On October 1, the Ethereum mainnet went live with something called zkAPI: users first deposit ETH or USDC into an on-chain vault, and then use a zero-knowledge proof to obtain a temporary key to call an AI model—payment and identity are separated.
This gives USDC an additional “machine payment” use case, but it doesn’t mean buy-side demand.
The real on-the-move capital signals come from two groups. Some say that whale addresses with a single transaction over $1 million saw 30-day cumulative stablecoin inflows rise from $21.7 billion to $30.5 billion—an increase of over 40%, after a prior peak that had exceeded $61.0 billion. The other is that the USDC Treasury minted 250 million USDC on Solana.
On one side, it’s payment rails expanding; on the other, exchange funds flowing back. The market bundles these two developments into a single story of “a surge in stablecoin demand.”
But zkAPI’s network layer may still be able to associate users via IP addresses and prompts. If the vault only receives and never exits, then it’s still just a test run.
There’s also a piece of news to be verified: Visa and the UK’s Lloyds Bank reportedly completed a cross-border settlement experiment based on USDC. If that’s true, then this would be the actual incremental development in the settlement layer.
So the question is: in this wave of USDC hype, should it be priced according to the “AI payment narrative,” or according to “exchange inflows”?
This gives USDC an additional “machine payment” use case, but it doesn’t mean buy-side demand.
The real on-the-move capital signals come from two groups. Some say that whale addresses with a single transaction over $1 million saw 30-day cumulative stablecoin inflows rise from $21.7 billion to $30.5 billion—an increase of over 40%, after a prior peak that had exceeded $61.0 billion. The other is that the USDC Treasury minted 250 million USDC on Solana.
On one side, it’s payment rails expanding; on the other, exchange funds flowing back. The market bundles these two developments into a single story of “a surge in stablecoin demand.”
But zkAPI’s network layer may still be able to associate users via IP addresses and prompts. If the vault only receives and never exits, then it’s still just a test run.
There’s also a piece of news to be verified: Visa and the UK’s Lloyds Bank reportedly completed a cross-border settlement experiment based on USDC. If that’s true, then this would be the actual incremental development in the settlement layer.
So the question is: in this wave of USDC hype, should it be priced according to the “AI payment narrative,” or according to “exchange inflows”?