a16z: The real opportunity for stablecoins lies not in disrupting, but in filling gaps

Writing: Noah Levine, a16z Investment Partner Compilation: Saoirse, Foresight News A few weeks ago, Citrini Research published an article stating that stablecoins will bypass Visa and Mastercard, directly causing a sharp drop in the stock prices of card organizations. The encryption circle is booming. This logic sounds very clear: AI will optimize every transaction, and the transaction fee is a kind of "tax", while stablecoins can bypass it. I have been immersed in the field of encryption all day, hoping that this statement is correct, but most of it is wrong. Not because stablecoins are not important, but because the real opportunity is not to replace bank cards, but to serve merchants who find it difficult to access traditional card payments. Citrini's argument that bank cards will dominate the vast majority of the market is based on the assumption that AI agents that break free from human habits will actively optimize card organization fees. But bank cards are not just transfer tools. It provides unsecured credit, pre authorization for uncertain transactions, and fraud protection through the right to refuse payment. Stablecoins can transfer money, but they cannot do the rest. Assuming your agent helped you book a hotel, the result is completely different from the picture. With a bank card, you can initiate disputes and recover money. With stablecoins, money cannot be returned once it goes out. 82% of Americans hold reward credit cards (referring to gifts such as cashback, points, air miles, hotel points, etc.), with a global circulation of up to 18 billion cards. For the vast majority of transactions, consumers will not voluntarily give up consumer protection and points to choose a payment method that has no benefits and is irreversible. Fraud detection is a huge advantage for card organizations: card networks can run models for billions of transactions in real-time. There is currently no network level anti fraud layer comparable to stablecoins. Small payments are often referred to as the weakness of bank cards, but card organizations have long adapted to such mismatched transactions. Visa has processed over 2 billion transportation tickets by consolidating multiple card swipes into daily settlements. The card industry has never given up on any type of transaction, it always invents new products to cover it. There is also a question: 'Intelligent agents cannot hold cards.' However, intelligent agents are essentially just new devices. Your phone, watch, and computer all hold independent tokens pointing to the same card, just like Apple Pay. Mobile phones have never undergone KYC, they only hold your token, and the same goes for intelligent agents. Visa has issued over 16 billion tokens, which will also be used by intelligent agents. Visa's intelligent commerce framework is currently being piloted, and Mastercard's Agent Pay has been launched for cardholders across the United States. The intelligent commerce protocol jointly built by Stripe and OpenAI has been integrated into Etsy, and over a million Shopify merchants are about to go live. The conclusion is clear: for existing merchants and consumers, bank cards are almost destined to dominate intelligent commerce. The opportunity for stablecoins lies elsewhere - in merchants who have not yet emerged. Every platform migration of merchants who have not yet emerged will give rise to a wave of merchants that existing payment systems cannot serve. When eBay appeared, individual sellers couldn't even open merchant accounts, so PayPal served them; Shopify grew from 42000 merchants to 5.5 million in 13 years; When Stripe was founded, many of its customers had not even been born yet. The pattern remains consistent: the winner serves merchants that existing giants cannot underwrite. The AI wave will generate such businesses faster than any previous platform migration. Last year alone, 36 million new developers joined GitHub. In the winter batch of YC 2025, more than 95% of the company's code repositories are generated by AI. On the popular AI programming platform Bolt.new, 67% of the 5 million users are not developers at all. People who couldn't write production grade code two years ago are now releasing software. They are both buyers and sellers of developer services. Imagine: an ordinary developer spent 4 hours using AI tools to create a financial data display tool for a listed company. There is no website, no terms of service, and no legal entity. Another developer's agent calls it 40000 times a week, at a rate of 0.1 cents per call, generating $40 in revenue. No one clicked on the checkout page throughout the entire process. I see developers creating such tools every week. The first question they always ask is: How do I collect money? For most people, the answer is: currently unable to receive it. It is difficult for existing payment institutions to access such merchants. It's not that the technology is not good, but once the payment institution passes through the merchant, it has to bear its risks. If the merchant engages in fraud or generates a large number of chargebacks, the payment institution will be held responsible. Without a website, entity, or record keeping tool, it is almost impossible to pass the risk control audit. The system is running according to its design, but it was not originally designed for this scenario. Payment institutions can certainly make adjustments, as they have done before. But it took PayPal 16 years from its launch to the industry's first underwriting guidelines for payment service providers. And these new merchants need to collect money now. For them, accepting stablecoins is like street vendors only accepting cash. It's not that cash is better, but these types of merchants have always had difficulty obtaining bank card acceptance qualifications. In this gap, stablecoins are currently the only feasible solution. Although the wallet experience is rough and the compliance framework is still forming, protocols like x402 can already embed stablecoin payments directly into HTTP requests: no merchant account, no processor, no login, and no chargeback liability. These merchants are not choosing between stablecoins and bank cards. They are choosing between stablecoins and not receiving money. New businesses will emerge from here, and every wave of new merchants will eventually be absorbed by traditional payment systems, and this time it is highly likely to be the same. But the order is always: the merchant appears first, followed by risk control. In the gap between these two periods, stablecoins are the infrastructure. The bank card service covers all merchants that can be insured by payment institutions; Stablecoins serve merchants that cannot be insured by any payment institution. The next wave of business will be born in this gap.