Major companies including Uber, Walmart, Amazon and Meta are starting to impose usage caps and monitoring on the generative AI tools they handed to employees, as the cost of scaling the technology strains corporate budgets.
Enterprise AI · The Reckoning
Big Companies Start Rationing AI as Token Costs Explode
After rushing company-wide rollouts, Uber, Microsoft, Meta and others are now imposing usage limits and tying AI spend to real business outcomes — as runaway "token" consumption blows through annual budgets in months.
3.2Q
tokens Google processes per month — 7× year over year
18%
of AI coding-tool spend that ties to code actually shipped
2,000+
companies analyzed on AI cost vs. results
Cost per pull request: light vs. heavy AI use
Same output, wildly different bill — heavy AI use costs ≈319× more per merged change.
Where it pays off
7+ months → under 2 weeks — Anthropic cites a complex task collapsed in time.
Real productivity gains reported on hard, well-scoped problems.
Where it backfires
AI costs exceeding human labor; heavy debug/rewrite burden on generated code.
Rehiring after quality drops: Klarna, Commonwealth Bank (+45 staff), Duolingo.
From land-grab to governance
Unlimited rollout
2023–25: avoid falling behind
→
Cost shock
budgets gone in months
→
Limits & ROI tracking
outcome-based controls
Meta orders a duplicate-tool review; Salesforce tracks tokens against outcomes; OpenAI rolls out enterprise cost controls — usage-based billing is the new battleground.
Continue reading The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in ✓ Signed in — this article isn’t included in your current plan.Unlocking the full article…