Firing employees due to AI implementation turns out to be economically impractical, as it costs more.

Short overview of the situation

Modern corporations are increasingly forcing their employees to use AI to boost work efficiency. However, expenses for these technologies are growing faster than their economic return, and many companies are already facing financial difficulties.

1. Microsoft: transition from Claude Code to GitHub Copilot CLI
- What happened

- Over six months Microsoft integrated the Anthropic Claude Code AI programmer into its projects, encouraging developers to experiment with it.

- Now the company is canceling most of the working licenses for Claude Code and moving engineers to its own tool, GitHub Copilot CLI.

- Why it matters

- Despite dropping the licenses, Microsoft continues to invest in Anthropic: $5 billion in investments and a commitment to purchase Azure compute capacity worth $30 billion.

2. Uber: budget “stream” of AI
- Situation

- Over four months Uber engineers exhausted the entire allocated AI budget for 2026.

- Earlier, management encouraged AI adoption, even creating department rankings based on technology usage.

- Takeaways

- The cost of AI implementation is becoming a serious barrier: Nvidia deep‑learning VP Brian Katanzaro noted that compute expenses exceed employee salaries.

3. Other major players
Company Incentive Measure Problem Meta ✴ “Clodonomics” leaderboard (named after Anthropic Claude) Token price rise with consumption-based billing Amazon Recommendation “maximise AI token usage” Increased compute costs

4. Forecasts and analysis
- Goldman Sachs

- By 2030, AI‑token consumption could grow 24×, reaching 120 quadrillion tokens per month.

- Even if the price per token falls, total AI spend will rise.

- Gartner

- Deploying trillion‑parameter models by 2030 may become 90 % cheaper than in 2025.

- However, AI‑agent‑driven models consume more tokens than standard solutions, so cost reductions do not always offset the higher usage.

5. Business implications
- Financial risk – companies may face unexpected compute costs if token consumption exceeds expected price drops.

- AI strategy planning – leaders must revisit budgets and monetisation models to avoid “wasted” AI investments.

6. Conclusion
AI adoption remains a priority for large tech giants, but rising compute and token expenses create new challenges. Companies are forced to balance the desire to boost employee productivity with real financial constraints.