanders. That knowledge doesn’t evaporate. What it requires is a substitute customer.

Here the comparison to Israel’s cybersecurity industry, almost universally invoked as the template, points in precisely the wrong direction. Unit 8200 alumni built Check Point, CyberArk, and Wiz by carrying skills and networks into commercial markets where the end customer was a corporate security team that paid on a SaaS contract. The procurement cycle was months. The buyer had a credit card. Defense hardware has procurement cycles measured in years, interoperability requirements demanding regulatory clearance, export licensing contingent on diplomatic relationships, and political dependencies capable of vetoing a signed contract. The conditions that produced a $20 billion Israeli cyber export industry are structurally incompatible with the conditions required to build a durable defense-tech one. Founders who treat the cyber story as a map will misread every signal.

The situation resolves in one of three ways. In the worst case, the IDF’s institutional conservatism reasserts itself the moment operational pressure lifts, and the 300-plus companies that built for wartime procurement discover that peacetime defense budgets move on political timelines, not tactical ones. Several dozen will fail. The rest will pivot toward export markets they are not yet structured to access. In the best case, remote but not impossible, the IDF formally institutionalizes fast-iteration procurement as permanent doctrine, Israeli defense-tech becomes the reference architecture for AI-enabled warfare among NATO allies, and export revenues generate a demand signal durable enough to sustain the ecosystem without domestic conflict.

The most probable outcome is neither. Call it the Procurement Armistice: legacy systems and fast-iteration startups coexist in permanent institutional tension inside the IDF, neither fully replacing the other. Export contracts — U.S. DIU pipelines, joint pilots with NATO units, integration into existing allied platforms — sustain the ecosystem’s most capable tier. The rest consolidate, pivot, or dissolve. This is not failure. It is what a real defense-tech industry looks like before it matures. Israel’s cyber sector spent a decade in its own version of this before Unit 8200 alumni collectively understood that the government was not their primary customer. The founders now pitching to Pentagon officials in Miami are beginning to understand the same thing about the IDF.

The question is whether they figure it out before the peace does.

Share Israel Tech Insider

His Russian-built Yakovlev Yak-130, a jet trainer first flown in 1996 and introduced into service in 2010, was flying over Tehran when the aircraft simply disintegrated. The explosion appeared sudden, inexplicable. There was no visible attacker. Just a flash, then debris falling through the sky.

The same thing happened hundreds of kilometers away to the sailors aboard an Iranian naval frigate returning through international waters. One moment the ship was underway. The next it was torn open by a torpedo launched from a submarine no one ever saw.

The weapons themselves were not extraordinary. A missile from an Israeli F-35I “Adir”, the Israeli version of the Lockheed Martin F-35 Lightning II. A torpedo fired from a submerged American submarine.

What mattered was the system behind them.

The F-35 is often described as a fighter jet. In reality it is a flying sensor network — a data center with wings and bombs. It collects data from its radar, infrared sensors, satellites, and other aircraft, fuses the information, identifies targets, and presents the pilot with recommended actions. The human still authorizes the shot, but the machine increasingly decides everything that leads up to it.

For the Israeli pilot, the kill was incidental. The aircraft’s systems flagged the target, calculated a firing solution, and suggested engagement. The pilot approved. The missile launched and guided itself. The pilot then moved on to his main target over Iran.

Iran is discovering what it looks like to fight a war against militaries that are quietly integrating generative AI into the machinery of combat.

Despite their public spat, the United States military is integrating large language models such as Anthropic’s Claude into the Pentagon’s Project Maven intelligence platform. Israel operates its own AI-assisted targeting systems developed during years of conflict with Hamas and Hezbollah.

The objective is identical in both militaries: compress the time between information and action.

Modern military systems now ingest enormous flows of information: drone feeds, satellite imagery, intercepted communications, logistics movements, radar tracks. AI systems process that data continuously, identify patterns, and recommend targets.

Human commanders still approve strikes, for now. Against an adversary that does use AI in war (like Russia and China), that “human in the loop” decision could change. We are entering the age of autonomous weapons of war.

Even small product features in commercial AI hint at the direction of travel. Anthropic recently announced an experimental “auto mode” for its coding tools that allows Claude to run extended tasks without constantly requesting permission from developers. In software development that eliminates interruptions. In warfare it reduces pauses in analysis.

But the battlefield is fighting back

The problem is that the battlefield itself is becoming hostile to data.

Electronic warfare now saturates modern conflicts. Jamming, spoofing and signal denial are constant features of the environment. Systems dependent on GPS navigation or external communications often fail once those signals disappear.

The war in Ukraine show that drones that function perfectly during testing frequently fail once exposed to sustained electronic warfare. GPS signals vanish. Command links collapse. Guidance systems malfunction.

Autonomous systems therefore need to navigate and operate independently — using onboard sensors, cameras and AI models rather than external signals.

But that autonomy introduces another vulnerability. Machines depend on data. And data can be manipulated.

Poisoning the machine

Artificial intelligence introduces a new domain of warfare: data integrity.

The same generative AI tools that help Western militaries process intelligence can also be used by adversaries to corrupt it.

Future conflicts may involve:

• synthetic communications traffic

• fabricated satellite imagery

• manipulated sensor feeds

• adversarial attacks designed to confuse targeting algorithms

• poisoned datasets inserted into intelligence systems

A machine trained on corrupted data can draw the wrong conclusions with perfect confidence. Instead of hiding from sensors, adversaries may increasingly attempt to feed those sensors misleading information. The battlefield becomes not only contested but deceptive.

Two war machines

The result is a collision between two fundamentally different models of warfare.

The American-Israeli system is built on information dominance. Sensors collect vast quantities of data, software platforms fuse the streams, and algorithms help commanders decide what to strike. Its strength is speed and precision.

The opposing model is designed to make that system blind.

Instead of advanced networks, it relies on concealment, decentralization, tunnels, cheap rockets and human couriers. Its objective is not technological superiority but opacity.

High-tech militaries attempt to map the battlefield. Asymmetric adversaries attempt to make the battlefield unreadable.

The next phase of AI war

For now the technological advantage remains firmly with the United States and Israel. Our systems process intelligence faster, identify targets more efficiently, and coordinate complex operations with unprecedented speed.

But the decisive contests of the next decade may not be fought only with better algorithms. They will be fought over the integrity of information itself.

Future wars may hinge on which side can poison the other’s data, manipulate its sensors, or deceive its machine-learning systems. In earlier eras the key question of warfare was who possessed the most powerful weapons.

In the era of artificial intelligence the question may become something more subtle.

The United States and Israel this week unveiled what they called a new strategic framework for artificial intelligence and critical technologies, signing a joint declaration under the banner Pax Silica, a U.S.-led coalition aimed at building a “coalition of capabilities” among leading tech nations (but which includes Qatar) to reduce dependency on hostile actors — especially China — in critical supply chains.

The setting was deliberately theatrical: the City of David in Jerusalem, surrounded by archaeological ruins more than three millennia old. The message was continuity—ancient resilience meeting modern power.

The document signed, however, was not a treaty, a funding agreement, or an industrial plan. It was a joint declaration of intent, committing both governments to “deepen cooperation” on artificial intelligence, semiconductors, space, robotics, energy, and other “critical and emerging technologies,” while promoting secure and trusted supply chains and expanded research, investment, and commercialization.

In diplomatic terms, it was a signal. In policy terms, it was thin.

The declaration creates no new institution, allocates no funding (“Nothing in this statement commits the participants to the expenditure of funds”), and names no projects. It does not establish a joint secretariat, but rather a Joint Economic Development Group as “the primary steering committee to provide strategic direction on the implementation on areas of cooperation.”

There are no timelines, milestones, or benchmarks. Nothing in the text binds either side to act, beyond a general aspiration to cooperate more closely across a very wide technological frontier.

This matters because the problems Pax Silica claims to address—AI leadership, semiconductor resilience, energy security—are not solved by alignment statements. They are solved by hard choices: procurement commitments, export controls, investment screening, research restrictions, and expensive infrastructure build-outs. The declaration gestures at “protecting sensitive technologies,” but does not define what is sensitive, who decides, or how that protection will be enforced.

Artificial intelligence, the centerpiece of the initiative, is treated more as an abstraction than as infrastructure. The declaration speaks of advancing AI research and innovation, but says nothing about compute capacity, shared facilities, data access, or procurement—the bottlenecks that now define global AI power. There is no reference to joint supercomputing resources, cloud access, or coordinated investment in the physical stack that makes advanced AI possible.

The same pattern repeats across other domains. Semiconductors are invoked without specifying whether cooperation is focused on design, manufacturing, packaging, materials, or workforce. Space is framed as “the high ground” without distinguishing between civil, commercial, or security use. Energy is described as a “base layer,” but without any indication of whether the focus is grids, storage, nuclear, renewables, or industrial power for compute-intensive systems.

Perhaps most conspicuous is the absence of money. There are no figures, no funding vehicles, no public-private instruments, and no procurement pledges. A strategic partnership without capital is not strategy; it is branding.

The rhetoric surrounding Pax Silica goes further, presenting technological strength as a pathway to peace. That is a comforting formulation, but an incomplete one. Building trusted supply chains and technological blocs also entails exclusion, friction, and retaliation. None of those trade-offs appear in the declaration, which avoids naming adversaries, chokepoints, or enforcement mechanisms.

Pax Silica may yet become meaningful. Declarations often precede substance. But at present, it functions more as a narrative wrapper than an industrial framework—a way to signal unity and ambition without committing to the costly, contentious work that real technological sovereignty requires.

When the ceremony fades and the cameras leave the ruins, what changes? Who is in charge, what is funded, and what gets built? Until those questions have concrete answers, Pax Silica remains a declaration carved in stone surroundings, but written in air.

Israel’s tech sector looks healthier than ever. That may be precisely the problem.

In 2025, the industry delivered numbers most countries would kill for. High-tech exports rose to 57.2% of all Israeli exports, the highest share on record. Private funding rebounded to roughly $15.6 billion across 717 deals. Exit values topped $70 billion. NVIDIA announced a massive new campus. Google bought Wiz, the fastest-growing startup in Israeli history. Capital-gains taxes surged, pushing overall tax revenue up 12% and sending the finance minister into a celebratory hora.

By any conventional metric, this was a boom. Israeli tech at ludicrous speed.

Yet something fundamental did not move. High-tech employment dropped. R&D roles shrank. More engineers left the industry—and the country—than entered it.

That gap matters. Employment is not just a labor statistic, it tells a big story, especially in a small, export-driven country with a thin base of skilled tech workers (around 10% of the labor force). When tech revenue and exits surge without a corresponding expansion of the domestic talent base, it signals a dangerous disconnect between the tech industry itself and the broader Israeli economy. (Eugene Kandel talks a lot about this.)

Three forces explain the divergence.

First, capital concentration combined with AI. The $15.6 billion raised in 2025 was spread across the fewest deals in a decade. Investors crowded into a narrow band of late-stage winners—AI, cybersecurity, semiconductors—rather than seeding a broad pipeline of new companies. Those firms deployed capital to scale infrastructure first, headcount second or third. AI-driven productivity gains allowed them to ship more with fewer engineers. The result is a quiet arbitrage: Israeli talent creates value, but global firms no longer need to anchor large teams—or long-term commitments—in Israel to extract it. As productivity is being decoupled from headcount through AI, innovation is increasingly decoupled from place. Startup Nation could be turning into Subcontractor Nation: ideas seeded in Tel Aviv, scaled and controlled abroad, with Israel creating a handful of millionaires, collecting fees and taxes but not the strategic value that compounds over decades.

Second, currency pressure and margins. A strong shekel --in itself not a bad thing-- eroded the value of dollar-denominated revenues just as global markets demanded profitability over growth. Export-led software companies responded by trimming costs, not expanding teams. R&D employment fell by 6.5% in the first half of 2025. This represented roughly 14,000 fewer R&D roles in an otherwise stagnant employment environment.

Third, talent migration—both sectoral and geographic. Banks, insurers, retailers, hospitals, government agencies and defense contractors absorbed engineers who once would have built SaaS startups. With tech wages no longer soaring and IPO dreams deferred, stability and perceived national purpose became powerful magnets. Wartime demand is masking what may become a lasting drain on the civilian export engine. These flows are not easily reversible. At the same time, senior engineers and founders are making jurisdictional decisions based less on pay than on regime risk: judicial uncertainty, reserve-duty burden, family calculus. This trend predates October 7, but the war accelerated it.

