Microsoft unveiled a new family of in-house AI models that it says cost up to 89% less than OpenAI equivalents for the same tasks. The announcement signals a shift: instead of relying almost exclusively on OpenAI models, Microsoft is diversifying its AI portfolio.
Microsoft had been criticized for outsourcing its AI brain to OpenAI. With its own data center hardware and the Phi model family as a base, it can now compete on cost. For enterprise Azure OpenAI customers, more options mean more bargaining power. For OpenAI, it's a warning sign — its biggest investor and cloud partner is becoming a competitor.
The real story isn't technical, it's strategic: Microsoft is following AWS's Nitro playbook — first use a third-party product to learn the market, then build your own with better margins.
Source: VentureBeat
