BlackRock paper frames blockchains as future rails for AI agent payments
DeFi & Yields ·
A new BlackRock research paper argues that machine-native settlement infrastructure, not human trading, will define blockchain's next phase as autonomous AI agents begin transacting on their own.
BlackRock has published a whitepaper examining how artificial intelligence and digital assets are converging, arguing that networks of autonomous AI agents will need financial settlement systems built for machines rather than people. The asset manager frames this as an urgent gap: as agents begin acting independently, they require rails that can move value without a human approving each step, a shift detailed in BlackRock's whitepaper.
The paper lays out three areas where AI and crypto infrastructure intersect. The first is rapid, automated payments between machines, where AI agents settle transactions directly in fiat-backed stablecoins using emerging protocols including x402, offerings from Stripe and Tempo described as MPP, and OpenAI's ACP. The second involves AI systems interacting with tokenized financial products, letting software execute trades and manage portfolios through smart contracts tied to tokenized real-world assets. The third centers on treating computing capacity itself as a financial asset that can be collateralized or traded.
Underpinning the thesis is a figure BlackRock cites for scale: adjusted stablecoin transaction volume has already topped $11 trillion in 2025. Combined with what the firm describes as growing regulatory clarity across the United States, Europe, and the Asia-Pacific region, BlackRock contends this sets the stage for blockchains to shift from primarily speculative, human-driven trading venues into settlement backbones for a machine-run economy.
The paper's conclusions are echoed elsewhere in the same research cycle, with separate coverage noting that BlackRock identifies autonomous AI agents and machine-driven commerce as a significant source of future demand for stablecoins and blockchain-based payment rails, as outlined in reporting from CryptoPotato. Related summaries of the whitepaper likewise point to tokenized compute markets and agent-to-agent transacting as core pillars of what BlackRock calls a machine-native economy.
What remains unclear is the pace at which protocols like x402, MPP, or ACP could reach the transaction volumes needed to test this thesis, and how regulators in the jurisdictions BlackRock cites will treat AI agents as independent economic actors capable of holding or moving stablecoins. Whether tokenized compute resources develop into a liquid, collateralizable asset class, as the paper's third scenario suggests, is also untested. Four distinct sources have now covered the whitepaper's release, but none report a timeline for when BlackRock expects these dynamics to materialize at scale.