Microsoft axes internal Claude Code licenses over token costs

// 1h agoNEWS

Microsoft axes internal Claude Code licenses over token costs

Microsoft has reportedly canceled its internal Claude Code licenses after token-based billing made the costs untenable for widespread use. The decision highlights the financial risks of deploying autonomous coding agents at enterprise scale without predictable per-seat pricing.

// ANALYSIS

Token-based pricing for autonomous agents is hitting the enterprise wall.

–Uncapped agentic loops can consume millions of tokens in minutes, leading to unpredictable bill shock
–Enterprises strongly prefer the flat, per-seat licensing model established by GitHub Copilot
–Microsoft's internal cancellation is a strong signal that usage-based pricing hinders broad enterprise adoption of agentic tools
–The move may force Anthropic to offer flat-rate enterprise tiers for Claude Code to remain competitive

// TAGS

claude-codeai-codingcoding-agentdevtoolpricing

DISCOVERED

1h ago

2026-05-22

PUBLISHED

2h ago

2026-05-22

RELEVANCE

8/ 10

AUTHOR

ylecun

// KEEP READING

More AI developer news from the feed

EXPLORE FULL FEED

BENCHMARK1h ago

Qwen3.7-Max tops BridgeBench Flappy Bird

Alibaba's Qwen3.7-Max was released on May 20, 2026 as a proprietary agent model built for coding, tool use, and long-horizon execution. This post highlights its BridgeBench Flappy Bird output, which suggests unusually strong UI and game generation quality.

INFRA1h ago

Morgan Stanley Puts Rubin Rack at $7.8M

Morgan Stanley’s estimate suggests Nvidia’s next-generation Vera Rubin rack-scale system is getting materially more expensive than Blackwell, with a single VR200 NVL72 rack coming in at roughly $7.8 million and memory accounting for more than $2 million of that total. The broader takeaway is that AI system cost is now being driven by the entire rack bill of materials, not just the GPU, as memory, PCBs, substrates, and other components rise sharply in value.

FUNDING2h ago

DeepSeek nabs $10.29B, commits to open-source AGI

DeepSeek is advancing a $10.29 billion financing round while founder Liang Wenfeng explicitly rejects short-term commercialization. The AI lab is doubling down on developing open-source models to pursue its AGI goals.