Onstage with Indian Prime Minister Narendra Modi at the India AI Impact Summit, Altman and Amodei seemingly refused to touch during a photo op. Modi grabbed Altman’s hand and lifted it for a group photo, but Altman and Amodei, standing side by side, did not clasp hands or make eye contact. 

While other leaders like Google’s Sundar Pichai and Meta’s AI chief Alexandr Wang held hands for the shot, the rival CEOs of OpenAI and Anthropic instead raised their fists in the air.

“[About] the thing on stage. We didn’t know—I didn’t know what was happening. I was confused, like Modi grabbed my hand and put it up and I just wasn’t sure what [we were] supposed to be doing,” Altman later told reporters.

Still, the visible on-stage tension between Altman and Amodei illustrates the extent of the pair’s fractured relationship and how it has escalated into a bitter rivalry, even though just a few years ago the two CEOs were on the same team. 

Super Bowl scuffle

The most recent conflict between Altman and Amodei came after Anthropic, earlier this month, released a four-ad Super Bowl campaign titled “A Time and a Place,” indirectly poking fun at OpenAI’s recent move to include ads in ChatGPT. 

Each commercial opened with a single, loaded word splashed across the screen: “betrayal,” “deception,” “treachery,” and “violation.” In each of the commercials, ordinary people seeking advice from an artificially-sounding person standing in for the AI chatbot were interrupted by unwarranted advertisements. 

The AI assistant would begin answering helpfully before abruptly pivoting to pitch an unrelated product. The tagline was blunt: “Ads are coming to AI. But not to Claude.” 

Altman was quick to respond publicly to the campaign, saying in a lengthy post on X the company would never run ads in the way Anthropic showed in their advertisements. He also called Anthropic’s ads “clearly dishonest” and accused the company of “doublespeak.” 

“Anthropic serves an expensive product to rich people. We are glad they do that and we are doing that too, but we also feel strongly that we need to bring AI to billions of people who can’t pay for subscriptions,” Altman wrote in the post.

Yet, several experts later praised the advertisement, including Australian marketing journalist Mark Ritson, who said in a post it was “the first piece of effective brand strategy the AI category has produced.”

Altman’s response didn’t counter Anthropic’s attack as he might’ve thought it did, added Scott Galloway, a New York University marketing professor, during an episode of the podcast Prof G Markets.

“When you’re the market leader … you don’t reference the competition,” Galloway said.

From partners to rivals

The two AI leaders weren’t always adversaries. Amodei worked under Altman at OpenAI from 2016 to 2020. His last role was vice president of research, where he focused on safety. During his time at OpenAI, he played an instrumental role in launching GPT-2 and GPT-3. He also co-invented “reinforcement learning from human feedback,” which uses human input to train large language models to deliver the best outputs.

But later Amodei, along with a small group of employees, became concerned about OpenAI’s direction and what he perceived as a lack of safety involved in scaling up ChatGPT.

In 2021, Amodei, along with his sister and cofounder Daniela Amodei and a dozen other ex-OpenAI employees started Anthropic with a mission to put safety first. The company released its large language model, Claude, in 2023.

The employees who jumped ship from OpenAI believed two things, Amodei told Fortune’s Jeremy Kahn at the Fortune Brainstorm Tech conference in 2023. First, they believed if you pour more compute into large language models, they will greatly improve, and there’s almost no limit to how advanced they can get. Second, they believed—in addition to scaling up the models—there needed to be a strong set of guardrails and safety considerations.

“There were a set of people who believed in those two ideas. We really trusted each other and wanted to work together. And so we went off and started our own company with that idea in mind,” Amodei told Fortune.

When Altman was abruptly ousted as OpenAI’s CEO in November 2023, the board reportedly approached Amodei to replace him, and even considered merging the two AI startups, Reuters reported. Amodei reportedly declined both offers.

Growing competition

From a dozen employees at its start to now more than 2,500, Anthropic has rapidly evolved into a formidable competitor to OpenAI.

The company has secured billions in funding from Google, Salesforce, and Amazon, among others, and last week raised $30 billion in a funding round that cemented it as one of the world’s most valuable private companies with a post-money valuation of $380 billion. 

The awkward hand-holding moment Thursday comes less than two weeks after the companies’ Super Bowl clash, suggesting a fallout is still lingering. 

While Altman said he was simply confused, the episode underscored that he and Amodei remain locked in a high-stakes battle for AI dominance that is increasingly playing out in public.