
Companies aren’t reducing artificial intelligence spending overall. They’re reining in waste. Global AI spending is still on track to grow sharply in 2026, but a growing number of companies, including Uber and Amazon, are putting caps on runaway AI bills after costs came in far higher than planned.
At the same time, many employers are cutting jobs and pay to free up money for AI. So the real story isn’t “AI spending is falling.” It’s that companies are getting pickier about where AI money goes, and workers are often the ones footing the bill. Here’s what’s actually being cut and what it means for you.
Is overall AI spending going down?
No. Research firm Gartner forecast in May that worldwide AI spending will hit $2.59 trillion in 2026, up 47% from 2025. It has raised its forecast more than once this year.
Most of that money comes from the biggest tech companies. Alphabet, Amazon, Meta and Microsoft have committed roughly $750 billion in combined 2026 capital spending, mostly on AI data centers and chips.
Gartner says AI spending so far has been driven mainly by tech companies and cloud providers, and that regular businesses haven’t fully opened their wallets yet. When those businesses do spend, Gartner expects them to favor smaller, practical projects over big bets.
Related: Claude Duet: What Anthropic’s New AI Agent Does, What It Costs and Who Can Get It · Gemini AI Hack: How Google’s AI Breached Three Real Companies
So why are companies cutting AI budgets?
Because the bills surprised them. Many AI tools now charge by usage, measured in tokens (small chunks of text the AI reads and writes). The more employees use AI, the bigger the bill, and some companies encouraged heavy use without a ceiling.
The numbers show how common this has become:
- A Futurum survey of 1,636 tech decision makers found 46.9% of enterprises are running over budget on AI. Just 5.6% are spending below plan.
- In Pigment’s third-quarter CFO survey, 83% of finance chiefs said usage-based AI costs came in higher than expected.
- A Mavvrik report found 25% of firms delayed or canceled AI projects because of unexpected costs.
Which companies are cutting back on AI?
Uber is the best-known example. Its chief technology officer said in April that the company had already used up its entire 2026 AI budget, four months into the year, driven largely by AI coding tools. Uber’s COO later questioned whether the spending was paying off, saying it’s hard to draw a line from AI usage to new features for customers.
Here’s a snapshot of the most visible pullbacks:
| Company | What happened | What they changed |
|---|---|---|
| Uber | Used its full 2026 AI budget by April | Capped AI coding tool spending at $1,500 per employee, per tool, each month |
| Amazon | Staff ran AI on pointless tasks to climb an internal usage leaderboard | Shut down the leaderboard; an executive told staff not to use AI just for the sake of it |
| Cisco | Rising AI tool costs | Moved to curb non-essential AI tools |
Notice what these companies did. None of them stopped using AI. They put limits on it and started asking for proof that it’s worth the money.
How much is Uber’s AI cap worth?
Uber’s cap shows how fast these costs add up. Reports put Uber’s AI coding costs at $500 to $2,000 per engineer each month before the cap, across about 5,000 engineers.
Here’s the math on the new $1,500 monthly cap for a single tool:
| Scale | Maximum spend per month | Maximum spend per year |
|---|---|---|
| One engineer, one tool | $1,500 | $18,000 |
| 5,000 engineers, one tool | $7.5 million | $90 million |
That ceiling only applies if every engineer hits the limit, which most won’t. But it shows why finance teams got nervous. One AI tool used heavily across a large team can cost as much as hiring hundreds of extra employees.
Are companies cutting jobs to pay for AI?
Many are, at least by their own account. AI was the top reason US employers gave for job cuts for much of 2026, according to outplacement firm Challenger, Gray & Christmas. Through August, AI had been cited in 116,175 announced job cuts, about 22% of all layoffs this year.
A ResumeBuilder survey of 866 US business leaders found 54% of companies had cut or planned to cut employee pay to free up money for AI in 2026. The survey was commissioned by ResumeBuilder, so treat it as a signal rather than a precise measure.
Some experts think the AI label is being stretched. Research firm Forrester has warned about “AI washing,” meaning companies blame cuts on AI when they’re really about saving money. Glassdoor economist Daniel Zhao has made a similar point: a company saying AI caused layoffs doesn’t prove that’s the real reason.
There are early signs the trend may be shifting. In August, AI dropped to the fourth most-cited reason for job cuts, with 3,462 cuts, its lowest monthly total since December 2025.
Is AI saving companies money yet?
For most companies, not yet. In PwC’s January 2026 survey of 4,454 CEOs, 26% said AI had lowered their costs over the past year, but 22% said it had raised them. Deloitte’s 2026 survey found 40% of leaders had cut costs with AI so far.
My take: 2026 is the year companies stopped treating AI as a blank check. The money is still flowing, but it’s being moved from “use AI everywhere” toward “use AI where it clearly pays.” That’s healthy for businesses. For workers, it means AI skills matter more, but so does showing that your work produces results AI can’t easily replace.
If you’re worried about what a pullback in AI spending could mean for the stock market, our guide to Warren Buffett’s advice for a market downturn covers how to stay steady. And for a look at where consumer AI assistants are heading, see our post on Meta’s Muse AI.
What to Watch Next
- Gartner IT Symposium/Xpo in Orlando (October 19 to 22, 2026): Gartner often updates its views on enterprise AI budgets and returns at this event.
- Big Tech third-quarter earnings (expected late October 2026): Watch whether Alphabet, Amazon, Meta and Microsoft raise or trim their AI spending plans for 2027.
- Challenger October job cuts report (expected early November 2026): This will show whether AI-linked layoffs bounced back after August’s dip.
- 2027 budget season (November to December 2026): Many companies set next year’s AI budgets now. Expect more usage caps and stricter return-on-investment rules.


