The AI race is creating a global competitiveness challenge and an infrastructure problem of a scale the technology industry has rarely had to confront. The more widely used AI becomes, the more computing infrastructure it requires – and the more electricity that infrastructure consumes.
The numbers are already enormous and growing exponentially. A recent Brookings Institution study estimates that investment in U.S. data centers and related AI infrastructure could reach $10.3 trillion between 2025 and 2032. The International Energy Agency expects global electricity consumption from data centers to more than double by 2030, driven largely by AI.
Much of the response has focused on how to produce enough energy to sustain that growth. Technology companies are securing new power sources, utilities are planning for soaring demand, and governments are beginning to confront the pressure AI infrastructure could place on electricity grids. But this is no longer just an energy issue; it is becoming a question of global competitiveness. China is expanding its power capacity far faster than the US and other Western economies, potentially allowing it to achieve a lower cost per token even with less advanced chips.
And it doesn’t stop there – there’s an additional side to the problem: AI doesn’t just need more energy, it needs to use compute far more efficiently. Even in the best scenarios, GPUs and accelerators in AI data centers currently operate at only 20% to 30% utilization, meaning expensive and energy-intensive chips spend much of their time underutilized while still consuming power. As AI infrastructure expands, that inefficiency could translate into hundreds of billions of dollars in wasted compute over the next few years.
With AI computing facing nearly insatiable demand, winning the cost-per-token race will require not only building more AI infrastructure, but also fundamentally improving the economics of running it. The US, EU, and others need new computing architectures that deliver significantly more compute with the same, or less, power. And for Israel, that problem presents an opportunity.
Israel has spent decades building deep expertise in semiconductors and advanced computing, with Israeli engineers playing significant roles in the development of processors, networking technologies and computing infrastructure at AWS, NVIDIA, Intel, Google and other campuses. Alongside them, a new generation of Israeli companies such as NextSilicon is developing new approaches to address today’s challenges and scale AI computing and high-performance computing efficiently. In addition, Quantum Computing is around the corner, and Israel has the potential to lead.
That gives Israel a rare opportunity that extends far beyond the economic value of building another successful technology sector. At a time of growing geopolitical uncertainty and intensifying competition over the technologies that will shape the global economy, Israel needs to turn technological leadership into strategic relevance. Developing a solution to a constraint that every major AI economy will have to overcome would create something far more powerful than another area of Israeli innovation – it would give the world a reason to depend on Israeli technology.
For a country of such a small size, that position will not come from matching the United States or China in capital, infrastructure, or scale, but from owning a technological layer that is impossible to replace. Taiwan achieved this through TSMC and the Netherlands through ASML. Israel now has a similar opportunity- to leverage its existing expertise to develop a solution to one of AI’s most fundamental constraints, and make itself essential to the infrastructure on which the global AI economy will be built.
The world is preparing to spend trillions building the infrastructure for the AI era, and compute efficiency could become one of its defining constraints. Israel is unusually well positioned to help solve it. The challenge now is to turn that technological advantage into lasting strategic relevance. Opportunities to become a critical part of a global industry do not come often, and it’s highly important to not let this one pass.
Ben Rabinowitz is the Founder and Managing Partner of Amiti Ventures.