OpenAI, Anthropic new AI spending reality as users shift to efficiency

Flo Crivello’s expenses were running wild and there was only one way to get them under control.

Earlier this month, the 34-year-old CEO of artificial intelligence startup Lindy changed The company is abandoning Anthropic’s Claude models and moving 100% of its traffic to DeepSeek, a Chinese company that produces cheaper, exploit-heavy alternatives.

“We did it, and you could see the cost curve coming down, just crashing into the ground,” Crivello said in an interview from his company’s San Francisco headquarters. He said the decision will save Lindy millions of dollars over months, but he still expects the roughly 25-person company to spend more on AI than payroll.

“This is a matter of survival for the business,” Crivello said. “This is everything.”

Crivello, who previously spent almost five years here UberHe is among a growing number of founders and executives across the U.S. trying to rein in AI spending. Since OpenAI first wowed Wall Street with its ChatGPT chatbot in 2022, billings for AI have sometimes reached billions of dollars, setting off a rush for businesses to deploy the technology in areas like customer support, marketing and finance.

Costs have soared, especially in the AI-powered coding space, as developers pump tokens into the creation of new tools and services that previously required teams of coders. This has led to the era of tokenmaxxing and AI leaderboards, where employers encourage developers to use as much AI as possible without worrying about the consequences.

The raid continues. Uber said this month it was implementing a series of spending tiers for some AI tools, starting at $1,500 per month, but employees can request access to higher tiers. in April, Uber CTO Praveen Neppalli Naga clarified That the ridesharing company spent its entire annual AI budget in just four months.

OpenAI and Anthropic have been major beneficiaries of the spend-at-all-costs mentality that has fueled exponential growth rates and propelled both AI model leaders to valuations approaching $1 trillion.

Now, as they gear up for potentially historic IPOs (both filed privately in early June), the mood around AI is shifting, and business leaders like Crivello are no longer willing to throw money at Anthropic or OpenAI without a clear picture of the return on their investments.

“The current growth rates for Anthropic and OpenAI are the fastest they’ve ever been, and it’s mostly a matter of basic math,” Gil Luria, an equity analyst who covers tech companies at DA Davidson, told CNBC. “This is a good reason to go public now, as is concern that some of its largest institutional customers may begin to limit out-of-control token spending.”

Anthropic last reported an annual sales run rate of $47 billion in May, down from the nearly $10 billion in revenue it recorded in all of last year. OpenAI’s run rate was approaching $25 billion earlier this year. reportsThat’s an increase from its $13.1 billion in revenue in 2025.

While the numbers are still eye-popping, listing soon may be strategic.

“There needs to be a period in the future when it becomes possible for companies to rationalize their spending, and this could be a very important step for Anthropic and OpenAI.” Luria said in an interview. “This creates some sense of urgency to get this out to the public before we even see it.”

Antropik declined to comment for this story. OpenAI did not respond to a request for comment.

Crivello said he’s a big fan of Anthropic, but his company has long struggled with “unsustainable” AI costs.

Crivello said Lindy was built on the idea that the cost of tokens, or units of data processed and produced by AI models, will decrease significantly over time. This proved true for a while, but leading model developers, including Anthropic and OpenAI, have been slower to reduce prices in recent months.

Crivello said he would be open to converting Lindy back to Claude models if prices drop.

“Hopefully at some point they will reduce costs again, but until then we have options,” he said.

Jeff Henry, president of consulting at Highspring, said some of his firm’s clients are holding back until they “really start seeing a return on investment,” while others are still waiting another 12 to 18 months before making big spending decisions.

“Everyone is experiencing the same spending crisis for AI,” he said.

But he said there are countless mid-sized companies that haven’t even started experimenting with AI yet.

“AI is not going away,” Henry said. “There’s no way for the toothpaste to go back into the tube.”

Darren Kimura, CEO of enterprise AI company AISquared, said one area where AI spending is “absolutely” peaking is in the use of cutting-edge models. Frontier models for simple tasks that can be accomplished with cheaper alternatives.

OpenAI launched Analytics and updated controls for businesses released earlier this month allow managers to separate workplace credit spend, set usage limits and give employees visibility into their available budgets. Anthropic appeared a set of controls In August, it will allow customers to provision users at the organization and individual level, view analytics, and set spending limits.

Eric Glyman, co-CEO of expense management startup Ramp, said he’s paying close attention after finance departments were faced with surprisingly large AI bills.

“Most CFOs not only don’t plan for this (steep growth) in their annual plans, but they don’t have great tools to manage it,” Glyman said in an interview. “Suddenly the third pillar has emerged; spending through tokens and intelligence. This is not a clean area of ​​spending.”

As companies become more price-sensitive when it comes to AI, OpenAI and Anthropic are forced to compete with wealthy rivals seeking to develop lower-cost models.

Microsoft spilled more $13 billion for OpenAI and $5 billion for Anthropic, which introduced a number of new low-cost models earlier this month. The company also emphasized that the artificial intelligence coding product GitHub Copilot will be released. redirect users Finding the model best suited to a task.

One June articleMicrosoft CEO Satya Nadella said the industry should avoid concentrating power in a handful of large providers.

“The last thing any of us wants is a world where every company in every industry values ​​a few models who eat everything in sight,” Nadella wrote. “If all the value comes from just a few models, the political economy simply won’t tolerate that.”

Amazon And Google They are also increasing their investments in models aimed at business users.

Amazon’s top AI executive, Peter DeSantis, told CNBC this month that he hopes the company will be able to compete with leading models from OpenAI and Anthropic “next year.” Like Microsoft, Amazon is an investor in both these companies.

DeSantis said in February that Amazon would rely on its in-house chips to develop cheaper models than rivals.

“Artificial intelligence has a cost problem,” he said Wall StreetJournal in an interview. “Ultimately, if we want AI to transform everything, the costs need to be different.”

Google We made a concerted effort to highlight affordable AI offerings at our annual developer conference last month. The company showcased the Gemini 3.5 Flash, a lighter addition to its lineup that’s available at half, or in some cases close to a third, the price of similar flagship models, according to CEO Sundar Pichai.

“Microsoft and Google have the infrastructure and the capability (the entire stack) where they can intervene on both OpenAI and Anthropic,” PitchBook analyst Harrison Rolfes said in an interview. “They’re probably waiting on the sidelines to challenge, to see where they’re not good.”

When it comes to going public, none of the major model companies have offered a definitive time frame for their possible exits. New York Times reported OpenAI is inclined to delay until next year, people involved in the talks said Thursday.

The pressure to go public may revolve around the need for capital. With Anthropic and OpenAI increasingly competing with their biggest financial backers, the IPO market may be the best route for new money; especially as capital needs have become overwhelming for most startups and private equity firms.

“A lot of the traditional pockets of capital are drying up,” said Dharmesh Thakker, general partner at Battery Ventures. “All institutional investors who can invest in these companies have already received their share.”

WRISTWATCH: Anthropic accuses Alibaba of campaign to acquire AI talent