Larry Fink, the CEO of BlackRock — the world’s largest asset manager, overseeing over $11 trillion — has made a striking prediction: buying futures of compute could emerge as an entirely new asset class. At a time when artificial intelligence is reshaping every corner of the economy, Fink’s remarks cut to the heart of what’s becoming the defining resource constraint of the decade.
“The United States is short on power, short on compute, short on chips,” Fink said at an event. “There are going to be shortages in all three. And memory. That makes four things,” he added.
“I actually believe a new asset class will be buying futures of compute,” Fink said. “We just don’t have enough compute power right now.”
Fink’s argument is straightforward. The United States faces simultaneous shortages across four critical dimensions of AI infrastructure: power, compute, chips, and memory. Each of these constraints compounds the others — you can’t run chips without power, and you can’t run AI without chips. The scarcity across all four creates the conditions that have historically given rise to futures markets: a commodity in short supply, with intense and growing demand, and buyers willing to pay a premium to lock in future access.
The observation that compute could trade like a commodity isn’t entirely new, but Fink lending it the weight of BlackRock’s institutional authority is significant. When the CEO of the world’s largest asset manager says a new asset class is forming, markets listen.
The data backs Fink’s assessment. Data center construction has tripled since ChatGPT’s launch, rising to roughly $45 billion annually — and yet demand continues to outpace supply. Big Tech is projected to spend approximately $655 billion on AI infrastructure in 2026 alone, an unprecedented build-out, and still the shortages persist.
Power is perhaps the sharpest bottleneck. Datacenters now account for 7% of US electricity demand, and utilities are projecting continued double-digit growth in demand through the end of the decade. Microsoft CEO Satya Nadella has pointed out that his company has AI chips sitting idle because it simply cannot access enough electricity. Chips without power are just expensive paperweights.
The financial logic for a compute futures market is sound. If access to compute is uncertain, expensive, and rationed — and if it determines which companies can train and deploy AI — then organisations will want to hedge that exposure the same way airlines hedge jet fuel. Fink is essentially arguing that compute is becoming a strategic commodity, not unlike oil or electricity, and that financial instruments will eventually follow. Andrej Karpathy has made a similar structural argument, comparing LLMs to public utilities like electricity — homogenous, infrastructure-dependent, and essential. Utilities have futures markets. It stands to reason that compute, if it follows the same trajectory, eventually will too.
Whether the market gets there in two years or ten is an open question. But when the CEO of BlackRock says a new asset class is forming, it’s usually worth paying attention.