Israel's Sovereignty Stack Is Leaking
Capital, talent, and weapons are supposed to flow straight through. At every joint, they leak.
Israel’s Sovereignty Stack Is Leaking
Israel is building one thing in three stacked steps. Raise money. Use it to hire and keep engineers inside the country. Have those engineers build the technology and weapons Israel needs, so it owns and controls them outright instead of borrowing or licensing them from an ally. Call it a sovereignty stack: capital at the base, domestic talent in the middle, sovereign production at the top. The pitch, from the government and the funding headlines alike, is that each step converts almost fully into the next. Money into local jobs, local jobs into hardware Israel owns end-to-end. The evidence says it leaks at every joint.
The stack is being built now because the October 7 War made the old arrangement look dangerous. Twice during the current conflict, the allies who arm Israel used that supply as leverage: the Biden administration slowed shipments of 2,000-pound and 500-pound bombs mid-fight, and Britain and Canada froze export licenses for air-to-ground munitions. A supply chain a foreign government can pause during a war isn’t a supply chain. It’s a political relationship with a kill switch. The response was the largest industrial commitment in a decade, NIS 350 billion, about $110 billion, over ten years to build weapons Israel controls fully, alongside a broader push to lean less on any single foreign partner, currency, or cloud provider. Nice plan. Now let’s punch it in the mouth.
Start at the base, because the money is genuinely there. Israeli tech funding hit $7.6 billion in the first half of 2026, up 52% year over year, with $4.2 billion in the second quarter alone, according to a new IVC-LeumiTech Israeli Tech Review. Defense, Space & Quantum took a record 11.6% of all capital raised, roughly $846 million, on pace to double the whole of 2025 in six months. Defense exports hit a record $19.2 billion in 2025, up 30%. The foundation looks exactly like what was promised.
First, though, a bottleneck. The money isn’t leaking here, it’s pooling. Deal volume has been stuck near 100 rounds a quarter since 2023, down from roughly 140 in 2019-2020. Take the largest 10% of deals each quarter, the rounds that swallow most of the money: that slice now belongs to just 9 companies, against 22 splitting it at the 2021 peak. First-time investment, money going to a company raising for the first time, hit a record low of 35.5%, down from over 50% a decade ago. The vision was broad company formation. The reality is a handful of large winners absorbing the new money before it reaches a new firm, let alone a new hire.
The first real leak turns whatever survives into an Israeli salary, or what’s left of it. The Innovation Authority reports Israel lost roughly 3,500 R&D jobs in 2025, the first outright decline in a decade, even as funding recovered sharply, a finding independently corroborated by the Aaron Institute at Reichman University on separate data. Follow the money, and it runs into the exchange rate. The shekel has strengthened to multi-decade highs against the dollar since early 2024, so a dollar raised today in startup land buys fewer shekels of salary than the same dollar two years ago, eroding the purchasing power of the funding. Funding has more than doubled since October 7; the number of jobs it pays for in Israel hasn’t moved. A growing share of the rest goes to foreign-denominated cloud bills, not local paychecks. By March 2026, only 62% of employees at private Israeli tech firms were still based in Israel, down from 69% in 2019, the exact pipeline the top of the stack depends on, priced out by the exchange rate and pushed out separately by a thousand days of war, endless reserve call-ups, a rising cost of living, and a worsening political climate.
The second leak sits at the layer “independence” is supposed to describe. The NIS 350 billion arms program’s flagship projects are threaded with the dependency they were built to escape: the Israel Navy’s Reshef missile-ship hulls partly built by an American subcontractor funded by US security assistance, the Sa’ar 6 fleet’s hulls built at ThyssenKrupp in Germany, and a Raytheon-Rafael joint venture that opened a plant in Arkansas in November 2025 to build Iron Dome interceptors on US soil. That plant is the template for a new security memorandum now under negotiation, built on “reciprocal production”: each country making pieces of the other’s weapons on its own soil, replacing the one-way flow of American aid with something closer to a trade. The vision was full control of the weapons Israel fights with. The reality is dependence relabeled as partnership, on terms still written mostly in Washington.
The third leak is the most literal. Israel wants to be a regional AI hub, and its own grid can’t carry the load. On July 20, 2026, the Electricity Authority froze all new data-center connection requests for 140 days after pending applications hit 27,000 megawatts, roughly three times the country’s average electricity consumption. The 1,500 megawatts already committed would alone eat 10% of national power by the early 2030s. Money and talent mean nothing if a data center can’t be plugged in, and every stalled domestic project pushes AI workloads back onto the foreign cloud the stack was meant to escape.
The blunt freeze is the wrong tool for a real problem. Halting every request at once treats a financed project ready to break ground exactly like a speculative claim staking out capacity it may never use. The regulator’s own reported plan points the sharper way: charge developers an annual fee to hold reserved capacity, which prices out the paper claims and reveals who is actually building. Better still, let a serious project connect years sooner in exchange for a firm promise to cut its power draw during the handful of hours when the grid is most strained, running on its own backup instead. That separates builders from speculators without pretending the grid can serve everyone’s peak at once. Of the three leaks, this is the one with a clean policy lever within reach, which is exactly why it is the one worth acting on first.
None of this means the stack is broken. Exits hit a historic $64.75 billion this half in deals over $5 billion alone. Roughly NIS 6 billion, about $1.9 billion, is flowing to Israeli employees before taxes from the Wiz, CyberArk, and Armis cyber deals alone, Calcalist reports, money the state taxes and that Israeli tech reliably recycles into the next round of angel checks and startups. High-tech already provides roughly a quarter of Israel’s direct tax revenue. Foreign investors still supply about 69% of active VC funds, unmoved in a decade, and multinationals aren’t fleeing: 575 of their R&D centers remain active, seven more opened than closed in the first half of 2026. The stack works. It just loses more at every joint than the number at the base suggests.
Run it end to end. A dollar raised in Tel Aviv clears the bottleneck, survives the exchange rate, then reaches a defense program that still can’t build its own hulls without American or German hands, or a data center that can’t get on the grid. What arrives at the top, technology and weapons Israel genuinely owns and can power, is a fraction of what the base promised. No single leak is fatal. The three of them compounding are the story.
Time is money. The NIS 350 billion is already being spent, the connection queue already runs three times the grid, and the memorandum that defines what “reciprocal” means is unsigned. Israel is pouring concrete, signing contracts, and standing up production lines around a version of independence that neither Washington nor its own power grid has yet agreed to underwrite. Every round closing this quarter and every plant breaking ground is being priced against a sovereignty that is still, right now, a negotiation.