Key points
BlackRock expects standardized compute contracts to trade as exchange-traded futures, with capacity claims pledged as collateral and settled on programmable infrastructure The forecast appears in The Machine-Native Economy, a research paper the asset manager published last week It cites Stripe's agreement to buy OpenRouter, reported at more than $7 billion in August, as the first strategic signal of that shift
BlackRock expects computing power to become a traded financial asset, with exchange-traded futures contracts, collateral claims and settlement on blockchains, according to a research paper the world's largest asset manager published last week.
Its forecast goes further than any major institution has publicly gone on compute.
In a research paper, BlackRock argued that processing capacity is "becoming a distinct, large, and increasingly investable economic resource that could support a new class of digital assets," and that the contracts to trade it are a matter of design work rather than doubt.
A Commodity Market In Everything But Name
The paper draws the comparison to commodities directly. Large resource markets have historically built trading infrastructure as they scaled, the authors write, and "compute may follow a similar path as AI adoption scales."
They expect "standardized products, including exchange-traded compute futures, to support more transparent price discovery and more effective hedging for both providers and consumers of compute capacity." Those standardized contracts, the paper says, could create claims on capacity that can be "represented, transferred, pledged as collateral, and settled through programmable infrastructure."
BlackRock acknowledges the obstacles. Chip generations differ in productivity, energy costs vary sharply by region, and no standard exists for how contracted capacity gets delivered or settled in cash. It calls these "important but ultimately resolvable design considerations."
Also Read: Bitget Reopens Withdrawals After $388M Hack, Starting With Bitcoin
Stripe's $7 Billion Signal
The report's clearest piece of present-tense evidence is Stripe's agreement to acquire OpenRouter, which routes AI workloads across more than 400 models from over 80 providers.
Financial outlets had reported the price at more than $7 billion in August. Stripe has not disclosed terms.
BlackRock calls the deal "an early strategic signal that model routing and compute-usage optimization are becoming part of the financial infrastructure surrounding AI." The buyer is the tell. A payments company acquiring a compute router sits at exactly the intersection the paper describes.
Inference Is What Triggers It
The economics change because of what AI systems will spend most of their power doing.
Citing McKinsey estimates, the paper says inference rather than training becomes the largest workload by 2030, taking 43% of global data center power demand against 28% for training.
Training buyers are few and concentrated. Inference buyers are many and fragmented, which is the condition that produces a market rather than a procurement contract.
Consensus estimates cited in the paper put combined revenue across the major hyperscaler cloud segments near $1.1 trillion by 2030, a 29% compound annual growth rate from 2025 levels.
Read Next: Solana ETFs Add $188.1M In Five Days And Push Inflows Past $1.5B