Put together, the picture is unsettling. Israel’s tech sector is generating record revenue and blockbuster exits while the critical mass of talent—that wellspring of ideas and deep knowledge—is thinning domestically. Revenues, exits and tax receipts are lagging indicators of health; talent density, R&D depth and managerial experience are the leading indicators of future leverage.

The question is no longer why tech employment is flat (not enough women and minorities in tech, too centralized away from periphery, worsening education system, I can keep going). The question is whether Israel is drifting toward a model in which it supplies ideas, exits early, books short-term gains—and gradually loses control over where its technological future is built.

In 2026, a satellite can change orbit without being seen, an order can arrive in a general’s voice without coming from the general, and a strike plan can be drafted, approved, transmitted, and launched faster than a human can prove it’s a fake.

That is the throughline of the December 2025 Defense Tech Week newsletter (Hebrew), which lists transformative military applications that emerged in 2025 across Space, Defense, and AI.

Start in orbit. China’s Shijian‑25 spacecraft demonstrated orbital refueling in geostationary orbit, where communications, missile‑warning, and nuclear command satellites operate. Fuel has always been the limiting factor in space: once a satellite runs low, its behavior becomes predictable and its lifespan finite. Refueling breaks that constraint. A satellite that can top itself off can loiter indefinitely, shift position unexpectedly, or follow another satellite for inspection or interference.

At the same time, both China and Russia are deploying stealth satellites designed to be difficult to track. The newsletter describes systems with variable radar cross‑sections, meaning their surfaces and orientations can be adjusted so that radar signals sometimes bounce back clearly and sometimes scatter or disappear entirely. Ground stations that rely on radar to know where an object is instead see something flicker, shrink, or vanish. Russia’s Mozhayets satellite compounds the problem by being optically dim—too faint for many telescopes to track reliably. Space, once crowded but observable, is becoming more like urban warfare: unpredictable, dangerous, with rapid adaptations in tactics.

The United States is accelerating from a different direction: automation. The Department of War’s rollout of GenAI.mil replaces experimental chat tools with a mandatory, Google‑powered system. It speeds things up. The tradeoff is fragility. When AI systems generate plans at machine speed, errors propagate before humans can interrogate assumptions. The system does not need to be malicious to be dangerous—impersonating your enemy and issuing his orders is the Holy Grail of combat— it only needs to be wrong faster than correction is possible.

Meanwhile, reality itself is becoming unreliable. AI voice cloning has enabled impersonation of senior U.S. officials. Human authentication based on voice recognition is no longer reliable, even among experienced, security‑aware officials.

Deepfake videos—cheap to produce, increasingly convincing—have already reached tens of millions, including a fabricated coup video in France that spread faster than platforms chose to intervene. These are often reported as misinformation problems. In war as in life, timing is everything. A false order or surrender broadcast does not need to last long, it only needs to go viral enough to disrupt command decisions.

Individually, none of these developments end the world. Together, they change how wars begin. When satellites can move without being seen, orders can be issued without authors, and decisions are made faster than verification allows, escalation could come very quickly.

Civilian data centers process unicorn selfies. Military ones power the brain of the army: navigation, satellite imagery, secure comms, and targeting. Take them offline, and the modern military goes blind. No precision strikes. No coordination. Like horse cavalry charging machine guns.

The October 7 War exposed Israel’s reliance on foreign munition supply lines. But defense planners worry about a more subtle dependency. If an army loses the ability to crunch data—for navigation, code-breaking, targeting, and logistics—it is not merely slowed; it is lobotomized.

How To Attack a Data Center

The first problem is that data centers are easy to kill. They are usually hardened against missiles, buried in bunkers or reinforced concrete. But you don’t need to blow up a brain to kill it; you merely need to cut off its oxygen.

For a server farm, that oxygen is electricity and cooling. High-performance chips generate immense heat. Cut the municipal water supply or disable the circulation pumps, and processors throttle within minutes; within an hour, the facility goes dark. Sever the power lines or the fuel supply for backup generators, and computation ceases. Even if the lights stay on, cutting the fiber-optic cables turns a multibillion-dollar strategic asset into an isolated concrete mausoleum.

Thinking inside the box

The second constraint is physics. A tank commander under fire cannot afford the latency of the cloud. Sending sensor data to a distant bunker, waiting for an algorithm to process it, and receiving targeting data back takes milliseconds that the battlefield does not offer.

The solution, currently being pioneered by firms like Oracle in conjunction with Israel’s defense establishment (as well as startups like Poolside), is to invert the architecture. Rather than streaming data out, the models are deployed in. “Edge AI” puts the brain directly into the tank, the drone, or the soldier’s gun sights. The AI arrives in pre-trained, ready to execute decisions without phoning home.

The New Nightmare: Model Poisoning

This solution, however, introduces the most terrifying vulnerability of all: trust. Centralized models are easier to guard; distributing them across thousands of combat platforms creates thousands of attack surfaces.

The nightmare scenario is “model poisoning.” If an adversary can subtly corrupt the training data, they can turn an army’s strengths into liabilities. A tank’s targeting AI might be taught to ignore the thermal signature of a specific enemy tank, reporting “all clear” while an ambush prepares to strike. An autonomous supply convoy could be fed a corrupted navigation model that confidently routes it into a kill zone, all while displaying a safe green path on the dashboard.

Worse is the prospect of systematic friendly-fire, when the friend-or-foe identification algorithms are corrupted to only engage friendly forces. Drone swarms and artillery batteries could fire on their own troops faster than human operators can intervene.

The new arms race is therefore not just about building bigger bombs or autonomous killer robots, but also about verification. The challenge is to prove that a model has not been given a sleeper command, designed to trigger only under specific conditions. In the wars of the future, soldiers’ lives will depend on the integrity of the code in their weapons.

At the start of 2025, the dominant narrative was Israeli isolation: sanctions, boycotts, capital flight, brain drain, pariah-state dynamics. By Q4, the opposite materialized. Oracle, Nvidia, and other major Western tech firms tripled down on Israeli R&D. Investments and Exits soared. Gulf capital deepened. Defense procurement accelerated. AI collaborations flourished.

Startup Nation went to war in 2023. By the end of 2025, Israel became more embedded in Western and regional infrastructure, not less.

But that Silicon Shield is heavier and more brittle than anyone here admits.

This newsletter tracked one signal all year: Israel is both embedded and exposed—locked into critical Western tech, AI, and defense systems, while vulnerable to the politics, capital, and compliance structures that govern those systems. That duality—entanglement without autonomy—will decide who sets Israel’s strategic terms for the next decade: Israeli institutions, or foreign boardrooms whose interests may diverge, and whose circular economics may not survive the next geopolitical shock.

In 2026, I’m raising the bar for paid subscribers—more original reporting, sharper analysis, and earlier signals.

Subscribe now

3 Structural Shifts Defined 2025

Shift 1: Built In, Locked In

Israel didn’t become peripheral to Western semiconductors and AI—it became structural. As laid out in:

• Israel: The West’s New Chip Powerhouse

• Why Oracle and Nvidia Are Leaning Into Israel

• The Coolest, Hottest, Unsexiest, Most Globally Important Israeli Company You’ve Never Heard Of

Israel now sits on the critical path for global AI—chips, data centers, infrastructure, and defense systems. That brings both leverage and dependency—but not in equal measure, and not in Israel’s control. We’re critical to the world’s most important company NVIDIA, which is great. But if something bad happens to that company, that’s not so great. I worry about this.

At this point, our future depends on roadmaps set in foreign boardrooms, enforced by foreign regulators, and vulnerable to foreign political risk. Oracle’s next Israeli data center may protect domestic compute sovereignty—or Oracle will collapse under its own circular economy weight. Long Live NVIDIA and all, but putting so many AI eggs in Jensen Huang’s basket could go bonzai.

Subscribe now

Shift 2: Boards Set Borders

Foreign VC first investments dropped from 60% to 49%. Israeli VC struggled to raise new funds. European LP’s divorced Israeli tech. The Gulf is spending its way to AI supremacy. Israel, not so much.

In 2025, who funds you wasn’t just a financial question—it was geopolitical. Valuation decides exit. Sovereignty decides survival. Whether Israeli AI or defense startups align with European regulation, American cloud policy, or Gulf political priorities now depends on who sits on their cap tables and boards. The blurred line between commercial and defense tech is no longer theoretical, it’s operational. Most founders think “valuation.” The smart ones think “sovereignty.”

Shift 3: Super-Sparta As Operating System

Israeli tech stopped treating war as disruption and started engineering around it. Companies now design staffing, financing, and products assuming reservist call-ups, recurring shocks, and permanent mobilization. Foreign investors initially fled but they’re beating a path back. Combat-proven products, stress-tested management, “Founders Under Fire”—that’s the moat driving record-high Israeli tech activity.

The cost: Human capital stretched thin, capital tilting toward defense, liberal institutions under pressure from illiberal government, and tech employees who do all the military service, pay all the taxes, and get only more service and taxes as thanks.

Super-Sparta works under pressure. It fails over time. You can sprint on inherited infrastructure. But sprints don’t build futures. This has a stark effect on Israel’s most precious resources: the brains and hustle of its tech workforce. Doubly so in the all-important discipline of AI. Talent diverted from foundational R&D to short-cycle defense applications creates great middleware, but very little sovereignware.

Looking Ahead: 2026+VAT

2026 will clarify whether Israel is a strategic player or a sexy subcontractor—critical to the West, yet unable to shape its own strategic path. Here’s what I’m looking out for:

1. Permission To Compute?

Right now, Israel’s best AI work runs on ChatGPT, Claude, Gemini, Cursor, AWS, Azure, Google Cloud, NVIDIA chips, and the list goes on. That’s fine until Washington, Redmond, or Riyadh decides it isn’t, for whatever reason.

The 2026 question: Does any Israeli institution operate meaningful AI training capacity without needing permission from hyperscalers? If the answer is still “no” by December 2026, then “AI powerhouse” means talent exporter, not sovereign innovator.

2. 2023 Redux: Coalition Risk as Market Risk

Gaza isn’t settled. Neither is Lebanon. Syria looks promising until Turkey decides it doesn’t. And what have the Iranians done with those 400 kg of Highly Enriched Uranium? Don’t even get me started on the West Bank. Elections are likely October 2026. That returns Israeli tech to pre-war 2023: coalition overreach, judicial reform battles, government treating institutions as enemy. Foreign investors learned in 2023 that Israeli dysfunction is material to corporate decisions. Super-Sparta proved the ecosystem functions under external stress. It didn’t prove it survives internal institutional collapse.

In 2023, tech leaders spent more time fighting judicial overhaul than anything else. That energy wasn’t recoverable. If 2026 elections destabilize the coalition, same dynamic repeats: urgent displaces strategic. Super-Sparta proved the ecosystem can survive external shocks. It hasn’t yet proven it can survive internal chaos.

Signals to track: Do VCs build political risk clauses into term sheets again? Do acquirers demand new governance warranties? What if U.S. procurement officers get White House signals that Israel is liability, not ally? If 2026 looks like 2023 politically, 2025’s capital and partnerships won’t hold automatically.

3. Running on Fumes: The Human Capital Depreciation

Israel runs on scarce, high-end technical talent. That pool is aging, stretched, emigrating, politically and demographically squeezed. Does the political system treat human capital formation—military service inequality, education quality, research capacity, quality of life—as strategic priority? Or does another year pass where defense mobilization and culture wars consume political bandwidth while civilian infrastructure atrophies?

Super-Sparta works for companies needing resilient execution. It fails for nations and institutions needing long horizons and stable staffing. When your best people spend 200+ days per year in reserves, that’s not a productivity hit—it’s a constraint on what problems your ecosystem can solve. You get excellent at rapid iteration. You get worse at patient capital formation, basic research, deep infrastructure work that takes years to pay off.

