Circle's stock price doubles, paradigm shift of stablecoins

Author: Chloe, ChainCatcher

At the beginning of 2026, the payment industry is witnessing an asymmetric bet. From Circle launching the Arc blockchain and Nanopayments, to Stripe partnering with Paradigm to incubate the payment public chain Tempo and investing $1.1 billion to acquire Bridge, the giants share a common goal: they are no longer designing payment tools for humans, but rather building a new financial track for AI agents.

The logic of this arms race revolves around the idea that when AI agents become the main players in economic activities, the traditional credit card fees of 2-3% will gradually lose their necessity, replaced by stablecoin payment protocols with extremely low rates. In February of this year, a lengthy scenario report from Citrini Research predicted this future, leading to a one-day market value evaporation of 5-7% for Visa, Mastercard, and American Express.

Is this an overreaction from the market, or is capital pricing in future value ahead of time? It is still difficult to conclude. However, there remains a significant gap between ideals and reality: while the monthly transaction volume of the x402 protocol is still at $24 million, the global e-commerce market has already ballooned to $6.88 trillion. It raises the question: is this a layout for a new era, or a breathtaking gamble that overdraws the future?

The arms race in AI payment tracks has begun, but the battlefield has yet to take shape

In early March 2026, Circle CEO Jeremy Allaire revealed a detail: the company completed an internal settlement of $68 million among its eight business entities in just 30 minutes using its own USDC stablecoin, without the assistance of traditional bank wire transfers; by traditional financial standards, a cross-border transfer of the same scale typically takes 1 to 3 business days and incurs expensive bank fees.

This practical result sent a clear signal to the market: the potential of stablecoins as "modern financial infrastructure" cannot be ignored, proving that companies can bypass traditional banking networks to achieve large-scale cross-border or internal fund transfers at low cost and high efficiency.

This is not an isolated case; it is the infrastructure that the crypto industry has accumulated over the past years that is about to begin to exert its power.

Circle's dual-line bet, focusing on micropayments and agency commerce

Circle's layout showcases a clear dual-line strategy: first is Arc, a Layer1 blockchain that launched its public testnet in October 2025, using USDC as its native gas token. This blockchain features sub-second finality and predictable rates priced in dollars, precisely addressing the pain point of cost fluctuations in traditional EVM chains due to on-chain congestion in payment scenarios.

Secondly, there is Nanopayments, which went live in early March this year, supporting USDC transfers as low as $0.000001 with zero gas fees. Through batch on-chain settlements, Circle merges thousands of micropayments into a single on-chain operation, compressing the cost per transaction to nearly zero. The collaboration between Circle and OpenMind is also impressive, with a robotic dog named Bits autonomously paying for electricity, symbolizing the concept validation of "agency commerce" and showcasing the possibility of robots conducting transactions as independent entities in the physical world.

Stripe's strategic layout, focusing on ecosystem scale and integration with traditional finance

Stripe's layout also demonstrates ambition to the market. In September 2025, Stripe partnered with crypto venture capital firm Paradigm to launch the payment-specific Layer1 blockchain Tempo, aiming for 100,000 transactions per second and sub-second finality. The architecture of Tempo shares a similar design philosophy with Circle's Arc, both supporting gas fees paid in stablecoins and incorporating automated market makers (AMM) for cross-currency settlements.

However, Tempo's ecosystem partner list also includes top institutions and traditional financial giants such as Visa, Mastercard, UBS, OpenAI, and Shopify. On the other hand, considering Stripe's previous acquisition of Bridge ($1.1 billion), the Privy transaction, and the investment in Tempo, Stripe's total investment in the stablecoin infrastructure track is estimated to have likely exceeded $1.5 billion.

How Citrini Research ignited the Black Monday for traditional payment giants

In late February 2026, an independent organization named Citrini Research published a lengthy scenario report on Substack, titled “The 2028 Global Intelligence Crisis”.

The core narrative of the report is: when AI agents begin to make shopping and payment decisions on behalf of consumers at scale, "eliminating transaction friction and costs" will naturally become one of the goals of algorithm optimization. In the scenario designed by Citrini, by 2027, these agent systems will start bypassing traditional credit card networks, turning to settle in stablecoins on Solana or Ethereum Layer2, where the cost of a transaction is only a fraction of that in traditional finance.

