I just finished extending Binance’s USD1 campaign until August 7 yesterday. Today, I took a quick look at World Liberty Financial @worldlibertyfi’s recent products as well.
Originally I just wanted to see whether, after $USD1 , there are any new use cases—beyond continuing to use the $WLFI rewards to attract users to hold.
Turns out my attention got pulled in by the AgentPay SDK.
//////////////////////////
In simple terms, AgentPay gives AI agents a set of $USD1 wallets they can use themselves.
Agents can hold funds, complete payments, and even transfer money between different chains.
Developers can also set per-transaction limits, daily limits, weekly limits, as well as amount thresholds that require manual confirmation.
For example, routine API calls, buying data, or paying small service fees can all be handled automatically by the agent.
//////////////////////////
I think this design fits well with where AI agents are heading.
Many agents can already write code, do research, and carry out tasks.
But the moment they start actually running business, they quickly run into payment problems.
Calling models costs money. Buying data costs money. Using cloud services and on-chain tools also costs money.
If every time they need to wait for someone to manually confirm, the agent’s automation will be bottlenecked.
If you just hand over the wallet entirely, the risk of funds becomes very hard to control.
What AgentPay is trying to add is the missing layer of spending permissions in the middle.
Let the agent spend on its own, but only within the rules already written in advance.
//////////////////////////
So looking back at yesterday’s USD1 campaign extension, I think it’s actually interesting to put these two things together.
With Binance’s campaign, it continues to give regular users a reason to hold USD1.
As for AgentPay, it starts trying to make AI agents also become users of USD1.
Earlier, USD1 had already entered scenarios like wealth management, lending, contract collateral, and cross-chain liquidity.
Now it’s starting to extend further into the machine-payment direction.
AgentPay is still more of a developer tool for now. Whether it can truly take off depends on how many applications get integrated later, and how much ongoing payment demand they can generate.
But if, in the future, a large number of agents will end up calling services, buying data, and managing funds on people’s behalf, then the next wave of high-frequency stablecoin users probably won’t need to sit in front of the screen.
Originally I just wanted to see whether, after $USD1 , there are any new use cases—beyond continuing to use the $WLFI rewards to attract users to hold.
Turns out my attention got pulled in by the AgentPay SDK.
//////////////////////////
In simple terms, AgentPay gives AI agents a set of $USD1 wallets they can use themselves.
Agents can hold funds, complete payments, and even transfer money between different chains.
Developers can also set per-transaction limits, daily limits, weekly limits, as well as amount thresholds that require manual confirmation.
For example, routine API calls, buying data, or paying small service fees can all be handled automatically by the agent.
//////////////////////////
I think this design fits well with where AI agents are heading.
Many agents can already write code, do research, and carry out tasks.
But the moment they start actually running business, they quickly run into payment problems.
Calling models costs money. Buying data costs money. Using cloud services and on-chain tools also costs money.
If every time they need to wait for someone to manually confirm, the agent’s automation will be bottlenecked.
If you just hand over the wallet entirely, the risk of funds becomes very hard to control.
What AgentPay is trying to add is the missing layer of spending permissions in the middle.
Let the agent spend on its own, but only within the rules already written in advance.
//////////////////////////
So looking back at yesterday’s USD1 campaign extension, I think it’s actually interesting to put these two things together.
With Binance’s campaign, it continues to give regular users a reason to hold USD1.
As for AgentPay, it starts trying to make AI agents also become users of USD1.
Earlier, USD1 had already entered scenarios like wealth management, lending, contract collateral, and cross-chain liquidity.
Now it’s starting to extend further into the machine-payment direction.
AgentPay is still more of a developer tool for now. Whether it can truly take off depends on how many applications get integrated later, and how much ongoing payment demand they can generate.
But if, in the future, a large number of agents will end up calling services, buying data, and managing funds on people’s behalf, then the next wave of high-frequency stablecoin users probably won’t need to sit in front of the screen.