Israel Delivers No Matter What*

Israel’s world-beating edge didn’t emerge in 2022 or 2023. It was built on investments made decades ago—in science, national service, research culture, and open institutions. But that inherited capacity is depreciating, faster than it’s being replenished. Super-Sparta accelerates depreciation by pulling talent toward near-term defense priorities and away from patient R&D. You can run on inherited capacity for a while. But decline doesn’t announce itself with crisis. It shows as gradual competitive loss that becomes irreversible before it becomes obvious.

📬 Thank You

Thank you for challenging, correcting, and sharpening this work all year. The best parts came from your pushback, insights, and willingness to test these theses against your own deals and ideals.

Give a gift subscription

Lessons from a Defense Tech Week conference in Tel Aviv.

The battlefield in five years: Autonomous drone swarms navigate 1,000 kilometers without GPS, running on quantum sensing and hydrogen not batteries. They identify targets through AI models running on the platform itself—no cloud, no communication, no latency and no vulnerability to jamming. High-power lasers intercept dozens of drones per engagement at the cost of electricity. When interceptors run low, 3D printing production lines shift from Arrow missiles to Iron Dome interceptors via software update—same factory, different munition, hours not months. Rockets on demand. A 3D-printed tank’s targeting system fails. The mobile data center on the vehicle retrains the model using battlefield data, air-gapped from external networks.

War returns to industrial fundamentals: the side that manufactures faster wins.

Israel’s defense establishment is rebuilding around this constraint. MAFAT—the Israeli defense R&D directorate equivalent to DARPA—sets technology priorities for the military. Rafael, IAI, and Elbit are Israel’s three defense manufacturing giants, together accounting for roughly 80% of Israeli defense exports and $15 billion in annual revenue. Their CTOs rarely speak publicly about production bottlenecks and technology gaps. At a Defense Tech Week conference at Tel Aviv University in December, they described the same problem from different angles: manufacturing velocity determines strategic outcomes, and current production models can’t match threat tempo.

The stakes are global. China produces a million drones monthly through DJI and builds ships at 200 times the US pace. Russia and Ukraine burned through decades of munition stockpiles in months. Iran’s April saturation attack forced Israel to fire interceptors costing millions against drones costing thousands. The economic asymmetry is unsustainable. The countries that solve manufacturing at scale—high-mix, high-volume production with continuous technology refresh—set engagement terms. The countries that don’t face operational irrelevance.

The Directed Energy Inversion

Rafael’s Iron Beam intercepted dozens of drones and loitering munitions in October—operational deployment against actual enemies, not laboratory tests. The system combines high-power lasers with tracking systems and adaptive optics compensating for atmospheric conditions. Rafael now fields systems from Iron Beam’s strategic power down to Light Beam’s 10-kilowatt tactical units.

The economics invert. Kinetic interceptors—rockets that shoot down rockets—cost millions per engagement. Laser intercepts cost electricity, pennies per shot once fielded. When saturation attacks arrive, defenders with directed energy sustain engagement rates that deplete attacker inventory while preserving their own capacity.

The constraint isn’t weapon design. It’s manufacturing the components: integrated photonics consolidating laser components onto chips, and power electronics using gallium nitride handling electrical loads in compact form factors. Both require fabrication infrastructure that doesn’t exist at scale. Erez Berkovic, Rafael’s CTO, estimates multiple years to build production capacity even with the technology validated.

Edge AI and the Certification Trap

Autonomous systems must operate without continuous communication because adversaries jam frequencies. Edge AI—models running directly on platforms—eliminates dependency on external compute. But deploying AI to weapons creates a regulatory constraint Avital Schrift, IAI’s VP of Core Technologies, calls explainability. High-risk autonomous weapons require certification in Israel, the US, and Europe. Regulators demand proof algorithms do what they’re designed to do.

Most advanced AI algorithms are NP-complete—mathematically impossible to fully explain. A neural network with millions of parameters can’t articulate why it classifies specific objects as hostile. For commercial applications this is acceptable. For autonomous weapons making kill decisions, it’s disqualifying. IAI develops AI architectures trading performance for explainability—models whose decision pathways can be audited and certified.

The gap isn’t model performance. Commercial AI models exceed military accuracy requirements. The gap is achieving certifiable accuracy regulators approve for lethal systems. Companies solving this unlock certification competitors can’t match—a regulatory moat that matters more than algorithmic sophistication.

Missiles on Demand

The endpoint IAI’s Schrift envisions: push a button, receive an Arrow missile in hours. Push again, receive an Iron Dome interceptor from the same production line through software reconfiguration. Additive manufacturing combines composite materials, conductors, and insulators in single print runs. Two gaps remain: materials withstanding extreme conditions—rocket motors require materials surviving 3,000+ degrees Celsius—and interchangeable design where software shifts production between munition types without hardware reconfiguration.

This isn’t theoretical. Elbit implemented predictive maintenance at their ammunition facility using vision sensors, acoustic monitoring, and vibration analysis. The system detects anomalies minutes before failures—enough time to shut down safely and swap components, preventing cascading damage that would scrap hundreds of thousands of munitions. The technology has run operationally for a year, proving the concept outside laboratory conditions.

The Availability Equation

Gal Harari, MAFAT’s CTO, introduces the math that inverts defense procurement: Defense systems traditionally optimize for maximum performance—longest range, highest accuracy, greatest reliability under all conditions. This produces exquisite weapons that spend most of their time awaiting spare parts, stuck in maintenance cycles, or sitting idle because they’re too complex to deploy rapidly.

Harari’s equation exposes the strategic cost: multiply what the weapon can do (capability) by how often it’s actually available to use (availability) to get real-world effectiveness. A theoretically perfect weapon available 10% of the time delivers 10% effectiveness. A good-enough weapon available 80% of the time delivers 64% effectiveness—more than six times better operationally despite being “inferior” on paper.

This changes everything. Instead of designing the absolute best system and accepting low availability as inevitable, design systems good enough to win engagements but simple enough to keep operational and fast enough to manufacture at scale. The US lost seven $50 million MQ-9 drones to Houthi air defenses this year partly because each drone takes months to build—production can’t replace losses fast enough. An 80% solution you can build in weeks and keep flying continuously beats a 100% solution that takes months to produce and spends half its time grounded.

MAFAT’s Blue and White initiative implements this philosophy across the technology stack: photonic chips, infrared sensors, radar components, electronic warfare systems, new materials, production automation. Think of it as a Lego Army: building ecosystems where components from multiple vendors integrate without custom engineering and production scales when demand spikes.

Robotics at Scale

Israel has already proven the manufacturing model works. Colonel Yaron Sarig, Head of AI and Autonomy at MAFAT, presenting two decades of robotics development, demonstrated systems that transitioned from laboratory concepts to tens of thousands of operational units. Ground robotics, aerial drones with networking capabilities, and multi-environment deployment assets now operate as integrated networks across multiple domains.

This matters because it validates the thesis before the US defense industrial base has solved the same problem. Israel manufactures autonomous systems at scale while American defense primes are still debating requirements. The robotics production lines demonstrate that high-mix manufacturing—shifting between different platform types while maintaining volume—works in practice.

Production Tempo as Strategic Outcome

Russia-Ukraine proved production tempo determines outcomes when technologies converge. Both sides field similar capabilities. The differential is who manufactures faster. Ukraine received sophisticated Western systems performing well individually but couldn’t be produced quickly enough to affect strategic outcomes. Russia fielded less sophisticated systems in larger volumes, accepting higher attrition because production replaced losses faster than Ukraine destroyed inventory.

Heven Drones—Israel’s first defense unicorn—addresses this by building infrastructure customers depend on. Their hydrogen refueling stations started as internal necessity, became a separate business serving military customers. Their 300,000 square-foot facility outside Washington combines mass manufacturing with continuous R&D because technology evolves too quickly to separate production from development. Bentzion Levinson, Heven’s founder, describes the timeline: first US Army contract five years ago, unicorn valuation took six years. The customer’s question is always identical: technology validated, when can you deliver 1,000 units?

Military force equals quality times quantity to an exponent. Past the “good enough” threshold, volume overwhelms precision. The defense industrial base that can’t manufacture at scale didn’t lose a production competition. It lost the war before kinetic contact. Stockpiles built with electronics from years earlier arrive obsolete. Production lines optimized for steady-state output can’t surge when conflict demands exponential increases. Supply chains dependent on adversary-controlled ecosystems fail when routes close.

China has manufacturing velocity and ecosystem depth. The US has technological edge and capital. Israel has innovation pressure forcing faster iteration—and a two-decade head start manufacturing autonomous systems at volume. The question isn’t who builds the best weapon. It’s who builds the best weapons factory.

Share

A new analysis by Jefferies, the investment bank, argues that the predicted isolation of the Israeli economy was a mirage. The headline figures are undeniably punchy: a record $39.2bn in merger and acquisition volume and a 54% surge in tech investment compared to the same period last year. But the real story isn’t the volume; it’s the nature of the buyers.

For Silicon Valley, Israeli cybersecurity has become too critical to consciously uncouple from. Google’s $32bn acquisition of Wiz and Palo Alto Networks’ $25bn purchase of CyberArk suggest that when the alternative is leaving cloud architecture undefended, corporate ethics committees discover the virtues of pragmatism. For companies like Oracle and Nvidia--combined market cap $4.87 trillion--Israel’s resilience/reward ratio outweighs its risks.

This logic extends to the region. The $35bn natural gas deal with Egypt, signed in August, is evidence not of the Camp David Accords’ warmth, but of their enduring utility. Blessed are the electricity makers.

This is the “Silicon Shield” in action: whether it is the code protecting a global bank or the gas powering a neighbor’s grid, deep infrastructural integration insulates against diplomatic freeze. For now.

Crouching Tiger, Hidden Drone

Beneath the bullish topline lies a narrative of extreme concentration. The Wiz deal alone accounts for more than 80% of the total M&A figure. Strip out this single transaction, and the picture shifts from a boom to a defensive reconfiguration. The Jefferies report highlights a doubling of defense-tech startups, framing this as a “battle-tested” advantage. The data confirms the growth, but the implication is structural: the ecosystem is pivoting. Capital is congregating in “hard” sectors—cyber, defense, enterprise infrastructure—where governments and multinationals have few substitutes, while implicitly moving away from the consumer-facing “soft” tech that is vulnerable to reputational radiation.

Israel is not diversifying; it is fortifying. The risk is that it becomes a Spartan Startup Nation—technologically lethal and export-rich, but increasingly decoupled from the open, liberal innovation economy of a decade ago. Remember food-tech?

This pivot to “hard” tech faces its own test in the AI revolution. There is a massive opportunity for Israel to become the “chip-to-cloud” foundry for the next generation of AI infrastructure, leveraging its deep expertise in semiconductor design, software optimization, and process efficiency. But the challenge is scale. Building Large Language Models requires massive compute and energy resources—areas where a small, isolated state faces structural disadvantages. Israel may excel at optimizing the AI stack, but without the sheer industrial capacity of Saudi Arabia and the UAE-- to say nothing of the US or China--it risks becoming an outsource outpost rather than a platform architect. Without a growing tech talent pool--the tech workforce shrank by 1.2% in 2024, the first decline in more than a decade--Israel, in the age of AI, risks having no globally competitive industry to support its economy.

If Trump wills it, it is no dream.

Ultimately, the bull thesis rests on a massive geopolitical wager: that a “new Middle East”—featuring an integrating Israel, a disarmed Hezbollah, a functional Lebanon, a contained Gaza, a political horizon for Palestinians, a stable Syria, and a non-nuclear Iran (did I miss anything?)—is not just a pipedream.

The real danger to Israel is not a sudden boycott (would we really miss Eurovision?) but the runaway attrition of its civic space. The Silicon Shield can deflect diplomatic heat, but it cannot block the exhaustion/emigration of human capital. A globally competitive and sought-after workforce that spends half the year in uniform and half the year at protests is not sustainable. A tiny, mighty, hurting, and vulnerable country living from ceasefire to ceasefire and unable to agree on what it wants and where it’s going is not sustainably investable.

For now, capital hates a vacuum more than it hates a controversy. Investors will bite the bullet for battle-tested innovation, but they will not wait forever for a country that cannot decide when the war ends and the business begins.

How this was left unaddressed all the way to 7 October 2023 needs to be investigated as part of a National Commission of Inquiry into Israel’s worst intelligence failure (kidding there’s not going to be a commission).

