Anthropic takes a dig at OpenAI, says it is building an AI business and not chasing headlines

Anthropic AI

• Anthropic focuses on enterprise AI adoption, avoiding flashy headlines and ads.
• Anthropic critiques OpenAI's ad strategy, emphasizing utility and reliability.
• Anthropic pledges $50B for data centers, while OpenAI plans $1T infrastructure.

Did our AI summary help?

Anthropic is not interested in “flashy headlines”. At least, that is the message its commercial leadership is keen to project as the rivalry with OpenAI turns increasingly public.

In an interview with CNBC, Anthropic’s commercial chief delivered a thinly veiled critique of OpenAI’s strategy, positioning Anthropic as a company focused on revenue growth and enterprise adoption rather than attention-grabbing announcements. The remarks follow a brief but pointed advertising skirmish between the two companies during Sunday’s Super Bowl.

Anthropic aired two commercials, including a 60-second pregame spot, built around a simple message: “Ads are coming to AI. But not to Claude.” The campaign was widely interpreted as a jab at OpenAI’s decision to begin testing advertising within ChatGPT. OpenAI chief executive Sam Altman responded by calling the ads “deceptive”, while conceding they were “funny”.

Anthropic has been explicit about why it wants to avoid advertising. According to the company, ads would push it towards optimising for engagement rather than utility, safety and reliability. Without an advertising business to serve, it argues, it can focus on making models that integrate cleanly into enterprise workflows and scale across organisations without conflicts of interest.

The contrast with OpenAI is deliberate. While OpenAI has built ChatGPT into one of the world’s most recognisable consumer AI products, Anthropic has concentrated on selling Claude and its related tools directly to businesses. The company argues that this keeps incentives aligned, with no need to compete internally for user attention or ad inventory.

The divergence extends beyond business models to infrastructure spending. As investors scrutinise ballooning capital expenditure across Big Tech, Anthropic has sought to portray itself as disciplined rather than restrained. The company has committed around $50 billion to US data centre capacity while continuing to buy compute from partners such as Google and Microsoft.

That figure looks modest next to OpenAI’s ambitions. OpenAI has outlined more than $1 trillion in future infrastructure commitments, backed by partners including Nvidia, Oracle and Broadcom. Recent deals include a $100 billion Nvidia-backed buildout and a $10 billion agreement with Cerebras for AI chip deployments.

Anthropic’s leadership insists that restraint is intentional. The company says it reviews compute spending daily and aims to buy capacity as close to demand as possible. Management argues that it is scaling aggressively, but not speculatively, and that its growth in products such as Claude Code and Claude Cowork reflects genuine enterprise demand rather than hype.

That distinction became relevant last week when software stocks sold off sharply after Claude Cowork gained traction. Investors worried that AI tools could replace entire layers of enterprise software. Anthropic dismissed the reaction as exaggerated, arguing that traditional software still performs specialised tasks that large organisations will rely on for years.

Some companies may consolidate tools, Anthropic acknowledged, while others will continue to invest heavily in existing platforms. Either way, the company says the market reaction overstated how quickly AI can unpick deeply embedded enterprise systems.