Microsoft Limits Internal AI Token Use, Shifting Focus from Volume to Business Output

According to IT Home, citing TechRadar, Microsoft is tightening internal AI usage after engineers reportedly consumed unusually large amounts of tokens through GitHub Copilot and related tools. Updated guidance from July 2026 gives departments AI token budget targets and lets employees view their individual spending. Microsoft reportedly plans to make the lower-priced OpenAI GPT-5.6 the internal default model. Executive vice president Jay Parikh said the goal is not simply to use fewer tokens, but to ensure each token produces meaningful customer or business value. Microsoft has not disclosed specific spending caps. The report also places the policy alongside similar cost-control concerns at Uber and Amazon.