During the October 7 attack, Hamas squads managed to disable Israeli tanks in the Gaza envelope and render them inoperable. Terrorists activated a kill-switch that was supposed to be concealed, causing the temporary neutralization of the vehicles. For months, the army tried to understand how Hamas knew about this function, which was considered classified. Only at the beginning of 2024 was the answer discovered, after IDF forces located an underground compound known as “The Pentagon” in Gaza. This turned out to be the intelligence nerve center for the Jihadi group. Now the army has revealed some of what was learned there, and I’m seeing red.

The IDF investigation published this weekend revealed the depth of Hamas’s intelligence collection apparatus. The tank kill-switch was just one example. For five years, Hamas tracked nearly 100,000 Israeli soldiers across social networks. Gaza’s rulers established a network of impersonator accounts that befriended soldiers and officers on Facebook and infiltrated WhatsApp groups mimicking IDF units, such as the “Egoz Recruits” group. They tracked combatants from initial recruitment through their progression to platoon and company command positions.

Hamas military intelligence cross-referenced this data to map the location of every company, every missile battery. They monitored each “classified” deployment in real-time—the moment soldiers complained about it in internal WhatsApp groups. They harvested photos and videos from bases to decode facility layouts, security camera positions, interior tank configurations, weapons systems. Training videos posted by soldiers at an armor training base. “Mundane” photos and clips from inside Merkavas. Instruction footage. Casual documentation that individually meant nothing. Then Hamas built full-scale Merkava Mark IV tank replicas and advanced simulator software to train what they called their “Tank Corps”.

Hamas had trained elite squads as IDF tank operators. Not just disable-and-destroy. Capture and operate. The plan: seize Merkava tanks during the October 7 assault and drive them into Gaza, to use against IDF forces. The capture-and-extraction component failed, but the tanks couldn’t fight.

I don’t enjoy complimenting sworn enemies, but kudos to them. Respect to whoever architected this apparatus from nothing and managed it operationally for years. Unfortunate they’re on the opposing side. There’s much to reverse-engineer. But assuming that senior officer is now dispersed across half of Gaza, we must decode the strategic implications ourselves.

First implication: Hamas is not merely a “terror organization.” This operation required 2,500 Hamas personnel dedicated to intelligence collection and analysis—sustained over years with disciplined patience and institutional rigor. Mounting such an operation demands the resources of a small state, or at minimum a city-state. That’s what Hamas was, and regrettably appears to be reconstituting: an adversarial state entity positioned on the border, observing, listening, waiting for optimal strike windows. Hezbollah is no different.

Second implication: We regard WhatsApp as highly secure—with valid reason. End-to-end encryption renders message interception theoretically impossible during transmission. Except: humans constitute catastrophic vulnerabilities. Elementary social engineering breached the unit’s “perimeter of trust”, and soldiers freely shared intelligence believing themselves “among friends.”

Perhaps most devastating: 100,000 IDF soldiers functioned as unwitting double agents—inadvertent Hamas collection assets. Every mundane Instagram story against a fence backdrop, every farewell photo near an armory, every TikTok barracks dance: each transmission fed the adversary incremental intelligence. Aggregated, these fragments rendered bases comprehensively compromised.

This exposes—once again— a fundamental flaw in prevailing information security doctrine. The IDF invests heavily in preventing classified document leaks. Meanwhile, massive amounts of intelligence hemorrhages through social networks while leadership throws up its hands with a “Kids these days…what ya gonna do?”.

But gather a steady stream of this content, rent a couple of servers in Malaysia, run pattern matching, geolocation, and other open-source analysis, and hey Presto, wipe out the entire Southern Division of the strongest country in the Middle East.

One officer acknowledged: “Hamas understood this base better than I did—and I served there for years.”

The remedy: Return to fundamentals. Zero information sharing on public networks. No TikTok content. No Facebook weapons photos from installations. Upload a video? Face court-martial.

Good luck enforcing that with a generation neurologically incapable of not documenting their existence.

Israel’s AI sovereignty is off to a late start

In a classroom in Tel Aviv, two dozen engineers, graduates of Israel’s most elite military units, are learning to build large language models from scratch. Not to fine-tune them or write clever prompts, but to construct transformers, attention mechanisms and gradient-descent algorithms in PyTorch. The Architect Track at APEX, a new private-sector training program, has recruited precisely the people who should be architecting a national AI capability: alumni of the country’s top military R&D programs like Talpiot and Psagot which takes exceptionally bright youth and turns them into little Tony Starks [[Iron Man]].

The curriculum is rigorous, the students exceptional. They would be prized employees or startup founders anywhere in the world. The problem is that Israel has built nowhere for them to work.

This is not for lack of raw capability. Israel ranks first globally in AI research citations per paper and first in AI patents per capita, according to data compiled by the Ministry of Innovation, Science and Technology. Yet it ranks 26th out of 83 countries in AI infrastructure readiness. 65th for operating environment. It has no national supercomputer. Under Biden-era export controls, still in force, Israel is capped at importing 50,000 high-end GPUs per year through 2027—an “economy class” allotment, as Nvidia’s local director put it. By contrast, the United Arab Emirates is negotiating for 500,000 chips annually. When Crown Prince Mohammed bin Salman and President Donald Trump convened Silicon Valley’s AI elite in Riyadh this spring to announce $600 billion in commitments, Israel was not invited. The reason wasn’t just political. When American and Gulf planners surveyed the Middle East for AI infrastructure, they found vast solar farms, natural gas plants, oil facilities, fast-tracked data-centre permits and sovereign-wealth backing in the Gulf states. In Israel they found talented individuals renting GPU time in Frankfurt.

Talent without infrastructure

The gap between human capital and physical capacity has opened a second time. Israel exports $10bn-worth of cybersecurity products and services annually—more per capita than anywhere on earth. It also suffers 3.5% of global cyber incidents despite representing 0.1% of the world’s population. Israeli hospitals are encrypted by ransomware gangs using credentials bought on Telegram for $50. Seventeen per cent of successful breaches exploit stolen passwords; another third exploit misconfigured access controls. Israeli firms manufacture the software systems that prevent 99% of these attacks. They export them worldwide. Domestic critical infrastructure, however, remains unprotected, because no ministry mandated their deployment at home.

The pattern is structural, not coincidental. Unit 8200 optimises for offensive capability and venture-backed exits—the high-margin work of finding software vulnerabilities and building exploits. No institution optimizes for the unglamorous operational discipline of mandatory two-factor authentication and quarterly access audits. Israel monetised cybersecurity expertise while exempting itself from cybersecurity hygiene. It became the world’s foremost cyber-arms dealer and simultaneous proof that technical brilliance without institutional discipline is expensive theatre. Now it is repeating the experiment with artificial intelligence.

If you build it, they will come

APEX was founded to solve what its architects, themselves alumni from the IDF’s elite tech units, call the “talent-to-capability conversion failure”. Elite engineers emerge from classified military units with deep technical skill but no defined pathway into civilian AI leadership. The innovation model that built Israel’s cybersecurity industry—secretive, hardware-light, optimized for tactical advantage—cannot scale to the compute-heavy, open-data requirements of foundational AI. Foundation models like ChatGPT, Claude, and Gemini need public datasets, international collaboration and energy infrastructure measured in gigawatts. The top-secret playbook produces brilliant point solutions and lucrative exits. It does not produce sovereign capability. The result is that raw talent dissipates into incremental software features or emigrates to California. APEX’s co-founder Avishag Bohbot describes the dynamic as “talent rising, position falling”—Israel’s global AI ranking declines even as its per-capita research output climbs.

The Architect Track teaches what military research-and-development does not: large language models from first principles, transformer architectures, model deployment at scale. The parallel Founder Track updates the commercialisation playbook for the AI era: term-sheet psychology, American market entry, investor narratives that translate defense instincts into venture credibility. Together, the tracks aim to function as what APEX calls “a national talent router”, converting technical density into deployable capability. The curriculum is serious, the cohort model effective and the network effects genuine. But will it be enough to stem Israel’s slide in global AI power?

What the state will not build

In August 2025 a national committee chaired by Yaakov Nagel, a former national security adviser, issued its (blunt) assessment. “Israel is not at the right point, and not even at the desirable trajectory,” it concluded. “Without swift and adequate action, the country will experience a further and rapid decline in its global status.” The report quantified the shortfall. Israel fields roughly 120 researchers working on core AI—foundation models, training infrastructure and fundamental algorithms. The University of California, Berkeley, alone has 70 working in comparable areas. The committee proposed a remedy: NIS 25bn ($6.6bn) over five years for a National AI Directorate with procurement authority modelled on America’s Defence Advanced Research Projects Agency (DARPA), 60,000 GPUs for academic use within 24 months, fast-tracked grid connections for data centres, a sovereign energy strategy and open government datasets with unified privacy frameworks. The budget amounts to 0.25% of GDP annually, less than cost overruns on Tel Aviv’s light-rail project.

The government’s response has been to appoint Almog Cohen, a member of the far-right Jewish Power party, as deputy minister for AI. Mr Cohen’s CV includes yeshiva education, militia leadership and pizza-parlour management. No AI credentials appear. His mandate is to shift Israel’s AI “centre of gravity” to the Negev and Galilee—a goal that collides with the fact that Israel’s electricity grid was designed for peacetime residential loads, not desert data centres requiring cooling systems in a water-scarce region and transmission upgrades tangled in bureaucratic approvals. No budget has been appropriated. No infrastructure plan exists. No compute strategy has been articulated. The distance between APEX’s human-capital production and state infrastructure investment now defines the paradox.

What’s at stake?

Israel’s technical contributions to global AI and semiconductors are not marginal. Amazon Web Services runs on chips designed by Annapurna Labs, an Israeli acquisition that AWS’s chief executive called “probably the most important moment in AWS history”. Microsoft’s Azure cloud security was architected by engineers who later founded Wiz, now the fastest-growing cybersecurity startup in history, acquired by Google for $32B. Apple’s Vision Pro spatial-computing headset relies substantially on Israeli-developed technologies. Nvidia is expanding its Israeli operations to absorb engineers Intel can no longer retain.

Yet there is a difference between being an AI power and being a component supplier. Powers set technical standards, anchor supply chains and seed next-generation infrastructure. Suppliers generate revenue but cede governance. Without sovereign compute, energy infrastructure and data frameworks, Israel defaults to supplier status. APEX graduates will still build important systems—on Amazon’s cloud, using datasets hosted in Dublin, within regulatory frameworks defined in Brussels, generating returns that flow to Seattle shareholders.

APEX puts it even more starkly: “AI is central to defense, the economy, intelligence, and global standing. Losing leadership means loss of sovereignty and security edge. Since October 7, ongoing isolation shows sovereignty is a must. The stakes are existential.”

They’re not wrong, as I explored in my post on the Israel-Iran 12 Day War.

Israel’s AI Snapshot

The failure mechanism is economic, not technological. AI capability rests on three non-substitutable inputs: talent, compute and data. Israel has invested exclusively in the first. Talent without compute forces researchers to lease cloud credits, embedding dependency into foreign-governed infrastructure. Compute without data means training on synthetic or foreign datasets, forfeiting advantages in domains where Israel is defensible—medical imaging, defense systems, multilingual natural-language processing for populations fluent in Hebrew, Arabic, Persian, Russian, Amharic and French. Data without talent leaves expensive infrastructure idle. Economists call this complementarity collapse: each input becomes less valuable when others are absent.

The UAE, meanwhile, has committed $100bn to sovereign AI infrastructure, including foundation models in Arabic. Saudi Arabia’s state-backed AI firm, Humain, is deploying 18,000 specialised Nvidia servers. The Gulf is fusing solar capacity, nuclear reactors and subsidised grid upgrades to create low-cost power for hyperscale GPU clusters. Fast-track permitting, cheap land and sovereign-wealth underwriting eliminate the bottlenecks—planning delays, energy tariffs, risk premiums—that slow deployment elsewhere. America’s AI giants committed $600bn in Riyadh not because Gulf states have superior researchers (they do not), but because electron-to-inference supply chains now matter more than ARXIV citations. The world’s focus has shifted from talent to transformers—the electrical substations, cooling plants and grid interconnects that deliver power at scale. Israel is still competing on the old metric while the race has moved.

