Sam Altman Bristles at Anthropic’s Claude Super Bowl Ads
Staff Reporter
05 February 2026, 17:06
Anthropic’s Super Bowl debut quickly drew attention in Silicon Valley. Nowhere was the reaction sharper than at OpenAI.
Sam Altman, OpenAI’s chief executive, responded with an unusually personal rebuke. His reaction highlighted rising tensions inside the artificial intelligence industry.
One of Anthropic’s four commercials opens with a single word: “BETRAYAL.” It flashes boldly across the screen.
The ad then cuts to a man asking a chatbot for advice. The chatbot is clearly meant to resemble ChatGPT. The man wants help talking to his mother.
The bot, portrayed by a blonde woman, starts with familiar guidance. Listen carefully. Take a calming nature walk.
Then the tone shifts. The chatbot suddenly promotes a fictional cougar-dating site called Golden Encounters. The exchange becomes a parody of intrusive advertising.
Anthropic ends the spot with a clear message. Ads may be coming to AI chatbots. They will not be coming to Claude.
A second commercial follows a similar approach. A young man asks a chatbot how to build a six-pack. He provides his height, age, and weight. The response does not offer fitness advice. Instead, it serves an ad for height-boosting insoles.
The ads were widely seen as a response to OpenAI’s plans. The company recently said ads will be added to ChatGPT’s free tier. Headlines quickly said Anthropic was “mocking” OpenAI.
Altman said on X that he laughed at the commercials. His tone soon changed.
He followed with a long post criticizing Anthropic. He called the ads “dishonest.” He later described the company as “authoritarian.”
Altman argued that advertising is necessary. He said it helps keep ChatGPT free for millions of users. ChatGPT remains the most widely used chatbot in the world.
He rejected the idea that ads would distort conversations. He also dismissed suggestions of inappropriate promotions.
“We would obviously never run ads in the way Anthropic depicts them,” Altman wrote. “We are not stupid and we know our users would reject that.”
OpenAI says ads will be clearly labeled. They will be separate from chatbot responses. They will not influence the conversation itself.
At the same time, OpenAI has acknowledged a key detail. Ads will be conversation-specific. That is the core claim highlighted in Anthropic’s ads.
In a blog post, OpenAI said it plans to test ads at the bottom of answers. The ads would appear when a relevant sponsored product is connected to the conversation.
Altman’s criticism went further. He claimed Anthropic serves “an expensive product to rich people.” He contrasted that with OpenAI’s goal of reaching billions.
The pricing tells a more balanced story. Claude offers tiers at $0, $17, $100, and $200. ChatGPT’s tiers are $0, $8, $20, and $200.
Altman also accused Anthropic of exerting tight control over its tools. He said it blocks access to Claude Code for some organizations. OpenAI, he claimed, is among them.
Anthropic was founded by former OpenAI employees. From the start, it has emphasized “responsible AI.” Safety and usage limits are central to its identity.
Both companies maintain strict policies. Both enforce guardrails. Both block certain categories of content.
OpenAI allows some erotic content. Anthropic does not. Both restrict content related to mental health risks.
Altman escalated his language further. He described Anthropic as “authoritarian.”
“One authoritarian company won’t get us there on their own,” he wrote. “It is a dark path.”
The language drew criticism. The word carries weight in today’s global context. In some countries, protesters have been killed by their own governments.
Against that backdrop, some observers saw the comment as excessive. The dispute began with a satirical Super Bowl ad. Corporate rivals have long traded barbs through advertising. The Super Bowl has often been the stage. But this exchange struck a deeper nerve.
The response suggested a larger concern. How AI companies monetize matters. So does trust. As chatbots scale to billions of users, the balance between access, revenue, and control remains unresolved.
