At the World Economic Forum in Davos, Binance founder Changpeng Zhao laid out a compact thesis linking cryptocurrency, blockchain infrastructure, and autonomous artificial intelligence. He argued that crypto rails and smart-contract ledgers are the most practical way for AI agents to transact without leaning on traditional financial intermediaries.
Zhao made the point in direct terms, saying, “The native currency of AI agents will be cryptocurrency,” and adding that “blockchain will become the most natural technical interface for AI agents.” He framed the idea as part of a broader roadmap that also prioritizes asset tokenization and payments as parallel frontiers.
BINANCE FOUNDER CZ JUST SAID LIVE IN DEVOS THAT AI WILL USE CRYPTO FOR PAYMENTS NOT BANK CARDS
IT’S COMING pic.twitter.com/oY7zoTyiHK
— Vivek Sen (@Vivek4real_) January 22, 2026
Zhao’s three-part roadmap
Zhao’s first theme centered on machine-to-machine commerce, where autonomous agents use cryptocurrency for programmable payments and economic decisions while relying on blockchains for immutability and contract execution. In his framing, agents would treat crypto as a default settlement layer because it is software-native and always on.
His second theme focused on tokenization at the state level, saying he was engaged with more than a dozen governments on proposals to tokenize public assets. He positioned tokenization as a way to broaden liquidity, enable fractional ownership, and allow 24/7 transferability for traditionally illiquid assets.
The third theme was payments, where Zhao suggested the user experience can remain familiar while the backend increasingly settles via stablecoins and blockchain rails. His core message was that crypto becomes “invisible” to end users even as it absorbs more of the settlement workload.
Operational and supervisory implications
Within the same session, Zhao contrasted exchange infrastructure and established stablecoins with traditional banking by pointing to operational resilience, including a single-day $7 billion withdrawal episode he said was handled without disruption. He used that stress event as shorthand for scalability and readiness under peak load.
He also expressed skepticism about the durability of most memecoins and projected a reduced role for traditional retail banks over the next decade as digital financial services expand. Taken together, his comments implied a strategic tilt toward utility-first rails rather than hype-driven token cycles.
From an operational perspective, each pillar in Zhao’s roadmap introduces distinct control requirements. Tokenized state assets raise sovereign custody, disclosure, and valuation questions, while agent-driven payments shift counterparty risk toward code execution and automated settlement pathways. Stablecoin-based settlement, meanwhile, concentrates attention on reserve transparency, redemption mechanics, and liquidity management for treasuries and custodians.
Looking ahead, stakeholders will likely track the concrete implementation signals Zhao referenced: outcomes of government tokenization discussions, the plumbing behind stablecoin settlement, and real-world tests that demonstrate agent payments at scale. Those deliverables will be the practical KPI for whether the AI–crypto linkage translates into measurable changes in custody, reporting, and prudential practice.