Uber Blew Its AI Budget in Four Months: What That Means for OpenAI

Uber's CTO revealed in April that the company burned through its entire annual AI budget in just four months. The disclosure, first reported by The Information, came as Uber implemented spending tiers on AI tools starting at $1,500 per month per employee.

That single fact captures the shift underway. After two years of breakneck token consumption, corporate America is waking up to a hangover. Lindy CEO Flo Crivello moved his entire traffic from Anthropic's Claude to DeepSeek this month. The switch will save Lindy millions of dollars, he said, calling it "a matter of survival." Crivello said Lindy had been dealing with "unsustainable" AI costs.

The math is simple. Anthropic's annualized run rate hit $47 billion in May, up from roughly $10 billion in revenue in 2025. OpenAI's run rate was near $25 billion earlier this year. Both companies filed confidentially for IPOs in early June, hoping to list while growth still obscures the spending fault line.

Analysts see the window closing. D.A. Davidson's Gil Luria said current growth rates are "the fastest they will ever be." The pressure to go public now is real. "There has to be some period of time in the future where there's some rationalizing of spend by companies, and that may be a blip ahead for Anthropic and OpenAI," Luria said. He believes the urgency to list before that blip arrives.

Executives across the economy feel it. Jeff Henry, president of consulting firm Highspring, said many clients are pulling back until they can prove ROI. Others are waiting 12 to 18 months before making big spending decisions. "Everybody is experiencing the same spend crunch on AI," Henry said.

The response from the model providers has been tactical. OpenAI launched analytics and enterprise controls this month, letting administrators set usage limits and credit caps. Anthropic rolled out similar tools in August. Both are trying to keep large customers from defecting.

Cheaper alternatives are flooding in. Microsoft released low-cost models this month and is routing GitHub Copilot tasks to the most appropriate model, not just the most powerful. Google showed Gemini 3.5 Flash at its developer conference, priced at half or one-third the cost of comparable frontier models. Amazon's top AI executive told CNBC the company hopes to compete with frontier models within a year.

PitchBook analyst Harrison Rolfes said Microsoft and Google have the full infrastructure stack to "stiff-arm both OpenAI and Anthropic." He described them as waiting on the sidelines to see where the pure-play model companies struggle.

For now, the big model makers still command 95% of enterprise AI usage, according to Glean CEO Arvind Jain. That share will shrink as model routing and cost discipline spread. AISquared CEO Darren Kimura said the current use of frontier models for simple tasks is "untenable" for most companies in the long run.

The IPO timeline remains uncertain. OpenAI is leaning toward waiting until next year, The New York Times reported. Battery Ventures partner Dharmesh Thakker said traditional venture capital is drying up, making the public markets the last source of new capital for both companies.

Crivello, for his part, said he is open to returning to Anthropic if prices come down. "I hope that they cut the costs again at some point but, until then, we've got options," he said.

MSFT stock page](https://www.alphascala.com/stocks/msft) showed Microsoft shares down 3.46% to $352.83 on the day the article was published, with an Alpha Score of 50 out of 100. The score reflects a mixed outlook as the company balances AI investments with rising competition from its own partners.