AI Spending Shifts: OpenAI, Anthropic Face New Realities

Flo Crivello, the 34-year-old CEO of AI startup Lindy, found his company’s expenses spiraling out of control, necessitating a decisive shift in strategy. Earlier this month, Crivello made the bold move to migrate Lindy’s entire operational load away from Anthropic’s Claude models, redirecting all traffic to DeepSeek, a Chinese company offering more cost-effective, open-weight AI alternatives. This strategic pivot, Crivello explained from his San Francisco headquarters, has already resulted in a dramatic reduction in costs, projecting savings of millions of dollars within months. Despite this significant financial optimization, Crivello anticipates that AI-related expenditures will still outpace payroll for his approximately 25-person team. “It’s a matter of survival for the business,” Crivello stated, underscoring the critical nature of this decision. His prior experience at Uber, where he spent nearly five years, has undoubtedly informed his pragmatic approach to managing high-growth technology ventures.

Crivello is representative of a growing cohort of founders and executives across the United States grappling with the escalating costs of artificial intelligence. Since the unveiling of OpenAI’s ChatGPT in 2022, which ignited a fervent race among businesses to integrate AI into customer support, marketing, and finance, AI expenditures have surged, with some companies reporting astronomical bills. This surge was particularly pronounced in AI-assisted coding, where developers leveraged AI to rapidly create new tools and services, leading to phenomena like “tokenmaxxing” and AI leaderboards that incentivized excessive AI usage without a commensurate focus on efficiency or output quality.

A clear correction is now underway. Uber, for instance, recently implemented monthly spending caps on certain AI tools, starting at a baseline of $1,500, with provisions for employees to request access to higher tiers. This measure follows a revelation by Uber CTO Praveen Neppalli Naga to The Information, detailing how the ride-sharing giant exhausted its entire annual AI budget in just four months, a testament to the unchecked consumption of AI resources.

The unchecked spending spree has disproportionately benefited OpenAI and Anthropic, the two leading AI model developers, fueling their rapid growth and propelling their valuations towards the trillion-dollar mark. As both companies reportedly file confidentially for initial public offerings (IPOs) this June, a palpable shift is occurring in the market’s perception of AI costs. Business leaders like Crivello are increasingly demanding a clear return on investment before committing further capital to these dominant players.

“Current growth rates for Anthropic and OpenAI are the fastest they will ever be, which is mostly a matter of basic math,” explained Gil Luria, a tech equity analyst at D.A. Davidson. “That is a good reason to go public now, as is the concern that some of their largest enterprise customers may start limiting their out-of-control token spend.”

Anthropic recently reported an annualized revenue run rate of $47 billion in May, a significant leap from its approximately $10 billion in revenue for the entirety of last year. OpenAI’s annualized revenue run rate was reportedly nearing $25 billion earlier this year, a substantial increase from the $13.1 billion generated in 2025. This strategic timing for their IPOs, while their financial metrics remain impressive, suggests a proactive move to capitalize on current market conditions before potential cost rationalization by enterprise clients.

“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 elaborated. “That creates some sense of urgency to go public before we see that.”

Anthropic and OpenAI have not responded to requests for comment on this story.

Crivello, while expressing admiration for Anthropic’s capabilities, emphasized the unsustainable nature of his company’s AI costs. Lindy was founded on the premise of rapidly declining token costs, a trend that has slowed significantly as leading model developers like Anthropic and OpenAI have become more conservative with price reductions. Crivello indicated a willingness to revert to Claude models should prices become more competitive, stating, “I hope that they cut the costs again at some point but, until then, we’ve got options.”

The sentiment of cautious spending is echoed by industry consultants. Jeff Henry, president of consulting at Highspring, noted that several of his firm’s clients are either pausing AI adoption until a demonstrable return on investment (ROI) can be established or delaying significant spending decisions for another 12 to 18 months. “Everybody is experiencing the same spend crunch on AI,” Henry observed, while also acknowledging that a considerable number of mid-sized companies have yet to embark on their AI journeys.

Despite the current cost pressures, the pervasive influence of AI is undeniable. “AI is not going away,” Henry affirmed. “There’s no way that toothpaste ever goes back in the tube.”

