In a significant escalation of tensions between artificial intelligence rivals, Anthropic has cut off OpenAI's access to its Claude family of AI models, marking a new chapter in the increasingly competitive landscape of AI development.
The Decision
According to sources familiar with the matter, Anthropic made the decision to revoke OpenAI's API access after discovering that the ChatGPT maker was using Claude for internal benchmarking purposes. The move, first reported by Wired, represents one of the most direct confrontations yet between two of the industry's leading AI companies.
OpenAI had been connecting Claude to internal tools that allowed the company to systematically compare Claude's performance against its own models across multiple categories, including coding capabilities, writing quality, and safety measures. This practice, according to Anthropic, constituted a direct violation of their commercial terms of service.
Terms of Service Violation
An Anthropic spokesperson confirmed the decision in a statement to Wired, explaining that "OpenAI's own technical staff were also using our coding tools ahead of the launch of GPT-5," describing this as "a direct violation of our terms of service."
Anthropic's commercial terms explicitly forbid companies from using Claude to build competing services, a restriction designed to prevent rivals from leveraging their technology for competitive advantage. The company did indicate, however, that it would continue to provide OpenAI access for "benchmarking and safety evaluations" - suggesting some level of cooperation will remain for research purposes.
OpenAI's Response
OpenAI pushed back against the characterization of its usage as problematic. In a statement to media outlets, an OpenAI spokesperson described the company's usage of Claude as "industry standard" practice, suggesting that such competitive analysis is commonplace in the AI sector.
"While we respect Anthropic's decision to cut off our API access, it's disappointing considering our API remains available to them," the OpenAI spokesperson added, highlighting what the company sees as an asymmetric approach to access.
Pattern of Resistance
This decision is not Anthropic's first move to limit competitors' access to its technology. The company has previously demonstrated resistance to providing access to other competitors, most notably when it cut off access to Windsurf, a coding assistant that was rumored to be an OpenAI acquisition target before being acquired by Cognition.
Anthropic Chief Science Officer Jared Kaplan previously justified such decisions, stating, "I think it would be odd for us to be selling Claude to OpenAI." This philosophy appears to reflect a broader strategic approach of limiting how competitors can leverage Anthropic's technology.
Market Context
The clash comes at a time when both companies are experiencing significant growth and market positioning changes. Recent reports indicate that Anthropic has been gaining substantial ground in enterprise markets, with the company holding 32% of enterprise large language model market share by usage - a dramatic reversal from just two years ago when OpenAI commanded 50% of the enterprise market.
Anthropic has been particularly successful in coding applications, capturing 42% of the enterprise market share in that category, more than double OpenAI's 21% share. This success has been driven largely by the performance of Claude 3.5 Sonnet, released in June 2024, and the more recent Claude 3.7 Sonnet launched in February 2025.
Meanwhile, both companies are reportedly operating at significant losses despite growing revenues. Anthropic is expected to lose approximately $3 billion in 2025 on $4 billion in annualized revenue, while continuing to invest heavily in model development and infrastructure.
Industry Implications
The decision to cut off access highlights the increasingly competitive nature of the AI industry, where companies must balance collaboration on safety research with protecting their competitive advantages. While the AI community has historically emphasized open research and collaboration, commercial pressures are creating new tensions.
The move also underscores the strategic importance of API access in the AI ecosystem. Companies like Anthropic generate substantial revenue from API usage, but providing access to direct competitors creates a complex dynamic where rivals can study and potentially reverse-engineer competitive advantages.
Looking Forward
This development signals a potential shift toward more restrictive access policies among AI companies as the industry matures. While safety research and academic collaboration may continue, commercial applications appear increasingly likely to face restrictions.
The tension between Anthropic and OpenAI reflects broader questions about how AI companies will navigate competition while maintaining the collaborative spirit that has historically driven AI research forward. As these companies continue to develop more powerful AI systems, the balance between competition and cooperation will likely remain a defining challenge for the industry.
Both companies continue to pursue aggressive growth strategies, with OpenAI reportedly preparing to launch GPT-5 and Anthropic recently securing additional funding to support its expansion. The revocation of API access may be just the beginning of more assertive competitive positioning as the AI market continues to evolve.