AI 广告之争：OpenAI 与 Anthropic 的路线分歧

Tanuj Joshi is CEO of Eulerity, an AI-based multi-location digital marketing platform serving franchise and service-based businesses.

​The loudest debate in AI right now isn't about artificial general intelligence (AGI) timelines or model benchmarks, but whether your next conversation with a chatbot will come with a sponsored message.

In January 2026, OpenAI announced it would bring advertising to free ChatGPT users. Three weeks later, Anthropic ran a Super Bowl ad mocking them for it. Sam Altman fired back on X. Two of the most powerful companies in the world, each spending billions to shape the future of intelligence, were feuding over banner ads. This showcased differing theories of what AI is for and who gets to use it.

Since AI has always been at the core of Eulerity's DNA, the industry's shift toward ad-supported models validates our long-standing vision and allows us to lead clients through the transition from AI automation to AI-driven discovery. As CEO, I see this evolution as a critical expansion of the conversational ecosystem we have already mastered. This is especially important as marketers voice concerns that the "confessional" and personal nature of AI conversations could lead to incentive misalignment or the mishandling of sensitive user context.​

Understanding The Bill They’re All Paying

The intense debate around ads stems from the high cost of running AI systems. In 2025, OpenAI generated $13 billion in revenue (paywall), but infrastructure costs, including chips, cloud, data centers, electricity and talent, totaled an estimated $22 billion. They spent significantly more than they made.

Every AI conversation demands huge compute. Training a frontier model costs hundreds of millions, and running it at scale to serve millions of daily queries costs hundreds of millions more monthly, with electricity bills rivaling that of a small city.​

​OpenAI has reportedly committed to an estimated $600 billion in infrastructure spending through 2030, making deals with Nvidia, AMD, Microsoft Azure, Amazon Web Services, Oracle and CoreWeave that dwarf anything seen in prior technology. To put that in perspective: AI-related data center spending in 2025 already matched the dot-com peak as a share of U.S. GDP.

Anthropic projects spending roughly $200 million on compute over the next three years, which will cover model training, inference infrastructure, Google Cloud, Azure contracts and direct chip purchases. In February 2026, they made headlines by pledging to cover consumer electricity price increases caused by its data centers, an acknowledgment that its infrastructure footprint is significant enough to affect utility bills for ordinary Americans.

​Both companies are burning cash at a rate that makes even the most optimistic subscription growth projections look insufficient. This is why the ads conversation is happening now.

Why OpenAI Might Be Right

Following the blueprint of platforms like Google, Instagram and YouTube, the integration of advertising is essential for AI to scale toward a billion users, as the current model leaves roughly 850 million free users who are unlikely to pay a $20 monthly subscription fee. OpenAI is currently exploring sophisticated, intent-based ad formats that surface during the discovery phase, mirroring the utility of sponsored search where the advertisement itself provides the answer.

This evolution suggests a future where AI agents manage end-to-end commerce by fulfilling purchase intent through a transparently ranked combination of sponsored and organic options. Ultimately, shifting to an ad-supported model ensures that AI remains an accessible tool for the global majority rather than a luxury restricted to those who can afford a premium price.​

Why Anthropic Might Also Be Right

​The argument for advertising in AI often overlooks that many AI interactions are deeply personal "confessions" rather than simple, transactional search queries. Anthropic’s internal analysis reveals that users frequently treat AI as a trusted advisor for sensitive topics like career changes or family health, creating a private context that should remain shielded from ad targeting systems.

Beyond privacy, a fundamental concern is incentive misalignment. While the most useful AI interaction is often the briefest, ad-supported models are financially incentivized to maximize engagement and keep users in the app. History suggests that once revenue is tied to ad performance, platforms inevitably begin to prioritize advertisers over users, potentially turning helpful assistants into tools that subtly manipulate recommendations to drive profit.

What This Means If You’re A Marketer​​

In light of the shift from search to conversation, the traditional marketing funnel is being restructured into a high-intent, decision-heavy interface. For marketers, this is not just a new place to buy ads but a new way to influence the "moment of choice" where intent signals are extraordinarily high and AI answers carry more influence than any traditional advertisement.

To prepare, brands should immediately audit their conversational footprint by testing how major models describe their products and prioritize structured data to ensure AI agents can accurately parse their unique value propositions.

Teams must ask themselves if they have enough high-quality content to ensure favorable automated summaries. Also consider how to handle the "confessional" nature of these interactions, asking how to ensure brand presence appears in helpful, transactional moments rather than intrusive or sensitive contexts.

Marketers should evaluate whether they are optimized for decision making over simple search by identifying the specific data points that would make an AI prioritize them over a competitor. Ultimately, the goal is to partner with platforms that maintain the integrity of the user experience, turning this fragile period into a powerful customer acquisition channel through transparency, relevance and genuine utility.​​

​AI Advertising

For brands, this is simultaneously the most exciting and most fragile moment in the history of digital marketing. Done right—meaning with transparency, relevance and genuine utility—AI-native advertising could be the most powerful customer acquisition channel ever built. Done wrong, with the same engagement-optimization logic that hollowed out social media, it poisons the well for everyone.

The question isn't whether ads belong in AI. They belong somewhere in the ecosystem. The question is who controls the incentive structure when they arrive.​

Forbes Agency Council is an invitation-only community for executives in successful public relations, media strategy, creative and advertising agencies. Do I qualify?