The choice ahead

Israel now faces a binary decision. It can fund the infrastructure backbone—national compute grid, energy strategy, supercomputer access, data governance that allows startups to train on Israeli medical records and transportation telemetry under privacy-preserving frameworks. That converts APEX’s output into deployable sovereign capability. Or it can continue producing world-class engineers who scale models on Google Cloud Platform, train on ImageNet, deploy on AWS and generate returns that accrue to foreign balance sheets. The former requires the NIS 25bn and cross-ministerial coordination that successive Israeli governments have proved incapable of sustaining outside wartime. The latter requires accepting that “Startup Nation” has become a talent-export terminal with a flag.

The cybersecurity precedent suggests which outcome is more probable. Israel built the world’s most sophisticated offensive cyber capability, exported it profitably and exempted itself from the defensive discipline that capability implied. The result was not a failure of intelligence but a failure of the assumption that excellence obviates structure.

The broader lesson extends beyond one small country. AI capability is increasingly defined not by the smartness of individuals but by the density of physical infrastructure: chips, watts and fibre. Nations that treat AI as a software problem to be solved with talent and venture capital will find themselves supplying components to nations that treat it as an infrastructure problem requiring energy policy, grid upgrades and sovereign compute. Israel pioneered the former approach in cybersecurity and is now repeating it in AI, despite clear evidence that the model produces profitable companies but not strategic autonomy. APEX trains the infantry. The state must build the base.

Meanwhile, the gap is closing. Not because Israel is building faster—because everyone else is. The Gulf states are converting oil wealth into electron wealth at sovereign scale. The U.S. is integrating AI into defense procurement and grid planning. China is mandating AI deployment across industrial policy. Israel is appointing pizza parlor managers and hoping Unit 8200 alumni can bridge the distance with talent alone. The global AI race will not pause for Israel to reconcile the contradiction between the individual brilliance it celebrates and the institutional coordination it cannot get right.

The author advises technology companies on strategic communications and previously served as EMEA Technology Editor at The Wall Street Journal.

Why Oracle and Nvidia Are Doubling Down on Israel

While many global tech firms tread carefully around war-time Israel, Oracle and Nvidia are leaning in. Their decisions are not acts of virtue-signaling but calculated bets on a country whose innovation engine keeps running through crisis. For both firms, Israel is a strategic asset that outweighs their corporate reputational risk.

Oracle: Sovereign cloud, extreme infrastructure

Oracle’s stance is both explicit and embodied. After the October 7 attacks, Oracle posted “Oracle stands with Israel” in multiple languages across its global websites, provided Israeli employees with double salary payments, made a $1 million donation to Magen David Adom, and launched a global employee fundraising campaign for Israeli relief causes, matching every dollar raised.

The company’s leadership, including former CEO Safra Catz—one of the highest-profile Israeli-Americans in global tech— and founder Larry Ellison, have repeatedly stated that supporting Israel is integral to Oracle’s values. Catz has repeated the “We stand with Israel,” ever since, positioning the country as both a moral and strategic partner.

The company operates an underground data center in Jerusalem, 50 meters below ground and hardened against missile strikes, and has confirmed plans for a second, deeper facility. The goal: guarantee sovereign-capable cloud services for governments, hospitals, and defense clients that cannot afford disruption. Oracle is also working more directly with Israeli defense bodies on edge-AI and mobile data-centers that bring AI to the battlefield.

While some of these deployments are operationally secret, the overall embrace of wartime Israel is public, explicitly so. Oracle wants people to know that Israel is where Oracle stress-tests the future of mission-critical computing for regulated industries like defense, government, health, and finance. Israel becomes a “proof-lab” for resilience and sovereign-cloud services that Oracle intends to export globally. With global defense spending booming, this is a smart bet by Oracle.

The Humans of Huang’s Nvidia

Nvidia’s support for Israel has been quieter but no less sweeping. After October 7 2023, the company led a $15 million humanitarian fundraising drive—the largest in its history. This year it began scouting land in northern Israel for a new mega-campus of up to 180,000 m², expected to employ thousands of engineers and anchor Nvidia’s chip-to-AI-system ambitions.

Then came the story that gave all this weight. Avinatan Or, a 32-year-old Nvidia electrical engineer kidnapped at the Nova music festival, was released in October 2025 after 738 days in captivity. For two years, Nvidia Israel’s HR chief sent weekly emails counting the days, and every quarter Jensen Huang mentioned Or’s name in global meetings to remind 40,000 employees that one of their own was still missing.

When Or finally returned, his manager wrote: “I hope you have a good excuse for not showing up to work for 738 days… and I also have some pretty good news about what’s happened with your stocks over the past two years.” It was a note equal parts humanity and capitalism—proof that loyalty and value are not opposites in this business.

Big Tech, Small Backbone

Other major tech players hold Israel operations but their posture is commercial rather than values-driven. Amazon had its own Israeli hostage, Sasha Troufanov, a 29-year-old AWS electrical engineer who was kidnapped by Hamas from Kibbutz Nir Oz, along with his family. Troufanov was finally released in February 2025 after 498 days in captivity, as part of a ceasefire-hostage deal. Amazon did not publicly comment on Troufanov’s abduction while he was captive, explaining that any statements could risk his safety and negotiations for his release.

While Google hasn’t expressed anything publicly one way or another, in 2024 it fired dozens of Google workers who protested the company’s cloud contract with the Israeli government. In July this year, Google co-founder Sergey Brin called the United Nations “transparently antisemitic” after it published a report on companies, including Google, that profit from working with Israel, The Washington Post reported.

Meanwhile, Microsoft expanded its Israeli Azure cloud region in 2023 but has buckled to internal pressure to sever its Israeli military sector links.

Risk, Resilience, Reward

Oracle and Nvidia alone have made Israel central to their corporate identity rather than a peripheral operation. The upside: Oracle gains proximity to premier engineers, sovereign-cloud demand and resilience credentials. Nvidia embeds itself in a deep talent pool and locks in scale for next-gen AI architecture. But the risks are real: Oracle’s open alignment may alienate customers or employees in markets sensitive to the Israel-Palestine conflict. Nvidia’s high cap-ex campus in a volatile region could suffer disruptions from geopolitical, supply-chain or regulatory shocks. But they are both betting that the future of compute and cloud will not just run through Silicon Valley but also through Tel Aviv and Haifa.

Oracle and Nvidia showed that resilience pays dividends, that in Israel, the moral and the material are converging, and that the companies willing to stand firm in hard times will also be the ones holding the most valuable stock when the dust settles.

The (War) God of Little Things

Israel’s defense establishment is undergoing its fastest structural rewiring since the 1973 Yom Kippur War. Then, as now, the initiative is reactionary — born of failure rather than out of doctrine. The Israel Defense Forces (IDF) Southern Command’s surprise defeat on October 7, 2023 exposed critical flaws: the vaunted US$1 billion Gaza-border sensor-network collapsed under cheap drones and basic hacks; the initial response was slow and disjointed; Hamas’ tunnel network exploited blind spots; and fighting while protecting hostages paralyzed decision-making.

The failure was as much organizational as technological. Over-reliance on technology, entrenched group-think, and mis-read intelligence meant doctrine didn’t match battlefield reality. In response, Israel is turning directly to “network resilience” as its new priority. On the front line, commanders are deploying real-time live-map overlays to track every soldier, automated sensor-software to detect guerrilla ambushes and AI-augmented portable command-kits that keep units fighting even if headquarters is knocked out.

Behind the scenes, engineers have compressed R&D cycles from years to weeks. The defense-tech boom is unmistakable: hundreds of startups founded since October 7, and a state-backed, billion-shekel investment scheme now channels government guarantees into smaller firms once overshadowed by Elbit, Rafael and IAI. The logic: wars now are software problems deployed through hardware constraints, and advantage flows to the nation that integrates fast and recovers lost momentum faster.

Yet speed introduces new vulnerabilities: cyber-exposure, power grid fragility and an over-reliance on an off-shore manufacturing base (AKA America). At the same time, global arms-embargo pressure mounts even as Israel’s defense-industry exports continue to rise. The future of Israeli security will hinge less on missiles and more on the industrial, digital and human web linking sensors to soldiers, factories to battlefields, code to combat.

For more of my writing chronicling how Israel’s defense establishment is retooling for AI war, check out:

As always, feedback to amir@israeltechinsider.com

Israeli Hero outguns Switchblade

Real-world performance is overtaking domestic favoritism in Pentagon procurement

The Pentagon just handed a $982 million loitering munitions (AKA killer drone) contract—not to an American defense giant, but to a Maryland proxy for an Israeli drone maker, UVision. The contract runs through 2030.

That matters because UVision’s Hero drones beat out AeroVironment’s Switchblade—a system the U.S. has already invested nearly $1 billion into—and Anduril’s Altius, backed by the most hyped defense startup in America.

According to Defense Express, it wasn’t even close. The Hero outperformed on range, loiter time, and general bad-assiness, exactly the metrics that matter when drones replace close air support.

So why did the Pentagon override the usual default to domestic firms?And does this mean that the core assumption that “Buy American” always wins? The answer: when timelines and performance diverge, even billion-dollar legacy bets get sidelined.

If this marks a significant shift in procurement strategy then that’s good news for Israeli defense tech firms, who face a labyrinth of American protectionism that would kill most foreign companies: the Buy American Act mandates federal agencies purchase domestically-made equipment, export controls treat defense technology like nuclear secrets requiring US citizens and facilities, and security clearances demand American corporate structures—meaning even a company winning a billion-dollar Pentagon contract must operate through a Maryland intermediary rather than sell directly from Tel Aviv.

Despite all this, Israeli defense tech startups doubled in 2024. The reason? Ukraine and Gaza became brutal proving grounds—battlefields where systems either worked or got soldiers killed. That real-world validation now matters more to Pentagon buyers than decades-old relationships with Boeing or Lockheed. For example, Reuters reports “Around 50% of anti-drone technology used by Israel’s military during the conflict came from startups”.

American venture capitalists are flooding into Israeli defense tech precisely because the traditional defense giants are slow, bureaucratic, and optimized for lobbying rather than innovation, while Israeli startups are shipping battle-proven drones and AI-targeting systems that actually perform when missiles are incoming—creating a possible shift where the next wave of defense unicorns is emerging from a country smaller than New Jersey rather than the Washington Beltway’s contractor fortresses.

The IDF’s Killer Map App

In Gaza, city blocks disappear in hours. Buildings collapse, landmarks shift, and street grids once used for targeting no longer exist by nightfall. For decades, Israeli infantry entered combat zones with compasses, plastic-covered paper maps and grease markers, relics from an offline age. During the October 7 War, that analog era ended.

TheMarker (Hebrew) reports that the vast majority of IDF ground forces now rely on OLAR, a smartphone-sized digital mapping device built by Israeli startup Asio Technologies, which provides real-time battlefield visualization even under GPS jamming or comms denial. The features-packed pocket platform offers the kind of situational awareness typically reserved for air and drone units—now in the hands of frontline infantry. Israel’s air dominance is well established. What’s new is its ground dominance—faster, more precise, more coordinated than anything on record.

But don’t call it a map though: Asio says OLAR is a “geo-spatial position acquisition systems for defense and paramilitary markets.”

Also known by its original name, Orion, the device fuses real-time 3D terrain visualization, positional tracking, mission planning, and intelligence overlays into a single handheld tool. It doesn’t just show where a soldier is—friendly positions are marked in blue, enemy sightings in red. A new route flashes across the screen, avoiding a suspected ambush site identified by drone feed just seconds ago. No radio call. No paper. Just tap, confirm, move.

Think of it as a battlefield iPhone that works even when there’s no signal, no Wi-Fi, no GPS. No problem. The device works offline, drawing from cached satellite data, sensor inputs, and stored mission parameters. When connectivity returns, it syncs. It’s rugged, encrypted, and it never waits for permission to be useful.

This isn’t Google Maps for soldiers—it’s what Google Maps can’t be. Civilian platforms blur sensitive zones, rely on commercial satellites, require constant connectivity, and offer no tactical overlays or access controls. Worse, GPS spoofing or signal jamming—now routine in Gaza and Ukraine—renders most consumer apps useless. OLAR works where comms fail and terrain changes hourly. It’s not just a navigation tool. It’s like Waze for Warfighters.

