Anthropic is trolling OpenAI with Super Bowl ads targeting ChatGPT’s new ad model. We dive into the $60 CPM, the financial desperation of 95% free users, and why Claude is staying “pure.”
The AI industry just hit its “Instagram 2012” moment. As of February 2026, the era of frictionless, ad-free intelligence is dying. Anthropic knows it, and they’re using it to weaponize user trust.
Their Super Bowl LX campaign was a direct shot at OpenAI’s pivot toward monetization. The tagline wasn’t subtle: “Ads are coming to AI. But not to Claude.”
While Anthropic plays the “pure research” card, OpenAI is grappling with the harsh reality of hosting 800 million weekly users, 95% of whom have never paid a cent. To keep the lights on, OpenAI is rolling out its first ad-supported tier. The price tag for brands? A staggering $60 CPM.
The Trolling: Anthropic’s Super Bowl Blitz
Anthropic’s ads didn’t just showcase Claude’s technical chops. They directly lampooned the potential “ad-creep” in conversational AI. One viral spot showed a user asking for medical advice, only for the chatbot (a clear ChatGPT surrogate) to pivot into a sponsored pitch for orthotic insoles.
Sam Altman called the ad “dishonest,” but the data tells a different story. OpenAI is currently testing sponsored placements at the bottom of answers for free and “ChatGPT Go” users in the U.S.
At a $60 CPM (double the cost of a premium Meta ad), OpenAI isn’t just selling space. They’re selling your intimate conversational context.
Follow the Money: The Desperation of 95%
Why now? Because OpenAI reached a breaking point.
OpenAI hit an annualized revenue of $20 billion in 2025. In any other industry, that’s a triumph. In AI, it’s a rounding error. They’re projected to lose $14 billion in 2026. The “compute tax” is eating their lunch. As we noted in our breakdown of the $10B Cerebras bet, scaling isn’t just a research problem. It’s a literal GPU bottleneck.
CFO Sarah Friar, the architect behind Square and Nextdoor’s scale, had to plug the leak. With 95% of users burning expensive tokens for free, the “ChatGPT Go” $8/month tier wasn’t enough. Ads were the only lever left that didn’t involve cutting off the free tier entirely.
Revenue Models: Pure Research vs. High-Scale Commercial
The reason Anthropic can afford to “stay pure” isn’t just ethical superiority. It’s their customer base.
1. Anthropic: The Enterprise Engine
Anthropic is an “API-first” company. Roughly 85% of their revenue comes from enterprise token usage. Companies aren’t using the Claude chat interface. They’re piping tokens through backend servers for code migration and document analysis.
This is the SaaSpocalypse we predicted. Anthropic doesn’t need your $20 subscription. They need your company’s $200k monthly API bill. Only 15% of their revenue comes from consumers.
2. OpenAI: The Consumer Giant
OpenAI is the inverse. 73% of their revenue comes from consumer subscriptions. They are a consumer product company first. But as the inference costs for GPT-5.2-Codex scale, they need a secondary engine. Ads subsidize the R&D that subscriptions can’t cover alone.
3. Gemini: The Ecosystem Play
Google’s Gemini exists in a different dimension. As we saw with Gemini 3 Auto Browse, Google’s goal is to keep you in the Chrome ecosystem. They don’t need to show you an ad in the chat because they’ll get you in the search results later.
The UX Reality: Can AI Stay “Helpful”?
Conversational AI is a high-trust medium. When you ask a search engine for “best running shoes,” you expect a list of paid results. But when you ask an assistant for “advice on my career” and it suggests a sponsored MBA program, the illusion of an “impartial genius” evaporates.
OpenAI claims ads will stay at the bottom of the screen. But history suggests otherwise. We’ve seen “sponsored result creep” in Google Search for two decades. It starts as a footer. It ends as the entire first page.
If ChatGPT becomes a search engine with a chat interface (much like Perplexity‘s recent patent), the ad-free experience becomes a luxury good for the elite.
The $60 CPM: Why Brands are Salivating
Advertisers are tired of the blunt instruments of social media. OpenAI offers something better: intent.
In a world where GPT-4o is being retired for more powerful reasoning models, OpenAI knows exactly what you’re planning. If you spend 20 minutes discussing an upcoming move to London, an ad for a relocation service at the end of that chat is the most targeted placement in history.
OpenAI is betting that the utility of the model is so high that users will tolerate the friction. Anthropic is betting that users are so tired of being the product that they’ll switch to Claude for the peace of mind.
The Bottom Line
In the short term, OpenAI wins the revenue race. Adding a $60 CPM to 800 million users is a money printer that funds the path to AGI. It might even be the thing that keeps them independent after Amazon’s $50B investment.
But Anthropic’s “Pure Research” brand is a powerful moat. If they keep the lead in agentic coding, they become the professional standard.
In 2026, the choice is clear. You can pay $20 a month, look at an ad, or move to the “clean room” of Claude. My bet? The professionals will stay clean.
What about you? Are ads a dealbreaker for your AI, or is the power of GPT-5.2 worth the intrusion? Let me know in the comments.
