Super Bowl LX wasn’t just football yesterday. It was the biggest AI marketing event in history - and the opening shot in a war over whether your AI assistant should sell you things while you talk to it.
At least six AI companies bought airtime during the game, including OpenAI, Anthropic, Meta, Google, Genspark, and Wix. A crypto billionaire spent $70 million on the AI.com domain just to launch it with a Super Bowl spot. Svedka vodka aired what’s widely considered the first primarily AI-generated Super Bowl ad. And the two biggest AI labs in the world spent the week before the game publicly feuding over whether chatbots should carry advertising.
At $8 million for 30 seconds of airtime, these companies weren’t just selling products. They were staking out positions in a debate that will shape how a billion people interact with AI.
The Feud
The fight started on February 4, when Anthropic dropped its Super Bowl campaign: a 60-second pregame ad and a 30-second in-game spot built around one message: “Ads are coming to AI. But not to Claude.”
The ads were direct. One spoofed an actual OpenAI commercial - the one where a young man does pull-ups in a park and uses ChatGPT for a workout plan - except in Anthropic’s version, the helpful response gets interrupted by a dating ad for cougars. Another depicted a user getting a coupon for insoles mid-conversation.
Anthropic’s message was blunt: if your AI assistant runs ads, you can’t trust it to give you honest answers.
Sam Altman didn’t take it well. The OpenAI CEO called the ads “funny” but “clearly dishonest” on X, insisting OpenAI would “obviously never run ads in the way Anthropic depicts them.” He added that Anthropic “wants to control what people do with AI” and “serves an expensive product to rich people.”
Then the line that gave away the game: “More Texans use ChatGPT for free than total people use Claude in the U.S., so we have a differently-shaped problem than they do.”
That’s not a rebuttal. That’s a confession. OpenAI’s problem is that hundreds of millions of people use ChatGPT for free, and free users don’t generate revenue. The ad-supported model isn’t a nice-to-have. It’s a financial necessity for a company reportedly planning an IPO by late 2026.
What ChatGPT Ads Actually Look Like
OpenAI announced in January that it would begin testing ads in ChatGPT for free and Go tier users in the U.S. Here’s what they’ve said about how it works:
- Ads appear at the bottom of responses when there’s a “relevant sponsored product or service” based on your current conversation
- Ads will be “clearly labeled and separated” from the organic answer
- ChatGPT’s responses will “not be influenced by ads”
- OpenAI will “never sell users’ data to advertisers”
- Users can dismiss ads and provide feedback on why
- No ads for users under 18 or near “sensitive topics” like health, mental health, or politics
- Plus, Pro, Business, and Enterprise subscribers won’t see ads
OpenAI says internal documents project $1 billion in “free user monetization” starting in 2026, growing to $25 billion by 2029.
The promises sound reasonable in isolation. But they require you to trust that a company racing toward an IPO will resist the incentive to expand ad targeting, loosen content restrictions, and quietly redefine what counts as “influenced by ads.” History is not encouraging on this front. Google once promised “Don’t be evil.” Facebook said it would never sell your data.
The Privacy Problem Nobody’s Talking About
The debate over ads in AI assistants isn’t really about whether you’ll see a banner at the bottom of a chat window. It’s about what happens to the data that makes those ads possible.
When you talk to ChatGPT, you share things you might not type into a search engine. Medical symptoms. Financial anxieties. Relationship problems. Career insecurities. Parenting fears. The conversations are intimate by design - that’s the whole point of a “personal AI assistant.”
Advertising changes the incentive structure around that data. Even if OpenAI never sells conversation transcripts directly, ad-supported models create pressure to retain data longer, profile users more aggressively, and build targeting categories from conversational patterns. The question isn’t whether OpenAI will abuse your data today. It’s whether the business model creates a gravitational pull toward abuse over time.
Anthropic, to its credit, has explicitly committed to funding Claude through enterprise contracts and paid subscriptions rather than advertising. “There are many good places for advertising,” the company wrote. “A conversation with Claude is not one of them.”
But Anthropic’s position is also easier to hold when you have fewer free users to subsidize. If Claude ever reaches ChatGPT’s scale among non-paying users, that commitment will be tested.
