The "plateau" after reaching a new high: After $300 billion, what are stablecoins waiting for?

null

Stablecoins have been sitting at the $300 billion mark—the highest point in history—for a while now.

The numbers are impressive, but if you zoom in, you’ll see another side: over the past six months, stablecoin growth has started to faintly show the outline of a plateau. This doesn’t mean the market has lost its room for imagination; it means the logic that supported the scale expansion of the past few years is quietly approaching its own limits.

This means stablecoins need a new story—not just a new use case, but a deeper shift in their underlying attributes. If payment scenarios are no longer just about trading, and if the initiator is no longer only human, what role will stablecoins play?

I. A Stuck Ceiling: What Changes and What Doesn’t

Stablecoins aren’t the first asset class to reach a crossroads like this.

From USDT to USDC, and then to the various newer types of stablecoins, previous rounds of expansion were almost always accompanied by several familiar scenarios: higher overall market trading volumes, stronger DeFi activity, greater cross-chain liquidity, and broader global remittance demand.

On the surface, stablecoin growth comes from supply-side expansion. Stablecoins have reached crossroads like this before.

But in the past few years, nearly all of these core demands have come from human behavior. Whether it’s order matching and trading in exchanges, collateralized lending and borrowing in DeFi protocols, cross-border transfers and arbitrage routes, or short-term parking of risk-hedging capital—at bottom, everything revolves around “trading.” In other words, the growth of stablecoins in the previous phase was fundamentally driven by “people’s trading demand.”

The problem today, however, is that these demands haven’t disappeared—they’re increasingly nearing a “predictable ceiling.” After all, exchange volumes are still huge, but the competitive landscape is relatively stable; DeFi remains important, yet it’s difficult to create explosive incremental growth on its own the way it did early on. Cross-border payments and arbitrage are still expanding, but increasingly resemble a slow penetration process rather than a new story that can instantly reshape valuation expectations.

That’s exactly why market interest in “the next stablecoin system with incremental growth stories” is clearly rising.

At present, the new incremental growth mainly clusters in two directions.

One is on-chain yield-bearing stablecoins—combining stablecoins with Treasuries, RWA, protocol revenues, and other structures, repackaging their appeal through “holding earns yield,” similar to the yield-bearing stablecoin path that has been discussed repeatedly in the market in recent years;

The other is the direction that has clearly been hotter recently: the on-chain business of AI Agents, and the stablecoin payment and settlement demand built around it.

Actually, compared with the yield path, on-chain payments and the stablecoin rails fit the characteristics of this new demand even better, because stablecoins naturally have a set of conditions that traditional payment systems struggle to combine: operate 24/7, global standardized settlement, programmable execution, support for high-frequency micro-payments, and no need for complex intermediaries layered with authorization.

In other words, stablecoins may not be fighting only over today’s existing stock of cross-border payment volume—they may be competing for a larger incremental payment market in the future, especially when the initiator of payments is no longer only human.

II. From Yield-Bearing to AI-Driven: Exploring New Paths for Incremental Volume

Recently, traditional giants have been clearly doubling down on this new direction.

For instance, Visa Crypto Labs launched its first experimental product, Visa CLI, attempting to let AI agents safely complete the required fee payments when writing code and calling services. Put in a broader context, its significance isn’t simply that it adds another tool—it’s that, for the first time, the payment主体 (the subject of payment) is starting to shift from “human” to “program.”

In traditional payment systems, every transaction implicitly assumes a prerequisite: it must be initiated by a human. Whether it’s cards, e-wallets, or mobile payments, everything behind them relies on KYC, manual authorization steps, and finally fund transfers carried out by the banking account system.

In the end, this system is designed around “human behavior.”

But AI doesn’t belong to that system.

If an AI Agent needs to complete a task, it may need to automatically subscribe to data services, pay API fees based on call counts, purchase computing-power resources across different platforms, or even execute automated trading based on policies. For this kind of behavior, every step waiting for a human to manually confirm is unrealistic and also can’t adapt to the high-frequency, real-time operational rhythm. And traditional banking account systems were not built for native interactions between machines.

That’s precisely where on-chain payments have the advantage. Stablecoins like USDT and USDC, in a sense, are money that’s “born for AI”: they are borderless, programmable, and support instant settlement. They perfectly match AI’s ultimate pursuit of “high speed, low cost, and no friction,” which also means that combining stablecoins with wallets gives these payments genuine programmability for the first time.

