BlackRock says AI compute could be tokenized in the future

BlackRock’s Digital Assets Research team just published a white paper that reads like a roadmap for a world where GPUs are the new gold bars. The paper, titled “The Machine-Native Economy: How digital assets connect intelligence, commerce, and compute,” lays out a vision in which AI compute power gets priced, financed, and traded as tokenized digital assets.

What BlackRock is actually proposing

The paper identifies three areas where AI and programmable finance are starting to converge: automated machine-to-machine payments, tokenized financial assets, and emerging markets for computing capacity.

That last one is the headline grabber. BlackRock’s researchers argue that standardized digital contracts, essentially claims on compute resources, could become tools for financing and programmable settlement. AI agents would transact based on factors like price, performance, and latency, choosing compute providers the way an algorithm picks the cheapest flight.

The scale of the opportunity isn’t trivial. The paper estimates that revenue from hyperscaler cloud services will hit roughly $1.1 trillion annually by 2030, representing a compound annual growth rate of 29% from 2025.

“Compute is emerging as a new and potentially large market for digital assets,” the paper states, a sentence that carries a bit more weight when it comes from a firm managing north of $10 trillion in assets.

The research also flags agentic payment protocols like x402 and the Machine Payments Protocol as foundational infrastructure for machine-initiated transactions. These are the plumbing systems that would let AI agents autonomously pay for the compute they need without a human clicking “confirm purchase.”

Context: BlackRock’s digital asset playbook

This isn’t BlackRock wandering into crypto territory on a whim. The firm has been methodically building its digital asset footprint for years, launching the BUIDL fund and tokenized money-market products that put traditional financial instruments on-chain.

The white paper doesn’t announce any new products or reference existing tokens. It’s a research document, not a product launch.

The fine print and what it means

BlackRock is careful to note that these markets are “largely speculative at this stage.” The firm emphasizes that robust infrastructure around permissions, identity, and compliance needs to be built before any of this becomes real.

The emphasis on machine-to-machine payments suggests BlackRock sees a future where AI agents are economic actors in their own right, not just tools that humans use. If autonomous systems need to purchase compute, data, and services continuously, they’ll need financial rails that operate at machine speed.

The $1.1 trillion revenue projection for hyperscaler cloud services by 2030 provides a concrete target for how large this addressable market could become. Even if tokenized compute captures a small percentage of that total, it would represent a market larger than most existing crypto sectors.