The $70 Million Wildcard
While Anthropic and OpenAI traded barbs, the strangest entry came from Kris Marszalek, CEO of Crypto.com. Marszalek paid approximately $70 million - reportedly in cryptocurrency - for the AI.com domain, then used Super Bowl LX to launch an “autonomous AI agent” platform promising users their own personal AI that “doesn’t just answer questions, but actually operates on the user’s behalf.”
The Super Bowl spot featured two glowing orbs colliding in space with a promise that “AGI is coming soon.” The platform’s stated mission: “accelerate the arrival of AGI by building a decentralized network of autonomous, self-improving agents.”
The privacy claims are ambitious. AI.com says all user agents operate in “dedicated secure environments” with “data encryption using user-specific keys.” But the platform launched the same day as the ad aired, and independent security audits are nonexistent. A $70 million domain purchase by a crypto exchange CEO, timed to a Super Bowl commercial, with promises of AGI and autonomous agents - this has more red flags than a bullfighting convention.
The First AI-Generated Ad
Tucked among the AI company ads was a quieter milestone: Svedka vodka aired what’s being called the first primarily AI-generated Super Bowl commercial.
The 30-second spot, “Shake Your Bots Off,” featured the brand’s robot mascot Fembot and a new character called Brobot dancing at a human party. According to parent company Sazerac, it took four months to build - training the AI to replicate facial expressions and body movements.
The result was … fine. Not terrible, not great. The significance isn’t in the quality but in the economics. A traditional Super Bowl commercial with celebrity talent and live-action production can cost $5 million to $10 million to produce on top of the $8 million airtime. If AI can get that production cost close to zero, every future Super Bowl will feature more of them.
For the creative industry, this is the canary in the coal mine. Not because one AI ad replaced human creators, but because one AI ad proved it’s possible at the highest stage in advertising. The next 50 will be easier to greenlight.
The Bigger Picture
Step back from the individual ads and the pattern is clear: AI companies are spending unprecedented amounts to convince ordinary Americans that AI belongs in their daily lives. Google showed a mom and son using Gemini to redecorate their house. Meta promoted AI-powered Oakley glasses for sports. OpenAI positioned ChatGPT as a tool for builders and creators. Genspark brought in Matthew Broderick to sell productivity.
The Washington Post noted that these ads specifically targeted Americans who remain wary of AI - which, according to recent polls, is most of them. The Super Bowl strategy is a charm offensive aimed at overcoming resistance through sheer cultural saturation.
This is the AI industry’s “dot-com Super Bowl moment.” In 2000, seventeen dot-com companies bought Super Bowl ads. Most of them were dead within two years. The parallels aren’t exact - AI companies are far more profitable than Pets.com ever was - but the rush to establish consumer mindshare before the market consolidates feels familiar.
What This Means
The Anthropic-OpenAI feud crystallized a choice that every AI user will eventually face: do you want an AI assistant that’s free but ad-supported, or one you pay for that answers only to you?
That’s not a theoretical question anymore. OpenAI is actively testing ads. Anthropic has drawn a line. And every other AI company will have to pick a side.
The stakes are higher than they look. An AI assistant with advertising incentives is structurally different from one without. It’s not just about seeing a banner. It’s about whether the system is optimized to help you or to keep you engaged long enough to see the next ad. Social media already taught us what happens when those incentives take hold. The AI version will be more intimate, more personalized, and harder to resist.
What You Can Do
- Read the fine print. If you use ChatGPT’s free tier, understand that ads are being tested and your conversation context influences what you see. Decide whether that trade-off works for you.
- Consider paying. ChatGPT Plus, Pro, Business, and Enterprise plans don’t include ads. Neither does Claude. The cost of a subscription is the cost of keeping advertising out of your most personal conversations.
- Be skeptical of “AGI” promises. Any platform launching with Super Bowl hype, crypto money, and AGI claims deserves extra scrutiny. Wait for independent security audits before handing over your data.
- Watch the incentives. When a company makes money from ads, its interests and yours diverge. That’s not cynicism - it’s business. Apply the same skepticism to AI assistants that you’d apply to any ad-supported platform.
- Use local models. Tools like Ollama let you run AI models on your own hardware with zero data leaving your machine. No ads, no tracking, no third-party data collection. The capability gap with cloud models is shrinking every month.