AI-POWERED NEWS

30+ sources. Zero spin.

Cross-referenced, unbiased news. Both sides of every story.

← Back to headlines

Corporate AI Budgets Are Blowing Up: Uber Burned Its Entire 2026 AI Budget by April

Corporate AI Budgets Are Blowing Up: Uber Burned Its Entire 2026 AI Budget by April
The AI spending hangover is here. Companies that gorged on AI tools in early 2025 are now drowning in token bills they never saw coming — and the industry is scrambling to build the cost controls that should have existed from day one. This was entirely predictable, and nobody in the hype machine wanted to say it.

The Bill Arrived. Nobody Budgeted for It.

Since we covered Anthropic's confidential IPO filing last week — with its $47 billion annualized revenue figure — the other side of that ledger has become impossible to ignore. Someone is paying those bills. Turns out, those someones are now in full panic mode.

Uber burned through its entire 2026 AI coding budget by April. That's eight months of budget gone in four. Microsoft revoked its developers' Claude Code licenses after only a few months of use. According to TechCrunch, a Priceline employee reported that a routine Cursor contract renewal came back 4-5x more expensive than before.

These aren't small companies. These are Fortune 500 firms with actual finance departments. And they still got blindsided.

How This Happened

The mechanics are straightforward. Per-token prices have actually fallen over the past year. The catch: AI adoption exploded, agentic tools multiplied, and consumption went through the roof. CEOs were demanding their teams use the best models and move fast — cost controls be damned.

New models released last November — Anthropic's Claude Opus 4.5, OpenAI's GPT-5.1, and Google's Gemini 3 Pro — brought major jumps in agentic capability, according to TechCrunch. Agents don't just answer questions. They loop, they call tools, they generate token after token in the background. Nobody's watching the meter while they do it.

The result? One unnamed company reportedly racked up a $500 million Claude bill after failing to set usage limits for employees. Five hundred million dollars. Because nobody remembered to flip a switch.

The Industry Response: Build the Controls Now

J.R. Storment, executive director of the FinOps Foundation under the Linux Foundation, told TechCrunch: "In April and May, I started hearing from companies: 'Oh my god, we are 3x over our entire 2026 token budget and it's only April.'"

The Linux Foundation this week announced the Tokenomics Foundation — a new standards body designed to bring the same cost discipline to AI tokens that FinOps brought to cloud spending. The comparison to cloud is intentional. Cloud costs also spun out of control before the industry developed frameworks to track and govern them. AI is running the same script, just faster.

Alexander Embiricos, OpenAI's head of enterprise, confirmed the shift in client conversations at an event in New York City this week. Six months ago, every conversation was about capability — what can it do? Now, he told TechCrunch, "Our conversations are never about that now. Now the conversations are about, hey, we're spending so much. What visibility do you have? What auditability do you have? What token controls do you have?"

Chris Reed, senior director of IT finance at Priceline, had the most colorful take. He compared AI vendor pricing to crack cocaine — let you try it for free, get you hooked, then charge what they want. Priceline has already started placing token limits on certain employee groups.

What Mainstream Coverage Is Missing

Most tech media is treating this as a growing-pains story — a speed bump on the road to an inevitable AI future. That framing lets vendors off the hook.

The reality is that AI providers have strong financial incentives NOT to build cost controls into their products upfront. Every token burned is revenue. The $500 million mystery Claude bill didn't happen because Anthropic forgot to add usage alerts. It happened in an environment where the default is wide-open consumption.

Meanwhile, CNBC and others are still pitching the AI trade as the next big wave for investors — pointing to Taiwan and South Korea ETFs that have surged 67% and 109% respectively this year, per Goldman Sachs strategist Tim Urbanowicz. He's not wrong that there's runway left in the chip supply chain. But the emerging crack in the foundation is corporate demand discipline. If Fortune 500 companies start pulling back AI spend to plug budget holes, those chip multiples get stress-tested fast.

The ROI Question Nobody Wants to Answer

The Tokenomics Foundation and FinOps frameworks are showing up after the money is already gone. The discipline infrastructure is being built in response to a crisis, not ahead of one.

And the ROI case for all this spending? Still largely unproven. Anthropic just filed for an IPO on $47 billion in annualized revenue. The companies paying that revenue are reportedly in existential budget crises. Somebody's business model is working great. It's not necessarily the ones buying the tokens.

The AI gold rush analogy has always been that the pick-and-shovel sellers — chipmakers, model providers — make money regardless of whether anyone strikes gold. What we're watching right now is companies discovering they bought a lot of shovels and dug a lot of holes. Whether there's gold at the bottom is still an open question.

Sources

center-left TechCrunch The token bill comes due: Inside the industry scramble to manage AI’s runaway costs
center-left Bloomberg Morgan Stanley Sees AI-Related Funding Expanding to 15% of All Credit Deals
center-left Bloomberg GoldenTree's Tananbaum Spots a Credit-Equity Disconnect
center-left Bloomberg Credit Titans Warn of Shakeout on Debt Deals That ‘Don’t Make Sense’
center-left Bloomberg US Prosecutors are Scrutinizing Marks in Private Credit, Clayton Says
center-left Bloomberg `There Is No AI Bubble,' Says BI's Rob Schiffman
center-left CNBC Where investors may find the next 'big wave' for AI trade
center-left bloomberg AI Infrastructure Spending Sparks Concerns Over Corporate Debt Levels
unknown ft The High Cost of AI: Why Investors Are Demanding Better Returns