OpenAI Weighs Drastic Token Price Cuts to Blunt Anthropic's Enterprise Run

OpenAI is weighing significant cuts to what it charges for tokens, the metered unit AI companies use to bill for their products, in anticipation of similar cuts it expects at Anthropic, the Wall Street Journal reported June 10, citing people familiar with the matter. The discussions are still in flux, the Journal said. They follow OpenAI's confidential IPO filing on June 8, one week after Anthropic's.

A price cut from a market leader usually signals strength; this one follows a quarter in which OpenAI lost both the valuation lead and the revenue-growth lead. The contemplated cut is a defensive repricing of an enterprise position OpenAI has already lost on revenue growth and valuation, and because both companies entered confidential SEC review within the same ten days, whatever pricing moves follow will surface in the prospectuses both companies must eventually publish.

Sam Altman acknowledged the pressure at a recent company event, where he said AI costs had become "a huge issue" for business executives. "I think we'll have a lot of ways we can help people get more value for less spend," he said, per the Journal's account.

The $113 billion gap behind the discussions

Anthropic closed its Series H round on May 28 at a $965 billion valuation. OpenAI's last mark, set in March, was $852 billion. That $113 billion gap, the first time the younger company has priced above its rival, is the backdrop the Journal's sources describe: OpenAI trying to catch up with Anthropic in the race for enterprise customers after Claude Code went viral among software engineers.

The revenue trajectory explains the valuation flip better than any single product launch. Anthropic put its annualized run rate at roughly $30 billion in April, a figure CEO Dario Amodei said had outrun the company's own forecasts by a factor of eight; the run rate stood near $1 billion at the start of 2025, and analyst estimates cited in syndicated coverage put it near $47 billion by May. OpenAI contests the comparison. Chief revenue officer Denise Dresser told employees in April that the company considers Anthropic's financials inflated by its accounting method, according to a company memo reviewed by Reuters.

"Approximately a year ago, OpenAI's strategy was swing for the fences, whereas Anthropic's strategy is make money first," Jenny Xiao, a partner at Leonis Capital and a former OpenAI researcher, told the Financial Times. "Now the two are converging, because both of them are trying to aim for an IPO and investors care more about money than dreams."

The $8 consumer tier and the enterprise meter

The cut under discussion targets tokens, the metered unit that prices API and enterprise usage, rather than consumer subscriptions. OpenAI already charges consumers $8, $20, and $100 and up per month across its GPT-5.5 tiers, CNBC noted, while Anthropic's Claude Pro runs $17 a month on an annual plan, so OpenAI's entry tier already undercuts its rival's. The contest the Journal describes is over the enterprise meter, where corporate AI budgets are tracked in tokens consumed.

Altman frames the discussion as customer relief, and the Journal's reporting carries the harder number next to that framing: both companies lose billions of dollars on the computing costs of serving queries. A lower token price under those economics bets that volume growth will outrun the loss per token. The Journal's sources did not name cut amounts, affected models, or timing, and OpenAI did not immediately respond to CNBC's requests for comment.

Uber's exhausted budget and a 20 percent drop in token use

An Uber executive said earlier this year that the company had maxed out its 2026 budget for agentic AI, and another corporate leader said last month it was hard to tie AI coding gains to features customers could see, the Journal reported. Those complaints have fed a Silicon Valley debate over "tokenmaxxing," the practice of burning tokens for productivity gains that do not return the spend.

Some AI tools are seeing token consumption drop by 20 to 30 percent as enterprises restructure their budgets, Crypto Briefing reported in its analysis of the pricing discussions. A price cut recovers revenue only if customers consume more at the lower rate, and the corporate examples in the Journal's reporting, Uber among them, are working to consume less.

Google's Gemini models, the budget Flash tiers in particular, undercut both ChatGPT and Claude on price, and its business plans cost nearly half of what OpenAI charges, Digital Trends noted in its read of the price-cut report.

Two confidential S-1s and a "within the next year" clock

The near-term sequence is partly on the record. The Journal's sources say OpenAI expects Anthropic to cut, which makes bilateral repricing the scenario OpenAI itself is planning around; Anthropic has announced nothing, and Reuters said it could not immediately verify the price-cut report. The deeper risk is one investors have flagged for years, in the Journal's telling: the two companies' products are interchangeable enough that customers can abandon one for the other with little friction. A price war tests that weakness in public, quarter by quarter, once the companies start reporting.

OpenAI is preparing the biggest overhaul of ChatGPT since launch, recasting the chatbot as a gateway to coding tools and agents, the Financial Times reported, citing more than a dozen current and former employees. "Chat is dead," one senior OpenAI employee told the FT. Codex has grown sixfold since its February desktop launch, to more than 5 million weekly users, and the 2 million businesses on OpenAI's products account for about 40 percent of revenue, a share the company expects to reach 50 percent by year-end. Cheaper tokens make more sense as the on-ramp to that mix than as the product itself.

OpenAI said in its filing statement there were "things we want to do that are likely easier as a private company," then declined to elaborate. Altman has told employees the company plans to go public "within the next year," in a Slack message earlier reported by The Information. The first public S-1, whichever company converts its confidential draft first, will show what a token price war does to gross margin. OpenAI's pricing deliberations, in the Journal's account, already assume that war is coming.

SpaceX Has Starlink Revenue. The Prospectus Makes xAI the Bill.

Bobby Peden took office in Starbase last May, after roughly 500 residents voted to make the Boca Chica site a city. SpaceX disclosed its public SEC prospectus on Wednesday with a ticker, SPCX, and the

The Implicator

Claude Needs Distribution. Wall Street Already Owns It

On Sunday evening, Lauren Thomas and Berber Jin put a price on Anthropic's private-equity problem. The Wall Street Journal reporters said the Claude maker was near a $1.5 billion joint venture with Bl

The Implicator

Anthropic shifts enterprise billing to per-token pricing. The flat-fee era is over.

Anthropic has restructured its enterprise plan to bill Claude, Claude Code, and Cowork usage separately from seat fees, moving its largest business customers to per-token pricing at standard API rates

The Implicator

Marcus Schuler

San Francisco

Editor-in-Chief and founder of Implicator.ai. Former ARD correspondent and senior broadcast journalist with 10+ years covering tech. Writes daily briefings on policy and market developments. Based in San Francisco. E-mail: editor@implicator.ai