OpenAI is Visa

OpenAI is on the verge of becoming the Visa of artificial intelligence. Visa’s success wasn’t just about building a payments network; it was about creating barriers that locked in customers and locked out competitors. And just as Visa faced threats from national payment networks and tech giants, OpenAI must contend with competitors like Google, Meta, and Amazon.

In 1958, 60,000 Californians got a fully working credit card in the mail. It was the first unsolicited credit-card drop, and it led to massive fraud and delinquency problems. Bank of America, which ran the campaign, realized it had to build a payments network with account verification and fraud detection. The network it built and licensed to other banks eventually became Visa, which IPO’d in 2008 at a $44 billion valuation. Today the company is worth about $600 billion.

But in the 2010s Visa faced numerous threats to its core business. Digital providers like PayPal and well-capitalized businesses like JPMorgan Chase and Apple were poised to threaten Visa’s de facto monopoly on payment processing. On top of that, other developed countries were rolling out their own national real-time gross settlement programs, facilitating instant interbank transfers at scale for free. Payment processing had become commoditized.

What Visa did in response recently got it sued by the Justice Department, which accused Visa of using aggressive tactics with companies like Costco and Apple to guarantee that a competitive payment network would not develop, The Wall Street Journal reported. It also spent tens of millions of dollars lobbying Washington for more favorable payment regulations. That could help explain why it took the US until 2023 to launch its own national payment network (FedNow), whereas countries such as Poland launched theirs in 2012, Denmark in 2014, and India in 2016. 

As Visa’s technological moat dried up, it built a legal moat, and there are already signs OpenAI is doing the same.

OpenAI’s revenue is projected to reach $100 billion by 2029, according to The New York Times, but there’s a major risk factor. The underlying technology behind its revenue growth is the large language model, or LLM, but similar to what happened with payment processing, such models will soon become so ubiquitous that they might as well be free. Earlier this month, OpenAI boss Sam Altman essentially conceded this API business would dry up: “There will be shockingly capable models widely available, used for everything… the AI itself — the reasoning engine — will become commoditized.”

Dozens of other companies, including Google, Meta, and most recently Amazon, have come out with their own foundational models. Some, including Meta’s Llama and Mistral’s 7B, are open source, meaning they can be downloaded and used in other companies’ products free of charge. Apple is rumored to be working on an LLM that can fit on your iPhone.

To counteract these threats, OpenAI appears to be taking a page from Visa’s playbook. Last year, Altman “stormed Washington,” urging lawmakers to regulate AI. And, in OpenAI’s latest funding round, participating investors were asked not only to abstain from investing in competitors including Anthropic and SSI but not to fund any application-layer companies such as Glean and Perplexity.

These efforts signal an attempt to dominate the market, not through superior technology, but by limiting competition through exclusivity deals, government contracts, and licensing requirements for advanced AI models. 

OpenAI could prevent rivals from competing fairly through securing government deals that would mandate it as the arbiter of AI procurement; limiting competitors’ access to talent, chips, data centers, or energy through exclusivity arrangements with partners; or long-term exclusivity contracts with large customers (OpenAI already claims 92% of the Fortune 500 as customers, according to CNBC).

But this strategy may face political and competitive roadblocks. With Elon Musk emerging as Altman’s chief rival and exerting influence through figures like White House AI czar David Sacks, government regulation is likely to relax rather than tighten. If OpenAI can’t build strong barriers to entry soon, it risks losing its edge in an increasingly crowded and democratized AI landscape. Either way, Visa and OpenAI seem to agree on one thing: that “competition is for losers.”

Read the other arguments for OpenAI's future here.

Taylor Lorenz publishes User Magazine, a tech and online culture newsletter. She is the author of the book “Extremely Online: The Untold Story of Fame, Influence, and Power on the Internet” and host of the podcast Power User. She has previously written for The New York Times, The Atlantic, The Washington Post, and more.