After the report was released over the weekend, it quickly went viral in the tech and finance circles. By the market opening on Monday, February 23, market sentiment began to deteriorate rapidly. On that day, Visa's stock price fell by about 4%, Mastercard dropped over 6%, and American Express declined nearly 8%, with several institutions primarily focused on credit cards and consumer finance losing hundreds of billions in market value in a single trading day.

As the report was shared by several tech and finance KOLs on social platforms like X, Citrini's originally "risk exercise" macro scenario memo was misinterpreted by the market in a short time as a "forthcoming fundamental prophecy," further amplifying concerns about the relationship between AI, stablecoins, and traditional payment networks.

Why stablecoins are the "natural currency" for AI agents

To understand why stablecoin issuers like Circle and payment infrastructure providers like Stripe view betting on "agent-native payments" as the next strategic direction, one must first recognize a key point: when the subject of transactions shifts from humans to AI agents, the design premises of traditional credit card networks begin to loosen.

The high fees and settlement delays of the traditional financial system make it difficult to accommodate the massive micropayments generated by AI agents. Stablecoins, with their extremely low costs and second-level settlements, provide an efficient and smooth underlying environment for 24/7 operating AI. Through programmability, AI agents can automatically execute details under budget rules set by humans, achieving automated settlements without legal identity, becoming the indispensable digital lifeblood in agency commerce.

$24 million compared to $6.88 trillion: the gap between ambition and reality

However, all the previous technical optimism is built on one premise: demand will eventually emerge. Current known data supporting this premise remains quite limited.

As an early reference for agent payment standards, the x402 protocol has recorded on-chain data over the past 30 days, with only about 100,000 buyers and over 20,000 sellers, accumulating a transaction volume of approximately $24 million. In comparison, the global e-commerce market is expected to reach $6.88 trillion in 2026, making $24 million only about 0.00035% of that figure. For a technology expected to "reshape the payment track," it currently resembles more of a concept validation rather than widespread adoption.

Even Circle's own USDC shows a similar scenario: with a circulation of about $75.3 billion, the on-chain transaction volume in Q4 2025 surpassed $11.9 trillion, but primarily stemmed from institutional settlements, DeFi protocols, and internal circulation within exchanges. Truly consumer-driven scenarios that can be clearly marked as "agent payments" are nearly impossible to identify within this total.

Chris Donat, head of fintech at BWG Global, also pointed out: it is hard to expect consumers to actively demand payments in stablecoins, and retailers lack strong incentives to reconstruct the entire payment collection system due to this demand gap. The lukewarm demand side is not due to inadequate technology, but rather the evolution of user habits and business ecosystems, which has always been much slower than the iteration of infrastructure.

Stripe's approach perfectly reflects this tension: on one hand, it offers a merchant acquiring fee rate of about 1.5% for stablecoin payments in the U.S., which is indeed attractive compared to the nearly 2.9% rate for credit cards; on the other hand, Stripe treats stablecoins as an optional channel, patiently waiting for regulatory frameworks to mature, consumer education to be completed, and enough killer application scenarios to emerge, while currently, these three conditions are far from being in place.

The gamble of infrastructure first, how to calculate wins and losses?

To understand the current actions of Circle and Stripe, one must frame it with the logic of the infrastructure industry rather than the consumer goods industry. When AWS launched S3 and EC2 in 2006, the demand for cloud computing was virtually nonexistent; at that time, no company knew it needed elastic computing resources until the option was available.

Perhaps the pipeline comes before the water flow; this is the fundamental logic of infrastructure betting.

From this perspective, Circle and Stripe's early investments, whether in R&D spending for Arc, the high financing costs associated with Tempo, or the billion-dollar acquisitions of Bridge-type assets, are closer to "location fees" rather than short-term recoverable business investments. They are betting that once the transaction volume related to AI agents climbs from today's tens of millions of dollars to the next level, the infrastructure providers that first complete technological validation and regulatory compliance will monopolize the new era's business.

However, the weakness of this logic is that if the arrival of demand is much later than expected, or if the actual form when it materializes is vastly different from current assumptions; for example, if the agent-native payment solutions are ultimately launched by existing traditional financial organizations like Visa rather than being disrupted by them, then all the pipelines laid out in advance may be regarded as irrecoverable sunk costs in financial reports.

The true contribution of Citrini's report lies not so much in the accuracy of its predictions, but in forcing the management and investors of the payment industry to seriously confront a question: when the decision-makers of transactions shift from humans to software, every seemingly solid premise in existing business models must be revalidated.