Microsoft reports expose AI's cost problem: The tech is more expensive than paying human employees
Fortune • • business
Key Points:
- Microsoft is canceling most of its direct Claude Code licenses just six months after encouraging widespread internal use, shifting engineers toward GitHub Copilot CLI instead, though this change does not affect its $5 billion Foundry deal with Anthropic.
- Uber and other tech firms like Meta and Amazon are also scaling back or reconsidering AI tool usage due to rapidly rising costs, despite initially incentivizing heavy adoption through internal competitions and leaderboards.
- The economics of AI adoption remain challenging, as increased use of token-based AI models drives up overall costs even while the price per token is expected to decline significantly by 2030.
- Gartner warns that falling token prices won’t necessarily reduce enterprise AI expenses because agentic AI models consume far more tokens per task and providers may not pass cost savings fully to customers, potentially leading to higher inference costs.
- Industry leaders like Nvidia’s CEO envision widespread deployment of AI agents alongside employees, but the escalating costs of token consumption could complicate the financial feasibility of such large-scale AI integration.