Founded in 2003 by brothers Tomer and Yaniv Malchi, Asio spent years in relative obscurity, refining its mapping tech for the harshest conditions. Then Gaza became the proving ground. The company grew from 20 to 50 employees in under three years. Revenues are nearing ₪100 million.

The implications extend beyond Israel. OLAR is part of a broader shift in military procurement: away from lobbyist-optimized giants and toward fast, interoperable, war-tested systems. And in this case, one of the most consequential frontline upgrades of the war didn’t come from Lockheed or Rafael—but from a Tel Aviv startup shipping battlefield software in a form factor a soldier already uses for things like Google Maps and Waze. The next war will be fought with handheld command and control systems, AI-augmented vision, and decentralized teams moving with real-time intelligence. And unless they adapt, most Western ground forces are still carrying the 20th century into the 21st.

A Financial Blueprint For Super-Sparta

Israel’s defense chief just made the “Super Sparta” thesis operational. Maj. Gen. (res.) Amir Baram told the Finance Ministry’s Accountant General Conference that Israel is fusing military power and industrial policy into one system—small population, permanent technological edge, economy organized around conflict readiness.

Since October 2023, Israel absorbed ₪205 billion in war costs while closing $2.5 billion in government‑to‑government export deals in a single week, created a National Armament Council to compress procurement cycles, creating a joint Defense‑Tech fund with the Chief Scientist, and is pulling critical production in‑house through vertical integration—controlling multiple steps from components to final assembly—to survive sanctions, vandalism of shipments, economic shocks, and geopolitics.

The plan runs on three tracks: immediate purchases to plug readiness gaps, decade‑scale force expansion, and long‑horizon “game‑changers” for future wars. This taps a roughly $2.7 trillion global defense market growing about 20% annually, where battlefield validation sells.

Baram set explicit goals: domestic independence in systems where foreign dependence creates vulnerability, faster decision cycles than adversaries can match, and using combat‑proven performance to win export contracts. Export revenue finances autonomy.

But capacity is the constraint: every system shipped abroad competes with domestic stockpiles as Iran rearms (Baram said further conflict with Iran is inevitable).

The picture Baram painted is the institutionalization of Super Sparta, an operating system where growth, deterrence, and foreign policy share the same assembly line—and where peacetime is just the resupply window.

Share Israel Tech Insider

As always, feedback to amir@israeltechinsider.com

Israeli AI Researchers Face New Publishing Dilemma

Israeli AI researchers are reconsidering a basic academic practice: uploading papers to arXiv, the open-access platform where scientists share work before formal publication.

arXiv serves two functions. It timestamps research to establish priority—proof that you had the idea first. And it makes work visible to the global research community immediately, rather than waiting months for conference acceptance.

For Israeli researchers, that visibility now carries risk. Academic conferences use double-blind review—neither authors nor reviewers know each other’s identities. But if you upload your paper to arXiv before the review process, your name and institutional affiliation become public. Reviewers can find it. And some Israeli researchers, like many others in the academic and broader tech industry, believe their nationality—until recently a byword for startup excellence— now works against them.

“I encountered the claim that it is recommended to avoid uploading papers submitted to conferences to arXiv, to avoid exposure to reviewers with anti-Israeli tendencies,” one researcher wrote in an Israeli tech forum. Another asked how to present postdoctoral work without mentioning their Israeli affiliation after facing rejection.

The debate reflects a bigger shift: almost 2 years into the October 7th War, Israel’s global reputation, its war economy, and severe internal stresses, are taking a toll on its tech industry.

“Since the events in Gaza it’s harder to find collaboration with foreign universities—both the atmosphere and funding have dropped.”

Investors who previously engaged with Israeli startups now ask more questions. Some Israelis report feelings of being “singled out” or treated differently in international research forums, conference reviews, or funding decisions, prompting advice on how to present one’s institution and nationality more neutrally. Some mention being passed over for foreign roles or losing remote contracts.

“Does anyone know how to present my postdoctoral work without mentioning I am from Israel? I got a rejection last time.”

The growing concern over global isolation—supercharged by Netanyahu’s recent Super-Sparta remarks—is reflected across Startup Nation’s tech Facebook groups, which are a virtual water-cooler where you can tap into the dominant conversations on the minds of Israeli techies. [[If you’re new to my reporting from these groups, check out Israeli Tech’s HR Horror Show]].

Academic researchers face competing pressures. Upload to arXiv too early, and you expose your identity during blind review. Wait until after acceptance, and you lose the timestamp that protects your work if someone else publishes similar research first. Upload anonymously—stripping your name and affiliation—and you risk someone taking credit for your ideas if the paper doesn’t get accepted.

AI research, because of its growing salience in everything from chatbots to killer satellites, requires showing up openly in global forums and being peer-reviewed without political bias. The question isn’t whether Israeli researchers can adapt—they clearly are. The question is whether adaptation under isolation produces the same results as operating at full capacity in global networks. The answer will show up in publication rates, collaboration networks, and funding flows over the next year.

The fact that this trade-off even exists tells you something about what it’s like out there for Israeli talent.

CEO Selfie With PMO

The guy with the mustache is Guillermo Rauch. Argentina-born Guillermo is the founder and CEO of a Silicon Valley-based startup called Vercel that helps developers build websites and applications, including tools powered by artificial intelligence. Its customers include OpenAI, Anthropic, PayPal, Nike and Walmart.

On Monday (September 29, 2025) Guillermo posted a photo from a New York meeting with Prime Minister Netanyahu.

“Enjoyed my discussion with PM Netanyahu on how AI education and literacy will keep our free societies ahead. We spoke about AI empowering everyone to build software and the importance of ensuring it serves quality and progress. Optimistic for peace, safety, and greatness for Israel and its neighbors.”

Empowering literacy, free societies, building quality software, optimism, peace, safety…who could argue with that??

A lot of people, apparently. The post prompted backlash from developers accusing Rauch of supporting Netanyahu’s policies (they used harsher words). A Vercel employee named Wesam resigned publicly, several users canceled subscriptions, and Google Trends showed a fivefold surge in searches for Vercel alternatives.

One of those alternatives is a company called Replit, run by a guy called Amjad Masad, who, you would be surprised to learn, is openly anti-Israeli.

Amjad jumped at his chance to globalize his intifada by offering to migrate, at his own expense, anyone quitting Vercel over to Replit. He made a video, turned it into a Migration Guide, and went viral. Respect. The dopamine hits must have been epic.

The next day Guillermo and Vercel delivered their answer to Amjad and his like-minded filter-bubble buddies.

How do you like them apples, Amjad?

Share

How important is tech to Israel’s economy?

Latest numbers by Jefferies, the investment bank, on just how critical Israel’s technology sector is to the country’s economy:

• ~20% of GDP

• 64% of exports

• 12% of workforce is employed in the high-tech sector

• 25% of income taxes

• 30% of the market cap on the TASE

Mr. Security Committee

When leaders can’t choose, they convene—and Benjamin Netanyahu has elevated this evasion into an art form. His latest creation: the Supreme Armament Council, a new committee to figure out how Israel can make enough bombs and bullets by itself, cutting its reliance on just one or two suppliers (thank you President Biden).

The immediate trigger was Netanyahu’s “Super Sparta” speech blaming civil servants for procurement obstruction, yet the real obstruction sits in his own office: three years ago, an expert commission delivered detailed recommendations for making Israel’s weapons production less dependent on foreign suppliers, but Netanyahu never implemented them—not because bureaucrats blocked the reforms, but because he refused to choose between his finance ministry’s budget discipline and his defense establishment’s spending demands.

The new council covers ammunition production but excludes major weapons systems like submarines, fighter jets, and warships—ensuring it can neither succeed completely nor fail in a way that assigns clear responsibility. Meanwhile, Israel’s ammunition storage remains exposed and domestic production capacity constrained, precisely what that earlier commission was meant to fix (Nagel Commission report). The pattern echoes America’s post-9/11 defense sprawl and Britain’s post-Falklands spending: emotional wartime expansion that erodes the economic base sustaining military power. What Israel needs isn’t another council—it needs Netanyahu to publicly rank defense priorities with locked budget allocations and mandatory parliamentary review. Sometimes effective governance just means forcing leaders to actually govern.

As always, feedback to amir@israeltechinsider.com

"We're going to be Athens. And Super-Sparta." --Benjamin Netanyahu, son of a historian, 15 September 2025.

Ancient Sparta dominated Greece for centuries through military excellence and a rigid economic model that prized self-sufficiency, a proto-Autarky. Warrior-citizens, when not actually fighting wars, trained for war; helots—state-dependent serfs—worked the land; perioikoi, free non-citizens, handled craft and trade. That system produced formidable soldiers and a mighty regional hegemon. But it also bred brittleness. When Theban general Epaminondas smashed the myth of Spartan invincibility at Leuctra in 371 BCE, Sparta could not adapt; within decades, it had slipped from hegemony to the margins. Collapse followed.

Benjamin Netanyahu, the son of a historian, invoked a version of this history on Monday, telling the Accountant General’s Conference that Israel may need to become “Athens and super-Sparta” and “adapt to an autarkic economy” amid rising diplomatic isolation in Europe. He linked that isolation to “unrestricted migration of significant and very militant minorities,” and said Israel must prepare to produce and procure more at home. Although his office later broadcast two rapid-fire clarifications—the prime minister only meant autarky in the defense industrial base— the markets reacted badly.

It’s unclear if Netanyahu was presenting to the accountants in the auditorium Israel’s new non-GAAP fiscal measures (Generally Accepted Autarky Practices), or whether he was giving a convoluted, pseudo-historic lecture cooked up on one of his recent tech-bro podcast appearances. Either way, he made the claim in a room full of accountants. But the numbers don’t add up. A brief look at Super-Sparta’s financials shows why.

• Israel imports far more goods than it exports — creating a large deficit in tangible items like oil, machinery, and raw materials. But its trade in services more than makes up for the gap. Israel exports around US$86 billion in services and imports about US$45 billion, which leads to a net surplus when you combine goods and services. In fact, exports of goods + services are about 31.9% of GDP, versus imports of 28.6%, producing a modest overall trade surplus of roughly 3-4% of GDP (mostly made up of 0’s and 1’s).

• Israeli engineers design software for European banks, provide cybersecurity for American corporations, and develop algorithms for global markets—intellectual property that travels instantly across continents without requiring ships or trucks. Autarky would sever these electronic lifelines, forcing Israel to abandon its comparative advantage in high-value services while attempting to produce low-value physical goods domestically.

• Energy is the hard stop. Israel is self-sufficient in natural gas, but imports essentially 100% of crude oil—about 200–250,000 barrels a day in 2022. Renewables are rising but modest: 2024 saw ~15% of electricity from solar, with gas ~70% and coal in the teens.

• Food is similar. Israel imports about half its supply, especially staples: ~80–90% of wheat and most fish. The result: a persistent agri-food trade deficit of $6–7B. Autarky would mean replacing those flows in a country where agriculture has already been hit by war and a labor exodus since October 7

• Manufacturing inputs aren’t any easier to secure. Turkey has suspended trade with Israel, cutting off imports of industrial goods like metals, cement, and machinery. Meanwhile, attacks in the Red Sea have disrupted shipping routes, forcing longer journeys and driving up freight costs. Israel depends heavily on sea routes for most of its food imports because overland alternatives are limited. Autarkic policies would make these vulnerabilities worse—because there’s no way to manufacture everything domestically, especially raw materials like copper or rare earths.

• Netanyahu’s vision would shatter the institutions that sustain Israeli prosperity. Israel participates in virtually every pillar of the global economy: OECD, WTO, WIPO, IMF, World Bank, ITU. These memberships secure IP protections, trade rules, and R&D partnerships. Autarky would mean abandoning them—strategic self-immolation disguised as independence.

