Bernstein argued that stablecoins could evolve into the internet’s native payment layer as AI agents begin transacting on their own. The core thesis is that programmable digital dollars may become the default settlement tool for machine-to-machine commerce.
That view also fed directly into Bernstein’s equity outlook. The firm tied the theme to a more constructive stance on Circle, lifting its target to 190 and implying roughly 60–70% upside if stablecoin adoption broadens meaningfully.
Stablecoins Are Being Reframed as Payment Infrastructure
Bernstein’s argument rests on the idea that stablecoins are naturally suited to automated, low-value, high-frequency transactions. For autonomous software agents, near-instant settlement and low-friction transfers make stablecoins a more practical payment rail than many traditional systems. In that framework, the firm sees Circle and Coinbase as the clearest public-market proxies because of their direct roles in stablecoin infrastructure and distribution.
At the same time, Bernstein did not present agentic commerce as the only growth engine. The firm said the current foundation for stablecoin expansion still comes from more conventional payment uses such as cross-border business settlement, remittances, card-linked products and neobanking. Those use cases, in its view, provide real demand while the AI-agent thesis develops over a longer cycle.
That existing base already appears to be growing. Bernstein estimated total stablecoin payment volume rose to $375 billion in 2025 from $213 billion in 2024, reinforcing the case that utility-led adoption is expanding even before autonomous payments become mainstream.
Commercial Adoption Still Lags the Narrative
Even with that bullish framing, the gap between infrastructure progress and actual usage remains significant. Independent analytics and Bernstein’s own discussion suggest that some early agentic-payment activity has looked far less robust after filtering out distorted or low-quality volume. The example is x402, where roughly $24 million in reported 30-day activity dropped to about $1.6 million after wash-trading adjustments.
The same pattern showed up in newer integrations. Early real-world traction still appears modest, with one Stripe and Tempo integration posting only about $5,000 in its first week. That kind of number suggests the technology may be moving faster than actual merchant demand.
Analysts were blunt about the current limitations. Chris Donat of BWG Global said consumers are not yet asking to pay with stablecoins, underscoring how early the market still is from a mainstream adoption standpoint. That weak pull from end users remains one of the biggest constraints on the thesis.
The infrastructure itself is also unfinished. Protocols and rails are still in an early stage, and many of the features needed for micropayments at scale are still being tested rather than proven in production. That leaves execution risk high even if the long-term opportunity is real.
A further obstacle is trust and protection. A Mastercard spokesperson warned that fraud prevention, dispute resolution and credit features become critical once AI systems start paying on behalf of users. Without those safeguards, payment automation may remain technically impressive but commercially limited.
The Theme Is Compelling, but the Proof Is Still Ahead
Bernstein’s note ultimately reframes stablecoins less as speculative crypto instruments and more as financial plumbing. If agentic payments do scale, issuers and platforms tied to stablecoin settlement could capture structural revenue from transaction flow, liquidity and distribution. That is the logic behind its positive view on Circle and Coinbase.
For now, though, the market still has to prove that vision in real commercial settings. The immediate challenge is not building the narrative, but demonstrating genuine demand, sustainable transaction volume and the protections needed for automated payments to work at scale.
