
What happened
BlackRock's research paper "The Machine-Native Economy," co-written by digital assets head Robert Mitchnick, argues AI agents need machine-native money and says stablecoins moved over $11 trillion in 2025.
Why it matters
If AI agents transact without human sign-up, today's card and bank systems may not fit, so crypto rails could become the payment layer those agents use.
What to watch
The paper itself admits AI-agent payments are still limited, and its evidence includes simulations, not real purchases. BlackRock's own iShares Bitcoin Trust held $67 billion on September 25 and is down about a third this year.
WHO IT HITSInvestors and digital-asset product teams weighing crypto exposure, plus payments and treasury staff at firms building AI-agent tools, face a new argument that machine-to-machine payments may need crypto rails.
Summaries like this, in your inbox every morning.
BlackRock's argument starts from a simple mismatch. AI agents are programs that finish tasks with little human help, but cards and bank transfers need a human to sign up, charge fees that make penny payments pointless, and some take a day or more to settle. Stablecoins, which the paper says never close and moved more than $11 trillion in 2025, are offered as the alternative, and BlackRock notes the plumbing is already being built: Coinbase's x402 lets software pay a website instantly, and Cardano joined x402 earlier this month.
The more ambitious part concerns compute. Analyst estimates cited by BlackRock put cloud revenue at Amazon, Microsoft, and Google near $1.1 trillion by 2030, and the firm suggests that power could be sold in standard contracts, like oil, and settled on blockchains. Stripe is already moving in that direction, agreeing in August to buy OpenRouter, which routes AI requests across more than 400 models; CEO Patrick Collison said tokens are the central currency for companies building with AI.
What keeps this from being a clean bet is BlackRock's own fine print. The paper admits payments by AI agents are still limited, and it cites tests where models chose stablecoins to spend and Bitcoin to save, though those were simulations, not real purchases. BlackRock also has its own money on the line: IBIT held $67 billion on September 25 and has lost about a third of its value this year. The thesis may hinge on whether real agent transactions arrive before skeptics read the simulations as the whole story.
Pick your industry and the AI tools you use, and get news related to your work every day.
Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →
Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.
ServiceNow trades on roughly 80.4x earnings, above the Software industry average near 29.7x and a peer group c…

Palo Alto Networks launched Cortex XCOR, an AI-native observability platform

John Deere announced JD, an AI assistant embedded in its Operations Center, at a press event at Iowa State Uni…

At a Sept. 24 Santiago summit organized by AmCham Chile and Honeywell, Honeywell's José Simón proposed expandi…

Salesforce trades at a 14.4x forward P/E versus Palantir's 85.5x, and Agentforce annual recurring revenue pass…

CrowdStrike Holdings has listed its Falcon cybersecurity platform in the OpenAI Marketplace for enterprise buy…
