According to Woofun AI, BlackRock (BLK.US) has established the shared origins of artificial intelligence and cryptocurrency in its newly released "Machine-Native Economy" report. The document, co-authored by digital assets head Robert Mitchnick, points out that AI is machine-native intelligence, while cryptocurrency serves as machine-native money, both designed to serve AI agents operating without human intervention.
The traditional payment system reveals structural shortcomings when handling machine interactions. Checks and bank transfers not only depend on manual operation, but their high fees and settlement cycles lasting up to a full day make small-value payments difficult to sustain. In contrast, USD-pegged stablecoins offer a seamless solution. Data compiled by Woofun AI shows that stablecoin circulation exceeded $11 trillion in 2025, a scale comparable to Visa (V.US) and Mastercard (MA.US).
To support this ecosystem, Coinbase (COIN.US) launched the x402 protocol, which enables instant software payments, and Cardano also integrated with the platform earlier this month, marking substantive progress in infrastructure development. A deeper transformation lies in computing power trading. As the core of AI operations, computing power is expected to achieve standardized contract settlement through blockchain. Analysts predict that by 2030, cloud service revenue for Amazon (AMZN.US), Microsoft (MSFT.US), and Google (GOOGL.US) will approach $1.1 trillion. Payment giant Stripe is actively positioning itself, having acquired OpenRouter in August, which can route to over 400 models. CEO Patrick Collison stated plainly that tokens are the core currency for AI development companies.
Meanwhile, BlackRock's own holdings also reflect market volatility. As of September 25, iShares data shows that IBIT holds $67 billion worth of bitcoin, although the fund's net value has fallen by about one-third this year. Despite the broad prospects, real-world applications still face limitations. Currently, the payment capabilities of AI agents remain incomplete. A BeInCrypto analysis last week noted that although AI models in testing tended to spend with stablecoins and store value in bitcoin, this was based only on simulated scenarios rather than actual purchase behavior, and there is still a distance to large-scale implementation.