You can’t improvise a defense-industrial base

Israel’s defense exports hit a record US$14.7-15 billion in 2024 — up about 13% from 2023 — driven largely by missiles, rockets, and air-defense systems, which made up roughly 48% of export volume. Over half of its defense contracts were with European nations (~54%), with Asia-Pacific and the United States making up smaller shares. Yet Israel faces growing diplomatic and commercial pressure: some European governments are suspending arms deals, trade shows are banning or restricting Israeli firms (e.g. bans in Paris, constraints at DSEI), and several countries have imposed embargoes or halted military exports in protest over its conduct in Gaza. This is why strategic autonomy in armaments is prudent, but autarky goes too far.

Programs like Arrow and David’s Sling rely on U.S.–Israeli co-production and multinational supply chains; production lines depend on imported energy, alloys, electronics, and propellants; and exports fund R&D. Cut Israel off from markets and inputs, and you starve the very complex you want to scale.

This is more than “just” mass manufacturing tanks, bullets, shells, ceramic vests, lasers, manned warplanes, unmanned warplanes, robot dogs, mosquito spies and Mossad dolphins in factories that don’t exist yet. Even if/when these factories are built, good luck keeping them all supplied and protected in a country the size of New Jersey surrounded by enemy missiles and drones.

Autarky’s Murky Record

The Spartan model held only while fresh conquests provided fields for farming and a subjugated workforce (the Spartans were not nice people). When they started losing wars, the base cracked. By the time of Alexander the Great, Sparta was a spent force. The analogy to modern Israel is limited, but the core lesson remains the same: a closed system is brittle. Small, trade-dependent states that tried “go-it-alone” fare poorly.

North Korea proves the point. Despite sharing identical starting conditions with South Korea in 1945, the hermit kingdom's self-imposed isolation has produced an economy worth just $27.86 billion, or just 1.7% that of South Korea's. The human toll is equally stark: 40% of the population remains undernourished, electricity blackouts plague everything beyond Pyongyang, and economic growth has averaged just 1.4% annually since 2000—barely keeping pace with population increases.

There is no credible model for Israeli autarky because the premise is unrealistic.

Super-Sparta Would Stifle Startup Nation

Israel's tech sector—the country's economic engine—faces a crisis that autarky would turn into catastrophe. Since October 7th, tech companies hace lost over 8,500 employees to emigration, while 30% cite military call-ups as their primary challenge. Projects stall, deadlines slip, and some soldiers return from duty to find their jobs gone. A "Yad Labanim" study found that 34% of first-degree relatives of fallen soldiers from the war have not returned to work. This isn't the brief conflicts that Israeli tech previously shrugged off—it's sustained bleeding of the talent that keeps the economy alive.

Netanyahu’s vision would finish what judicial reform and war started. Israeli tech thrives on global sales: 75% of $12B in 2024 startup funding came from foreign investors; exports worth $73.5B need global markets. Cutting lifelines would kill the sector. It’s already happening: nearly half of new startups (46%) are choosing to incorporate outside of Israel. In 2022, 80% of startups incorporated in Israel.

Israel cannot afford to destroy its cash cow in wartime. The country has already spent $87B fighting, debt has passed 70% of GDP, and growth crawled to 1%. Tech exports pay the bills.

The verdict

Sparta’s autarky produced military power but fatal fragility. Israel’s open, services-heavy economy wins by integration, not isolation. “Super-Sparta” is a slogan, not a strategy. Strategic autonomy in choke points like armaments and semiconductors is prudent; autarky is folly dressed up as stoicism.

It’s also important to note that while Israel may need to become an autarkic Sparta to survive international isolation, it cannot do so while its prime minister exempts the ultra-Orthodox—the fastest-growing demographic—from military service, taxation, and secular education, creating an unsustainable economy where those who refuse to fight, work, or learn modern skills are subsidized by those who do. Netanyahu compounds the strain with bloated ministries that serve coalition politics, not national resilience.

Netanyahu has conjured the perfect trap. His judicial overhaul triggered the brain drain, his war policies are decimating the workforce and international connectivity, and his coalition politics protect the ultra-Orthodox from contributing to either military service or economic productivity. Having engineered the crisis through political choices, he now prescribes autarky as the cure—the surest way to destroy what remains of Israeli prosperity. Sparta fell because it couldn’t adapt. Israel risks the same if it mistakes slogans for strategy.

As always, feedback to amir@israeltechinsider.com

Note: This article has been substantially revised following investor feedback that clarified the distinction between deal sheet visibility and actual investment activity, correcting the initial mischaracterization of foreign investor behavior as an "exodus" when evidence suggests continued engagement through larger, longer-stealth funding rounds.

An earlier version of this post said that foreign investor activity in Israeli tech has cratered to historically low levels, with established players like Index Ventures and Andreessen Horowitz completely absent from H1 2025 deal sheets.

But thanks to a smart investor friend, I’m humbly able to say that the picture is a more nuanced than I previously thought. H1 2025 data from the new IVC-GNY-KPMG report suggests foreign investor exodus—but the mechanisms of how this industry works reveal a more complex recalibration of risk, timing, and stealth capital deployment that challenges my initial surface-level panic narrative.

Foreign VC first investments indeed dropped to 49% of total deals, breaking the 60% historical average maintained from 2016-2024. Yet this metric captures disclosure moments, not funding decisions. While there are fewer of them, the early-stage rounds that have taken place have been dramatically bigger, allowing startups to extend stealth phases without immediate follow-on pressure.

And while Andreessen Horowitz and Index Ventures did not, as of the IVC report release, show up in H1 deal sheets, A16Z has in fact organized Israeli founder events during the summer. It apparently even caught flack on line for that.

Corporate venture capital presents the clearest retreat signal. Foreign corporate VC first investments crashed from 44 deals in 2022 to 4 in H1 2025. But context matters: Samsung Next and Verizon Ventures shuttered failing Israeli operations years ago, preferring centralized corporate venture strategies over dedicated local presence. Their absence reflects operational efficiency, not geopolitical flight.

Domestic funding shows genuine stress signals. Israeli institutional investors directed 79% of their $87 million toward follow-ons rather than new ventures—a defensive posture that prioritizes portfolio preservation over ecosystem expansion. Twelve new Israeli VC funds raised a modest $260 million in H1 2025, projecting toward a 60% annual decline from 2024's already-compressed $1.27 billion.

Investment concentration reveals strategic prioritization. Generative AI leads all sectors, while cybersecurity attracts "enormous sums"—Cato Networks' $239 million round and Sentra's $50 million raise exemplify capital flowing toward conflict-adjacent opportunities. Defense tech resurged as a focal point, driven by immediate security demands.

Life sciences suffered the most: only 23 new portfolio additions among Israeli VCs, down from 129 in 2022.

Israeli VC funds still hold approximately $11 billion in available capital—a war chest. The Israel Innovation Authority launched "Startup Fund" and "Yozma 2.0" initiatives to increase market liquidity, suggesting institutional recognition of opportunity amid crisis.

With larger rounds funding longer stealth periods, and defense/AI sectors attracting premium valuations, the ecosystem appears to be consolidating around geopolitical advantages. The question is whether Israel's innovation pipeline can maintain diversity while the world's attention—and money—fixates on its security innovations.

Share

I’ve been on a much-needed vacation so please excuse the lapse in publishing. I’m back at my desk now, with some fresh reporting and perspective. As always, feedback to amir@israeltechinsider.com

Let’s dive in.

War as a Feature: How Israeli Tech Learned to Thrive in Permanent Crisis

Day 669. Nearly two full years into the war with Hamas, Israel’s tech sector, like the rest of the country, faces one of its most critical inflection points in recent memory. As tens of thousands of reserve soldiers get called up again—many from startups, many from elite units—the weight of sustained conflict is no longer abstract. It is operational, financial, personal. It’s the new normal.

Over the past 20 years, the tech industry has treated war as a recurring bug—small, contained conflicts lasting days, weeks, or, in the case of the 2006 Second Lebanon War, a month max. Tech was teflon to these bugs. Companies were founded, scaled, bought and sold during these short wars. Nothing seemed to knock tech off its stride.

The October 7 war’s length—it could be heading into its third year next month—is Israel’s longest by far. It has become a feature, not a bug. And it is transforming the tech industry, as much as the industry is transforming war. This creates competitive advantages (chaos management, accelerated decision-making, dual-purpose dissemination) but also structural vulnerabilities (conflict dependency, talent fragmentation, risk-conscious capital flight).

Within this, my sense is that Israeli tech has evolved beyond crisis management into crisis optimization. What began as adaptation to temporary conflict became systematic transformation for permanent instability. Israel isn’t built for long wars. Israeli tech is building itself for just that.

The industry is adapting, hardening, and in some sectors, flourishing. In many places, it is also quietly fraying. Through private conversations with startup founders, investors, and ecosystem operators, seven core themes have emerged that point to a vibe shift: Israel's tech sector didn't just adapt to crisis, it is at once devolving back to its security DNA while also evolving into a more disciplined and mature version of its “traditional” peacetime operating models.

1. Nothing is Normal And That’s Totally Normal

Israeli tech operates on borrowed sovereignty. When the Department of Government Expenditure cuts U.S. federal budgets, the tremor hits Tel Aviv's enterprise sector within months. Healthcare buyers, for example, who purchased Israeli cyber solutions suddenly freeze procurement. The exposure is structural: Israeli companies are deeply intertwined with U.S. enterprise buyers and VCs, even as relationships with Europe reach breaking point. Founders report subtle shifts—longer decision cycles, softer demand, heightened due diligence. For some global financial institutions, Israel’s ongoing success in tech now sits uncomfortably alongside a political narrative they no longer understand—or trust.

“Even wise, honest people… they do not fully understand, LPs are concerned or confused about Israel’s trajectory, and it seems to them that something has gone really wrong.”

2. Cyber and Defense: Conflict as a Service

Cybersecurity and defense have proven…resilient. Even as the war drains national morale, global trust, and public coffers, these 2 sectors have attracted capital and credibility. Defense exports are up. Cyber valuations are strong. Defense technologies, especially ones in unmanned systems, are benefiting from a surge in global demand for battlefield-proven innovation. But this growth comes at a cost: it is narrow, concentrated, and often reliant on ongoing conflict for momentum. Founders quietly ask: what if our over-dependence on the US turns against us? What would our growth look like if and when the guns fall silent? (I, for one, think that would be a good problem for Israel to have).

“The huge demand in defense might make it a little bit challenging for some companies to do non-defense stuff.”

3. Human Capital: Missing in Action Items

War doesn’t just deplete the labor force—hundreds of thousands of Israelis are on semi-permanent military rotation—it scrambles leadership. Entire executive teams have been pulled into reserve duty. Founders vanish for months, returning distracted, “altered”, or emotionally drained. Some don’t return. Business continuity plans are stretched beyond design. Investors report delays in product delivery, missed fundraising windows, and operational slowdowns. While the talent pool remains deep, its availability is fractured, and companies are paying the price in execution.

“I have two CEOs in two companies who disappeared for four and a half months. They came back a little different.”

Yet something has emerged from the battlefields: Israel’s best talent is flowing into defense and cyber. This is benefitting the dual-purpose military-commercial application space. The question is: how long can the humans of Startup Nation carry on like this?

4. What Are You Doing For War’s End?

There is a quiet consensus among Israel’s top investors: if/when the war ends, innovation will roar back. Call it the roaring mid-twenties. But if/when we get to the end of the war, what is the scene we’re likely to see?

Unless you’re in cyber and defense, startup funding has slowed dramatically, and the circle of startups able to raise has narrowed. The number of new startups being founded in Israel has also dropped substantially, according to data from Startup Nation Central.

If and when political conditions improve, there will be time for a damage assessment. But the long-term outlook remains optimistic. New companies will emerge, talent will recycle, and infrastructure will be rebuilt. But this optimism is not evenly distributed. In foundational AI, for instance, Israel is playing catch-up. The current wave of innovation is thin on deep tech too, which is important enough to get its own section below:

“Post-crisis is the best time to invest.”

5. Israeli Deep Tech is in Deep Sh#t

What makes tech deep? Deep tech is built on hard-to-copy science like advanced chips, quantum hardware, new materials, robotics, energy storage, satellite systems, and biotech. These are the things that propel companies and nations forward, or keep them back if they’re not supported. They take years, cash, and sophisticated physical infrastructure to mature; they need labs, test rigs, clean rooms, and often factories rather than just cloud servers.