Darren Kimura, CEO of enterprise AI company AISquared, highlighted a critical area of overspending: the utilization of state-of-the-art “frontier models” for tasks that could be efficiently handled by less sophisticated, and therefore cheaper, alternatives. This inefficiency is exacerbated by the nascent adoption of “model routing,” a technique designed to match specific tasks with the most appropriate AI model. According to Arvind Jain, CEO of Glean, approximately 95% of current enterprise AI usage still relies on these resource-intensive frontier models, a practice Kimura deems “untenable” for long-term sustainability.

Luria of D.A. Davidson believes that the AI pricing landscape remains “unsophisticated,” but both OpenAI and Anthropic are actively adapting to a more budget-conscious market. OpenAI recently introduced enhanced analytics and controls for enterprises, enabling administrators to track spending across departments, set usage limits, and provide employees with budget visibility. Anthropic, in August, rolled out a suite of controls allowing customers to manage users, monitor analytics, and enforce spending limits at both organizational and individual levels.

The financial implications of unchecked AI spending are now a significant concern for corporate finance departments. Eric Glyman, co-CEO of expense management startup Ramp, stated, “Most CFOs not only didn’t plan for this in their annual plans — the steep growth — but don’t have great tools to manage this. Suddenly you have this third pillar that has showed up, which is spending through tokens and intelligence. It’s not a clean area of spend.”

“As companies become increasingly price-sensitive regarding AI, OpenAI and Anthropic face intensifying competition from well-funded rivals developing more economical models. Microsoft, a substantial investor in both OpenAI and Anthropic, recently unveiled a suite of new low-cost AI models, underscoring its commitment to providing accessible AI solutions. The company has also emphasized that its AI coding product, GitHub Copilot, will intelligently route users to the most suitable model for each task, thereby optimizing efficiency and cost.

In a recent essay, Microsoft CEO Satya Nadella articulated a vision for an AI industry that avoids concentrating power within a select few providers. “The last thing any of us want is a world where every company across every sector is ceding value to a few models that eat everything they see,” Nadella wrote. “If all the value is accrued by only a few models, the political economy will simply not tolerate it.” This perspective signals a strategic imperative for Microsoft to foster a more distributed and competitive AI ecosystem.

Amazon and Google are also significantly bolstering their AI investments, with a particular focus on models designed for business users. Peter DeSantis, Amazon’s senior AI executive, expressed optimism that the company will be able to rival OpenAI and Anthropic’s frontier models within the next year. Like Microsoft, Amazon is an investor in both of these leading AI firms. DeSantis previously highlighted that Amazon plans to leverage its in-house chip development capabilities to create AI models at a more competitive price point than its rivals. “AI has a cost problem,” he told The Wall Street Journal. “If we ultimately want AI to transform everything, the costs have to be different.”

Google made a concerted effort at its recent developer conference to showcase its affordable AI offerings. The company introduced Gemini 3.5 Flash, a more lightweight version of its model suite, available at a significantly lower cost compared to comparable frontier models. CEO Sundar Pichai noted that this model is priced at half, and in some cases, close to one-third, of the cost of its competitors.

“Microsoft and Google possess the requisite infrastructure and comprehensive capabilities to effectively compete with both OpenAI and Anthropic,” observed PitchBook analyst Harrison Rolfes. “They are likely observing the landscape, waiting for the leading companies to navigate their challenges before strategically entering the market where they perceive weaknesses.”

Regarding their IPO timelines, neither OpenAI nor Anthropic have provided definitive dates. Reports suggest OpenAI is leaning towards a 2027 debut, while the urgency for capital may drive both companies to the public markets sooner rather than later. As competition intensifies and their capital requirements grow beyond the capacity of most venture and private equity firms, an IPO appears to be the most viable avenue for securing substantial funding.

“A lot of the traditional pockets of capital are drying up,” noted Dharmesh Thakker, a general partner at Battery Ventures. “All the institutional investors who can invest in these companies have already taken their pound of flesh.” This indicates a maturing investment landscape where established players are being approached for significant, perhaps final, rounds of funding before potential public market debuts.