a16z crypto Analysis: Agentic Commerce Won't Kill Cards, But Will Open a Payment Gap

a16z crypto's Structural Analysis of Payment Infrastructure

Agentic commerce won't kill cards, but it'll open a gap

a16z crypto analysis on payment infrastructure for agentic commerce

a16zcrypto.com

On March 4, 2026, Noah Levine of leading VC fund a16z crypto published an analysis of payment infrastructure in the agentic commerce era. Titled "Agentic commerce won't kill cards, but it'll open a gap," the piece offers a measured structural analysis in response to the market turmoil that followed Citrini Research's February 2026 claim that AI agents would use stablecoins to bypass Visa and Mastercard interchange fees, which sent card-related stocks down 4-6%.

Why Card Networks Will Survive

The core of Levine's argument is straightforward. Card networks do far more than move money. They provide unsecured credit, pre-authorize uncertain transactions, and guarantee fraud protection through chargeback rights. Stablecoins can move money, but they do not yet offer these capabilities.

He illustrates this with a concrete example: if an AI agent books a hotel room that does not match the listing, a card payment allows you to dispute the charge. With stablecoins, the payment is irreversible and there is no refund mechanism. In the US, 82% of consumers carry a rewards card, and 18 billion cards are in circulation globally. Consumers will not voluntarily give up purchase protection and rewards for an irreversible payment that offers neither.

He also pushes back on the argument that "agents can't hold cards." An agent is simply a new device, linked to existing cards via the same tokenization technology as Apple Pay. Visa has already issued over 16 billion tokens. Visa's "Intelligent Commerce" is in pilot, and Mastercard's "Agent Pay" is live for all US cardholders.

The Real Opportunity: Merchants That Don't Exist Yet

The most significant part of Levine's analysis is his argument that the real opportunity for stablecoins lies not in replacing cards, but in serving emerging merchants that existing payment systems cannot accommodate.

Historically, every platform shift has created a wave of merchants that the existing payment system could not serve. When eBay created peer-to-peer commerce, those sellers could not get merchant accounts. PayPal filled that gap and grew to handle 40% of eBay auction payments by 2000. Shopify expanded from 42,000 merchants to 5.5 million over 13 years.

In the AI era, the emergence of these new merchants is accelerating further. GitHub gained 36 million new developers in the past year alone. In YC's Winter 2025 batch, a quarter of companies had codebases that were 95% or more AI-generated. On AI coding platform Bolt.new, 67% of 5 million users are not developers.

Consider a "vibe coder" who builds an API tool in a few hours that gets called 40,000 times a week by another developer's agent at a tenth of a cent per call, generating $40 in revenue. No website, no legal entity, no terms of service. Existing payment processors find it extremely difficult to underwrite and approve such merchants.

When a processor approves a merchant, it takes on that merchant's risk. If fraud or chargebacks occur, the processor is liable. It took 16 years from PayPal's launch for the industry to develop underwriting guidelines for the payment facilitator model PayPal pioneered. These emerging merchants need payment solutions now.

Stablecoins Fill the Payment Gap

Stablecoins fill this structural gap. Levine describes it as "the equivalent of a street vendor only taking cash." Not because cash is better, but because merchants with that profile have historically struggled to get approved for card acceptance.

The x402 protocol already embeds stablecoin payments directly into HTTP requests. No merchant account, no processor, no onboarding, no chargeback liability. Through Coinbase and the x402 Foundation, over 119 million transactions have been processed on Base and 35 million on Solana as of March 2026.

Meanwhile, the Agentic Commerce Protocol (ACP) co-developed by Stripe and OpenAI represents the card-network approach. Etsy is already live, with over one million Shopify merchants set to follow. Two parallel payment infrastructures, card-based and stablecoin-based, are now clearly emerging to serve different merchant segments.

Impact on E-commerce Operators

The key message of this analysis for e-commerce operators is not a binary choice between cards and stablecoins, but rather that the optimal payment method will vary by merchant profile.

For established e-commerce businesses, card networks will remain the primary payment channel. Adapting to Visa Intelligent Commerce and Mastercard Agent Pay should be a priority. However, for businesses offering API-based services, microtransactions, or individual developers selling services without a legal entity, stablecoin payments are becoming a realistic option.

E-commerce operators should verify that their checkout flows can accommodate AI agent access while also considering flexible payment architectures that could accept stablecoin payments in the future.

Looking Ahead

Levine of a16z crypto argues that the question of whether agentic commerce "kills or spares" cards is itself the wrong question. Card networks will dominate agentic commerce for established merchants, while stablecoins will rise among emerging merchants that cannot pass existing underwriting and onboarding processes.

The key factor to watch is how large this "gap" will grow. The speed at which the developer economy expands through vibe coding, versus the speed at which existing payment processors adapt their underwriting systems. This gap between the two will determine the market size of stablecoin payments.