AI Cost Crisis Emerges as Claude Usage and Agentic Coding Bills Spiral

Enterprise AI spending is outrunning corporate forecasts. Microsoft has canceled most internal Claude Code licenses, and Uber has admitted it exhausted its 2026 AI budget within four months.

Token-based pricing on agentic coding tools has produced bills that outpace headcount savings. Companies are now retrofitting financial controls onto rollouts that moved fast in late 2025.

Microsoft and Uber Crystallize the Trend

Verge reporting said Microsoft started winding down most internal Claude Code licenses in mid-May 2026.

Most access in its Experiences and Devices division ends June 30. Engineers had heavily adopted the agentic coding tool.

Token-based billing made consumption unsustainable at deployment scale, Fortune reported. The pullback sits beside Microsoft’s own AI workplace report on 80% productivity gains.

Uber went further. Chief Technology Officer Praveen Neppalli Naga said the ride-hailing firm exhausted its full 2026 AI budget by April.

The company had deployed Claude Code to about 5,000 engineers four months earlier.

Forbes reported per-engineer costs reaching $500 to $2,000 monthly. Roughly 70% of committed code now comes from AI tools, signaling a growing Claude reliance across major engineering teams.

Industry Data Confirms a Wider Squeeze

A 2025 survey from Mavvrik found 85% of companies miss AI cost forecasts by more than 10%. The same study showed 84% report AI spending cutting gross margins by over six percentage points.

“The AI cost crisis has started,” remarked trader and investor Crypto Rover.

Follow us on X to get the latest news as it happens

Big Tech AI capex hit $650 billion in Q1 2026. The number of FinOps teams managing AI spend doubled from 31% to 63% within a year.

Anthropic stands to benefit even as customers complain. The vendor’s $10.9 billion Q2 forecast would deliver its first profitable quarter. The spend story cuts both ways.

Companies are now layering quotas, internal leaderboards, and cheaper model routing onto deployments that ran open in late 2025.

The next quarter will show whether governance can keep consumption flat. Similar pressures could surface inside crypto AI infrastructure builds in coming months.