Anthropic's Super Bowl attack ads take aim at OpenAI's ads product — and the gloves are off
Feb 4, 2026
Key Points
- Anthropic launches a four-ad Super Bowl campaign attacking OpenAI's incoming ads product, depicting AI systems recommending payday loans and predatory services without naming competitors directly.
- The $70-$80 million campaign targets OpenAI's unproven promises to separate ads from content, amplifying market uncertainty about whether AI systems will prioritize advertiser interests over user safety.
- Anthropic's refusal to include download prompts suggests the ads aim to embarrass OpenAI and influence regulatory perception before an potential IPO, not win consumer market share.
Summary
Anthropic is running a four-ad Super Bowl campaign that attacks OpenAI's incoming ads product. The spots portray AI systems steering users toward payday loans, predatory dating sites, and other harms. While OpenAI and ChatGPT go unnamed, the timing is unmistakable—the campaign launches as OpenAI begins rolling out ads within ChatGPT.
One ad shows Claude recommending a high-APR payday lender to a woman seeking business advice. Another suggests a mature dating site as a solution to family conflict. A third mimics the ChatGPT voice pattern and cadence to pitch height-increasing insoles with a discount code. The fourth directly attacks the ads product itself.
The campaign exploits a real gap in the market. OpenAI has committed to gatekeeping advertisers and controlling ad context, but those promises exist only in blog posts and interviews. The ads weaponize consumer fear that OpenAI cannot or will not enforce those boundaries. The timing amplifies the effect, arriving as OpenAI's ads rollout begins and before the company can establish its own narrative.
Anthropric's decision to omit a call-to-action or download link suggests the campaign is not aimed at consumer acquisition. Claude averages around one million downloads per month and has not cracked the top 25 apps. Instead, the ads appear designed to embarrass OpenAI, shape regulatory perception, and create talking points for a potential IPO roadshow.
Super Bowl attack ads are rare and rarely this direct. The 1984 Apple ad portrayed IBM as authoritarian. The Mac-vs-PC campaign mocked Windows vulnerabilities. Those campaigns ran years into market maturity, not immediately before a competitor's major product launch. Bud Light's ads attacking Miller Lite and Coors Light on corn syrup work because corn syrup is verifiable on ingredient lists.
Anthropric's ads are speculative. OpenAI has promised to separate ads from content, likely using the Instagram model of clearly labeled sponsored results alongside organic responses. But Anthropric's spots suggest that boundary will collapse and that AI systems will be corrupted by advertiser influence. That claim exceeds what the facts support and echoes longstanding concerns from Mark Cuban that most observers believe have been overblown.
OpenAI cannot easily respond. Ad buys at this scale lock in weeks in advance. Production timelines and compliance review are immovable. Even a counter-attack before the Super Bowl is impractical.
The campaign signals a thaw in the cold war between Anthropic and OpenAI. Dario Amodei has publicly criticized ads-based models at Davos. The talent wars of summer 2024 left visible scars, including Mark Chen's leaked memo about something being stolen. Now Anthropic is spending roughly $70 to $80 million on a coordinated effort to slow OpenAI's expansion into consumer monetization. It reads less like a bid for market share and more like a direct attempt to damage a dominant player's growth.