BlackRock’s white paper doesn’t call for trades—it’s blunt about it: large-scale AI adoption could be the overlooked slice of demand for digital assets.

Filename (The Machine-Native Economy), subtitle How digital assets connect intelligence, commerce, and compute. Authors Will Su / Robert Mitchnick / Jay Jacobs / William Helm, document ID CE0926-M-5936357 (September 2026). Research ≠ product launch, and it’s not about which coin to buy.

Hard numbers first on the table:

1️⃣ Stablecoin circulating market cap as of September 2026 is over about $300 billion; in 2025, adjusted stablecoin transaction volume exceeded about $11 trillion—roughly in the same ballpark as Visa/Mastercard annual payment volumes; 2020–2025 CAGR is about 80% (vs. ACH in the same period of about $9.3 trillion)

2️⃣ The revenues of ultra-large cloud providers: analysts expect the market to exceed about $1 trillion per year by 2030; in the article, AWS + Azure + Google Cloud total about $1.1 trillion, roughly a 29% CAGR versus 2025

3️⃣ Estimated cumulative AI capital expenditures from 2025 to 2030 exceed about $5 trillion—compute itself is written as an economic resource that can be financed and settled

Argument, in three parts, is very restrained:

• Agentic AI needs a machine-native payment rail: high-frequency, sub-cent units, 7×24 M2M; stablecoins are more likely to handle the transaction layer, while native crypto assets and tokenized RWA can also move into programmable settlement

• Calling out protocols such as x402 (Coinbase HTTP 402) as an early machine-payment layer—protocol name ≠ product marketing

• Compute claims can be standardized into digital assets that are transferable, pledgeable, and tradable, so institutions treat compute as one kind of exposure—not just buying GPU stocks

My filter: this is an asset-management giant embedding the idea that “AI needs programmable money” into its institutional brief—not another interpretation that tries to hype up the story in a $BTC rush. And it’s separate from Grayscale renaming the Miners ETF to GCPU (an AI Compute stock basket)—one is a research outline, the other is a product rebrand.

Source: the original BlackRock whitepaper PDF text + Cointelegraph’s September 23rd recap.

#BlackRock #AI #稳定币 #RWA #DigitalAssets

Not investment advice.