AI Giants' Showdown Escalates: Anthropic Accuses OpenAI of Breach of Contract, Cuts Off Claude Access

August 03, 2025
TechCrunch, Wired, Slashdot
4 min

News Summary

According to multiple tech media reports, AI company Anthropic officially revoked competitor OpenAI's API access to its Claude series of AI models on Tuesday (August 1). This decision marks a further intensification of competition in the AI industry and exposes the increasingly strained relationship between the two top AI companies.

Core of the Incident

Anthropic accused OpenAI of violating its terms of service, specifically by having OpenAI's technical team use the Claude Code programming tool to assist in development work ahead of the upcoming GPT-5 release. Anthropic spokesperson Christopher Nulty told Wired magazine: "Claude Code has become the preferred tool for programmers worldwide, so it's not surprising to learn that OpenAI's technical staff were also using our programming tool before the GPT-5 release. Unfortunately, this directly violates our terms of service."

According to sources, OpenAI connected Claude to internal tools via a special developer API to compare Claude's performance with its own models in areas such as programming, writing, and security. The company also tested how Claude responded to safety-related prompts involving sensitive content (e.g., child safety materials, self-harm, and defamation).

Responses from Both Sides

Anthropic's Stance:

  • Business terms explicitly prohibit customers from using Claude to build competing products or services, including training competing AI models.
  • Will continue to provide OpenAI with "API access for benchmarking and safety evaluations, which is standard industry practice."
  • This move is seen as a necessary measure to protect its technological advantage and business interests.

OpenAI's Response: OpenAI Chief Communications Officer Hannah Wong issued a statement saying: "While we respect Anthropic's decision to cut off our API access, it is disappointing given that our API remains open to them." OpenAI described its usage as "industry standard," arguing that evaluating other AI systems to advance benchmarking and improve safety is necessary R&D work.

Historical Context

This is not the first time Anthropic has restricted competitor access. In June, when rumors circulated that OpenAI would acquire AI programming tool startup Windsurf for $3 billion, Anthropic immediately cut off Windsurf's direct access to its Claude 3.5 Sonnet and Claude 3.7 Sonnet programming models.

Anthropic co-founder and Chief Scientist Jared Kaplan frankly stated at a TechCrunch event: "I think it would be strange to sell Claude to OpenAI." He further elaborated on the company's strategic thinking, defining this decision as a necessary allocation of scarce resources.

Industry Impact

This incident reflects broader tensions in the AI industry, where companies are increasingly restricting competitors' access to their APIs. This strategy has previously been used by Facebook against Twitter's short-video app Vine, and by Salesforce against Slack competitors.

Technical Impact:

  • The Claude Code tool is facing service pressure due to surging demand, and Anthropic recently announced new usage limits.
  • Company data shows that some heavy users ran Claude Code for 24 consecutive hours, with a single user consuming tens of thousands of dollars in model usage on a $200 plan.

Market Impact:

  • This access revocation comes at a critical time as OpenAI prepares to release GPT-5, making the timing quite sensitive.
  • Industry insiders say OpenAI is currently in "full crisis mode," with only weeks remaining until the GPT-5 release.

Future Outlook

Despite the competition, the two companies still need to cooperate on certain standard industry practices. Anthropic stated it would continue to provide necessary API access for benchmarking and safety evaluations, indicating that even amidst fierce competition, industry standards and safety collaboration remain significant.

This incident highlights the complexity of competition in the AI industry, where companies need to both protect their technological advantages and bear responsibility for advancing the industry as a whole. As the GPT-5 release date approaches, the competitive relationship between the two companies will continue to be closely watched by the industry.