BlackRock Report: AI Agents Favor Bitcoin (BTC) for Long-Term Savings
BlackRock's report says AI agents picked stablecoins for payments and Bitcoin (BTC) for long-term savings, drawn from simulations rather than live transactions.
AI SummaryAI
- Robert Mitchnick, BlackRock's head of digital assets, signed the report
- Bitcoin Policy Institute research found AI agents prefer stablecoins for payments and Bitcoin for savings
- BlackRock says the findings came from simulations, not observed real transactions
- BlackRock expects growing stablecoin use to raise blockchain transaction volume and crypto demand
The Machine-Native Economy Report
BlackRock's digital assets division has published a report arguing that artificial intelligence could redraw the plumbing of money, and that the asset best suited to long-term machine savings is
Bitcoin (BTC). The paper, titled “The Machine-Native Economy: AI and Digital Assets,” carries the signature of Robert Mitchnick, head of BlackRock Digital Assets, and was released on October 7. It examines research from the Bitcoin Policy Institute (BPI), a nonprofit known for its Bitcoin-focused policy work. The BPI study posed a practical question: which assets would AI agents choose if left to manage payments and long-term storage on their own? Researchers then observed how the agents behaved, and the recorded preference split cleanly along the time horizon. For payment transactions, the agents generally selected stablecoins, whose predictable price structure and stable unit of account suit machines that transact at high frequency and in small increments. For long-term savings, the agents turned to Bitcoin. BlackRock reads that split as early evidence for a possible machine-native monetary architecture in which stablecoins act as the payment layer while Bitcoin serves as the store of value, a division of labor that echoes arguments long advanced under Bitcoin maximalism. The report lands at a moment when institutional research papers, rather than flows alone, increasingly shape the broader Bitcoin conversation, and it stops well short of a valuation call: the document makes no prediction about the Bitcoin (BTC) price.
The distinction between simulation and live behavior sits at the center of the report's caution. The BPI exercise asked agents directly which assets they would prefer for payments and long-term safekeeping, and BlackRock stresses that the answers reflect modeled choices rather than spending by deployed software. For that reason the firm rejects the shortcut reading that AI agents will pile into
Bitcoin (BTC), filing the results instead as early data on how an AI-driven monetary structure could take shape. The stablecoin side of the argument rests on a simple property: machines executing many small transactions need a fixed measuring stick, and Bitcoin's volatility rules it out for settlement even as it strengthens the case for Bitcoin as the savings layer. Whether the instrument in circulation is a plain coin or a yield-bearing stablecoin, the settlement role demands price predictability above all. BlackRock then extends the argument outward. AI systems, the report writes, are evolving from tools that produce information into autonomous systems that transact directly, settle payments and operate the required payment infrastructure without human intervention, a shift the firm expects to force significant changes in financial infrastructure. Rising stablecoin usage would lift transaction volume on blockchain networks, and that volume, in BlackRock's reading, can generate demand for cryptocurrencies more broadly. For
Bitcoin (BTC) holders, the long-term preference recorded in the simulations reads as a machine-age version of HODL behavior: agents holding BTC not for trading but for storage across long horizons. Institutions already treat the asset that way, as when the US government moved 833.6 BTC into Coinbase Prime custody.
Early Data, Not Accumulation
The caveat deserves as much attention as the headline finding. A simulation of agent preferences is not a flow, and nothing in the report measures actual machine purchases of BTC, so the note should not be read as an accumulation signal. What it does establish is that the world's largest asset manager now models a future in which stablecoins carry machine payments while Bitcoin stores the value those machines accumulate, and in which settlement volume, not sentiment, becomes the metric to watch. If that architecture emerges, instruments such as Wrapped Bitcoin would be the bridge letting agents hold BTC while transacting on programmable rails. The thesis also lands in a year when even the BIS cast Bitcoin's first decades as finance's open laboratory. BlackRock names no invalidation level for the scenario it describes; the nearer-term market read belongs to Bitcoin technical analysis, not to this report.
AI-generated, AI-reviewed, under COINOTAG editorial oversight.