What’s happening to Israeli deep tech? Not good. The numbers are blunt. In the first half of 2025, Israeli deep tech raised $597m across just 22 deals; all of 2024 saw $536m across 51. Specialist funds fell from 25 to six. Of 247 active firms, roughly 70% are still in R&D or only starting to sell. Money is returning, but only to a few. [[Source: IVC Data & Insights, Gornitzky GNY, and KPMG Israel.]]

Less money going to new and growing research-led companies means Israel falls behind on the really, really important stuff. Apart from capital clustering around winners working on national priorities—security, energy, and critical infrastructure—the government might need to intervene here in a big way to boost new deep tech company formation and growth.

6. Chaos Junkies: Ego Is The Enemy

Investors emphasized a shift from traditional competency to adaptability. The qualities investors are now prioritizing: coachability, flexibility, resilience. In today’s climate, fixed mindsets are liabilities. Executives who have managed companies through wars, navigated uncertainty as routine, and maintained operations during existential crises possess skills the global market increasingly values.

Smart companies now hold monthly leadership reviews asking: “If we were launching this company today, would we build it the same way?” DEI is DOA. Managers aren’t asking how you feel, they’re asking how fast you can ship. Execution cycles are shorter. Strategic plans are obsolete within months. Your top software engineer can be called up for duty at any moment. Leadership is just as much about absorbing reality as setting vision. Founders are expected to operate with humility, urgency, and data.

"We here know what it is like to work, to study, to live, during wars, in a routine that is not a routine. What happened to us in the last three years... manifests with much more determination and resilience, and in this age, it is very beneficial for companies."

7. First We Take Manhattan. Then We Stay There

Zoom is no longer enough. Customers—especially in the U.S.—expect Israeli startups to be on the ground, culturally in sync, and physically present. Founders are relocating again, bucking the pre-2023 trend of growing their companies from Israel. Sales and GTM functions are long gone. Flights are tricky and expensive. Many are between reserve duty and business trips.

The U.S. has become Israel’s primary lifeline. To scale in 2025, Israeli companies feel they need boots on the ground in America (whenever their boots are not on the ground in Gaza).

Venture fund 97212 reports a sharp rise in Israeli-founded or Israeli-led startups in New York, counting about 560 active companies plus roughly 80 in stealth, totaling around 640—significantly higher than earlier tallies. Political and security turbulence in Israel is cited as a catalyst for relocation, amplified by the support of a now-dense Israeli tech community in New York.

“The Americans like to feel it… not just the Israeli playbook, but to prove the business in the United States.”

Closing Thoughts

None of this should be surprising. The stress, and strength, are real: on the 699th day of war, Israel’s tech sector sits on a knife’s edge, with a knife between its teeth. Resilient but exhausted. Tapped out but with a few tricks still up its sleeve.

For now, the ecosystem is holding, hardening and narrowing—like a heart artery in a person with hypertension, interrupted sleep, lots of money problems, and an addiction to chaos. But with a fresh wave of reserve call-ups, global investors watching closely, and Israel’s international reputation withering, Israeli tech is bracing, either for a resurgence or reckoning.

If and when the fighting ends, the true test will begin: is Israeli tech able to beat its Swords back into SaaS?

As always, feedback to amir@israeltechinsider.com

It’s Now or Never Again

In the shadow of an escalating global AI arms race, Israel’s left AI hand doesn’t know what its right AI hand is doing—and the country’s AI startups are left scratching their heads. I’ve written about this before but every week there’s new evidence that Israel is heading for an AI Iceberg.

NB: AI ASAP

What’s Happening: The Ministry of Innovation, Science, and Technology’s April 2025 baseline report paints an Israeli AI success story. Israel ranks first globally in AI “citations per paper” and “AI patents per capita” over the last decade. But a brand new 120-page report (Hebrew) from the National Committee for Accelerating Artificial Intelligence, says we’re headed into an AI Iceberg and the whole ship will sink.

Chaired by former National Security Adviser Prof. Yaakov Nagel, the AI committee has issued what reads like an AI fire alarm:

“Israel is not at the right point, and not even at the desirable trajectory… Without swift and adequate action, the country will experience a further and rapid decline in its global status.”

Both reports can be true. Israel tops citations and patents per capita, yet its operating environment fell 42 places and infrastructure ranks 26/83—the research edge persists as the backbone erodes.

We Put Our People Fist

According to Nagel’s analysis, Israel has roughly 120 researchers focused on core AI (foundation models, training infrastructure, and fundamental algorithms). By comparison, a single U.S. hub like UC Berkeley fields on the order of 70-plus in similar areas. Israeli AI brain drain is rampant, doctoral pipelines are weak, and academic infrastructure is lacking. “The absence of coordinated national investment,” the report states, “has created a critical bottleneck in human capital.”

Unlike cybersecurity, where tactical skills can be trained on the job, generative AI demands deep, academic fluency. The committee insists: “It must begin in high school. It must continue through graduate research. There are no shortcuts.”

Translation: Israel (still) has the brains (for now) but no backbone.

Energia

Where the digital revolution was about talent and code, the AI revolution is about compute, data, and energy. Training a frontier model today consumes more electricity than cities—and Israel’s energy grid, the report warns, is ill-prepared. There are no dedicated provisions for AI data centers. Grid connection can take years. No national plan exists for green energy at the scale AI will require. Israel, despite its vaunted tech ecosystem, has no national supercomputer.

“The energy architecture of the future is becoming the sovereign enabler of AI power. Without it, everything else becomes irrelevant.”

Dejargonize: “Compute” is the factory floor of AI: clusters of GPUs that train and run models. “Time-to-GPU” is how quickly a researcher or startup can get capacity; long waits slow R&D cycles and deal flow. “Watt-per-FLOP” is simply cost-efficiency: how much electricity you spend per unit of AI work. Lower is better; countries with cheaper, stable power and modern data centers can build and run models more competitively.

And “open, de-identified state datasets” matter because health, mobility, finance, and security breakthroughs depend on high-quality, privacy-safe data only governments hold. If ministries hoard it, startups build on scraps—or go abroad.

In plain English: the more Israel depends on foreign power and foreign clouds, the more profits, intellectual property, and tax receipts migrate with them.

“I imagine, therefore I become”

Here’s a case study. Founded in 2023 by Technion alum Dr. Dean Leitersdorf and Moshe Shalev, AI startup Decart just raised $100 million at a $3.1 billion valuation to generate real‑time video “worlds”—instant, deployable content cheap enough for scale. That’s exactly the kind of company national strategies claim to want: core technology, high-capex training needs, global demand. The committee’s point is that Decart—and dozens like it—scale today on “foreign rails,” renting compute in Dublin or Oregon because getting reliable power and large GPU clusters at home is slow and costly. It works—until an energy shock, war damage, export control, or cloud capacity crunch turns a business hiccup into a national competitiveness, resilience, and security problem. “Companies can rent GPUs; countries can’t rent sovereignty,” Prof. Nagel writes.

Decart is far from alone. Israel hosts roughly “210–320 ‘core AI’ startups”—not just apps, but the plumbing: infrastructure, model tooling, training and inference platforms. Core AI firms need sovereign compute, cheap/stable power, and open data to iterate fast; they are the upstream that feeds everything else. This crop of new AI companies are the Jewish Nation’s Crown Jewels (Israel’s Crown Jews if you will) and it should build them the best possible operating system.

En-Gulfed

But…that’s not what’s happening. When the state can’t promise capacity and data access, as well as an AI education pathway, AI is passing Israel by. The global race won’t wait.

The UAE has committed $100 billion to create its own sovereign AI infrastructure, including foundational models in Arabic, and crucially, the clean energy infrastructure to power hordes of new data centers. The tiny Emirate is projected to rank third globally, after China and the US, in AI’s contribution to the economy. The UAE has a strategic vision and a Minister of Artificial Intelligence. Guess who Israel’s AI Czar is?

U.S. giants are in a historic buildout of data centers and the energy to run them.

Israel does not need to spend as much as the Gulf monarchies and Big US Tech firms. Also, it can’t. Israel maintains its edge through innovation and talent, despite smaller absolute spending levels. The country’s focus on research institutes, talent development, and leveraging existing tech sectors for new ideas, is still an effective approach to AI leadership.

But infrastructure matters. Sovereign AI infrastructure, including national power grids and data processing (compute) facilities are going to matter more and more. Relying on a small talent pool and the infrastructure of others poses national strategic risks.

Nagel Says

What the committee actually proposes is a tight operating plan with a scoreboard:

• Governance and authority

  • “Establish a lean National AI Directorate reporting to the Prime Minister,” with “DARPA‑style authorities to bypass standard procurement and hiring constraints.”

  • “One owner, one budget, one clock.” The Directorate should control funds, timelines, and cross‑ministry coordination to avoid diffusion of responsibility.

• Time‑bound deliverables (inside one budget cycle)

  • “60,000 GPUs for national academic use within two years,” alongside fast‑track access for startups (“time‑to‑GPU measured in days, not quarters”).

  • “A sovereign energy roadmap” that brings down “watt‑per‑FLOP” domestically, including expedited grid connections and data‑center interconnects.

  • “Open access to government datasets, with privacy preserved,” via stable APIs and a unified de‑identification regime.

• Budget and prioritization

  • “NIS 25 billion ($6.6 billion) over five years—about 0.25% of GDP,” framed as “a catalytic investment—no less strategic than national defense.”

  • Reallocate from slower, diffuse digital projects: “Architecture without appropriation is theater.”

• Deadlines and accountability

  • “If measurable outputs are not delivered within one budget cycle, shut it down and stop pretending.”

• Scope and sequencing

  • Start with “compute, power, data”—the “non‑substitutable enablers.” Talent and applications follow once the backbone exists.

  • Use “pilot‑to‑scale” execution: stand up one national research compute cluster, one expedited energy corridor, and one flagship open‑data domain (e.g., health) within 12–18 months, then replicate.

• Procurement and regulation

  • “Bypass slow procurement” with specialized authorities; adopt “outcome‑based contracts” tied to the KPIs above.

  • Mandate “interoperable data standards across ministries” and a common privacy‑preserving de‑identification framework.

• Strategic posture

  • “This is not an optional investment… it’s existential. If we are not in the top five, we are not in the game. Period.”

This is a departure from past theater. Israel’s 2018 AI plan—meticulous on paper, “never approved” in practice—died in budgets and bureaucracy. Nagel’s team insists on immediate execution with legal and budget authority, not another inter‑ministerial forum.

“This is an action plan, not a recommendation memo.”

Why YOU should care.

If Israel fails to be a major AI player—like it is in cyber and semiconductors—the consequences won’t only be felt inside its own borders.

This is a country that has contributed core advances in cryptography, chip design, big data analysis, and defense AI. Its researchers helped pioneer computer vision. Its companies power autonomous mobility, anti-money laundering, and medical diagnostics. If Israel becomes a secondary player in AI, the loss will be felt on Wall Street. For example: cybersecurity wunderkinds at WIZ built Microsoft’s Azure cloud security program before leaving to found the fastest-growing cyber startup of all time. Microsoft is now a cloud-first company in part because it bought Adallom, the WIZ founders’ first startup.

Amazon’s acquisition of chip design startup Annapurna Labs is “probably the most important moment in AWS history,” according to CEO Andy Jassy. AWS is powered by Annapurna-designed chips. NVIDIA’s Jensen Huang is expanding in Israel, replacing Intel’s footprint. The list goes on: Apple’s Vision Pro goggles, IBM’s AI, and Meta’s AR glasses are all developed here. Their stocks are driving global markets.

Sovereignty, or Supplier?

Israel can either be a top-five AI power or a top-tier AI supplier. Those are not the same thing. Suppliers make money; powers set rules. If Israel remains an AI power, it helps set standards, shapes supply chains, and seeds the next wave of tooling. If it becomes primarily a supplier, value still gets created—but “rule‑setting migrates.”

Right now, the private sector is winning deals. The state is losing the future. The single cheapest accelerator Israel could buy is policy: unify, de-identify, and open the state’s high-value datasets. Everything else multiplies from there. If this Directorate can’t deliver electricity, GPUs, and datasets within a budget cycle, shut it down and stop pretending.