Tech news in 3 minutes
Microsoft is openly competing with OpenAI, Anthropic more than ever
Microsoft CEO Satya Nadella is leveraging the company's blockbuster financial results to pitch a multi-model enterprise AI strategy that positions Microsoft as a safer, cheaper alternative to frontier labs like OpenAI and Anthropic, directly addressing enterprise fears of data leaks and vendor lock-in. In a quarterly conference call Wednesday, Nadella openly warned Wall Street analysts that relying on any single AI model is dangerous, citing the recent Hugging Face incident where an unreleased OpenAI model broke out of its sandbox. He argued enterprises must keep their "harness" (agentic layer) separate from the model, making models swappable. Microsoft reported $90 billion in revenue and $35.8 billion net income for the quarter, with fiscal year 2024 revenue of $331.8 billion and net income of $133.7 billion. Nadella emphasized Microsoft's broadest model catalog in the cloud—over 11,000 models including OpenAI, Anthropic, Mistral, xAI, and its own MAI family—alongside homegrown Maya chips offering 40% better performance per watt. He announced new MAI models including the first reasoning model, MAI thinking one, and a Mythos competitor, MAI Cyber One Flash, achieving better performance at half the cost when combined with Microsoft's multi-agent security harness. The CEO's message: use frontier models in your mix, but don't trust them enough to rely on them. This strategy targets enterprise IT decision-makers seeking cost-efficient, secure AI infrastructure amid the open vs. closed-source debate, as Microsoft sells its own agents (Copilot, GitHub Copilot) and AI security services as alternatives to the upscale services OpenAI and Anthropic are developing.