Anthropic vs OpenAI -- Who Owns the API? API Is Now a Weapon, Signal Shot of the AI Platform War: In August 2025, a small but symbolic incident shook the AI industry. AI startup Anthropic blocked OpenAI access to its representative service Claude API. According to multiple media reports, OpenAI used Claude code generation functions for internal testing -- specifically for GPT-5 feature benchmarking and capability comparison -- and Anthropic deemed this a violation of its Terms of Use prohibiting use for competitive product development, model training, or reverse engineering purposes. The API access block significance: this is the first known instance of a frontier AI company blocking a direct competitor from using its API; the API is now a competitive weapon, not just a developer service; the incident reveals the tension between the open AI ecosystem that all companies benefit from for research and development, and the competitive reality that each company wants to prevent competitors from using their capabilities to improve competing products. The broader API governance question: most AI companies allow API access to anyone paying -- the terms prohibit competitive use but enforcement is technically difficult; Anthropic blocking OpenAI access suggests companies are now actively monitoring who is using their APIs and for what purpose; this could shift the entire AI API ecosystem toward more restrictive access policies that fragment the developer ecosystem.