Will AI Agents Disrupt Payment Networks? Citrini Research Warning Triggers Visa, Mastercard, and AmEx Stock Sell-Off

Citrini Research Warning Sends Shockwaves Through Payment Sector

Visa, Mastercard, AmEx could be gutted by AI 'agentic commerce' threat, Citrini Research warns — but retail traders unfazed by selloff

From stablecoins to machine-to-machine payments, a new report argues Visa, Mastercard and AmEx face a structural shakeup, not just another tech cycle.

On Sunday, February 22, 2026, independent research firm Citrini Research published a report titled "The 2028 Global Intelligence Crisis" on Substack. The following Monday, markets reacted sharply: Visa fell approximately 4.5%, Mastercard dropped around 5.7%, and American Express declined roughly 7.2%.

The report was authored by Citrini Research and Alap Shah, CEO of AI payments startup Littlebird. Written as a scenario set in June 2028, the analysis depicts a future where AI agents bypass existing payment network fee structures through "agentic commerce." According to Gizmodo's coverage, the authors explicitly stated it was "a scenario, not a prediction," yet markets responded immediately.

Background and Industry Context

The reason this report had such a significant market impact is that AI-driven structural industry changes are already becoming a reality.

In early February 2026, a sharp decline in software stocks escalated into what has been called the "SaaS-pocalypse." IBM fell 13% following Anthropic's announcement of new AI tools, and selling pressure spread across SaaS companies broadly. Against this backdrop of heightened "AI threat" market sentiment, the Citrini report raised the same concerns for the payments sector.

Moreover, development of payment protocols by AI agents is already well underway. In September 2025, Coinbase and Cloudflare launched the x402 protocol, enabling AI agents to execute stablecoin payments instantly via HTTP. Google announced the "Agent Payments Protocol (AP2)", with over 60 companies including Mastercard and American Express participating in standardization efforts. OpenAI has also built an Agentic Commerce Protocol in partnership with Stripe. The concept of "AI agents making payments" is no longer theoretical.

Structural Threat to Card Fee Models

The core issue highlighted by the Citrini report is the question of fees in machine-to-machine (M2M) transactions.

In current card payments, an interchange fee of 2-3% is charged per transaction. These fees form the revenue foundation for Visa, Mastercard, and AmEx. However, in a world where AI agents transact with each other, this cost becomes "an obvious target for elimination," the report argues.

Under the report's scenario, by 2027, AI agents begin seeking payment methods that are faster and cheaper than cards, with many migrating to stablecoins (cryptocurrencies pegged to fiat currencies like the US dollar) on Solana or Ethereum L2 (Layer 2: technology that extends Ethereum's processing capacity). With stablecoin payments, transaction costs are less than one cent and settlement is nearly instantaneous.

American Express is identified as particularly vulnerable. The contraction of white-collar employment driven by AI adoption directly impacts AmEx's core customer base, and agents bypassing interchange fees would deliver a double blow to its revenue model. Blockonomi's analysis notes that while Visa and Mastercard have room to reposition within stablecoin infrastructure, AmEx faces a more difficult transition away from its card-centric business model.

At the same time, more measured perspectives exist. Seeking Alpha's analysis points out that the report is a "scenario," not a "prediction," and that Visa and Mastercard are already investing in stablecoin infrastructure. Indeed, Visa has announced its "Trusted Agent Protocol," and both Mastercard and AmEx are participating in Google's AP2. Payment giants are recognizing the threat while exploring new roles for the agentic commerce era.

Impact and Implications for E-commerce Businesses

The scenario described in this report carries important implications for e-commerce businesses as well.

Businesses need to prepare for changes in payment cost structures. When AI agents automatically select the cheapest payment method, the 2-3% card payment fee could become a competitive disadvantage. Beginning to explore stablecoin payment adoption now can lead to medium-to-long-term cost optimization. With the passage of the GENIUS Act in 2025, the regulatory framework for stablecoins in the US is becoming clearer, lowering the barrier to adoption.

Agentic commerce readiness cannot wait. Google's AP2, OpenAI's Agentic Commerce Protocol, Coinbase's x402, and other protocols enabling AI agents to search, compare, and purchase products are expected to be standardized during 2026. According to a Payments Dive report, Amazon, Visa, Google, PayPal, and Stripe have all presented agentic commerce protocols or plans. Making your product data compatible with these protocols is a prerequisite for capturing sales through AI agents.

Reward program restructuring should be on the radar. The Citrini report suggests that reward programs funded by card fees could shrink as agentic commerce expands. Businesses relying on points-based loyalty models for customer acquisition should consider alternative loyalty strategies.

Conclusion

The Citrini Research report was written as a "scenario," not a definitive prediction. However, the structural threat of AI agents bypassing card payment fees is steadily becoming more realistic as protocols like x402 and AP2 advance in implementation.

Major payment companies are also working to adapt to this shift, with Visa, Mastercard, and AmEx all participating in agentic commerce protocol development. Key milestones to watch going forward include the Mastercard earnings in Q1 2027, which Citrini identified as a "point of no return," and the trajectory of actual stablecoin transaction volumes. For e-commerce businesses, it is important to treat this change not as something that will happen "someday" but as something to start preparing for now.