What this produces is a new form—“Agent Wallet.” Wallets gradually evolve into AI’s asset interface and execution endpoint, and in practice several typical patterns emerge (Further Reading:《From “Collective Intelligence” to “Super Individuals”: How AI Is Reshaping DAO and the Ethereum Ecosystem?》):

Non-custodial authorization: You can create an independent, restricted sub-wallet for your AI Agent. It can trade autonomously within limits you set (for example, no more than 500 USDC per transaction) without you manually confirming each time. The master private key remains in your hands at all times, and the AI is only your authorization proxy;

Cross-chain asset management: AI can instantly query your assets across more than 100 chains, and rebalance, stake, or arbitrage according to strategies you set. You’re freed from tedious day-to-day monitoring and can focus on higher-level strategic decisions;

Human–machine collaboration: This is not “hands-off” completely. It supports flexible confirmation mechanisms—for example, small amounts auto-executed and large amounts prompting reminders. AI is responsible for finding opportunities and constructing transactions, and you provide the final button. This model can also perfectly combine human judgment with AI execution efficiency;

III. From “Who Issues Stablecoins” to “Who Organizes Them into a Network”

If Visa’s experiment represents a change on the demand side, then on the other side, the blockchain project Tempo—supported by Stripe and Paradigm—announcing the launch of a stablecoin mainnet is more like an upgrade on the supply side.

Its importance isn’t only that the market now has another stablecoin project. It’s also a reminder that the industry’s competitive focus has long since shifted from “who can issue stablecoins” to “who can truly organize stablecoins into a network that can run.”

In the past few years, the stablecoin industry first solved the issuance problem.

Mainstream stablecoins like USDT and USDC have completed the scaled on-chain supply of “on-chain dollars,” making “digital dollars” for the first time an asset class that can be used globally. But after supply becomes increasingly mature, what’s truly scarce is no longer stablecoins themselves, but the capability to connect on-chain accounts, merchant collections, enterprise payments, and fiat-clearing networks.

This also explains why, from Stripe to Mastercard, and then to Visa and PayPal, traditional payment giants have rolled out dense deployments around stablecoins over the past two years. Even native crypto platforms have started to seep into TradFi in return:

In October 2024, Stripe acquired stablecoin API service provider Bridge for $1.1 billion, setting a new record for M&A deal amounts in the crypto payments space at the time;

This March, Mastercard acquired stablecoin service provider BVNK for $1.8 billion, refreshing that record again;

At the same time, Visa continues to expand its partnership with Bridge, pushing stablecoin-linked cards to a broader market;

Looking further back, PayPal rolled out PYUSD earlier, and it has already clearly signaled its strategy;

In Hong Kong, the licensed compliant exchange OSL announced last year that it would pivot toward stablecoin payment and settlement infrastructure. This year, in January it completed the acquisition of Web3 payments provider Banxa, and in February it launched an enterprise-level dollar stablecoin, USDGO, which complies with U.S. federal regulation and can be distributed compliantly in Hong Kong;

Overall, Crypto and the payments-at-large industry’s attitude toward stablecoins has long shifted from “watching” to “positioning.”

That’s why projects like Bridge, BVNK, OSL/USDGO—and today, something like Tempo, which tries to build stablecoin network layers—suddenly seem so scarce. Their most valuable advantage is precisely their position: one end connects on-chain assets and wallets, and the other end connects merchants, enterprises, payment service providers, and the real-world clearing networks.

The industry has already moved past the initial phase of “who issues stablecoins” and entered the second half of “who can make stablecoins truly run.”

Written at the end

Stablecoin reaching a new high isn’t just a refresh of a size figure—it’s also like a watershed moment.

If in the past few years stablecoins addressed “how people make payments on-chain,” then next it must face the question: how to network, scale, and automate stablecoin influence.

When AI can autonomously call wallets, when payments are embedded into programs to run, and when stablecoins become the default settlement currency between global trade, stablecoin ceilings will no longer depend only on today’s market transaction volume, nor only on the speed at which it replaces existing cross-border payment volume. What it maps to might be a bigger new variable.

And that’s exactly why the next round of stablecoins that’s truly worth watching isn’t only whether the supply can keep setting new all-time highs—it’s whether it can further evolve into a “global settlement interface.”

And perhaps that is the real driver behind stablecoins breaking through this new plateau at the high end.