OpenAI Weighs Steep Price Cuts as Anthropic Pulls Ahead in Enterprise AI

OpenAI is weighing significant cuts to its API token pricing as competition from Anthropic heats up, according to an exclusive Wall Street Journal report. The discussions come as both companies have filed confidentially for IPOs this month and neither has turned a profit, putting margin pressure front and center.

"I think we'll have a lot of ways we can help people get more value for less spend," OpenAI CEO Sam Altman said at a recent event,,¹ which cited the WSJ report. The internal discussions are described as still in flux, but the direction is clear: OpenAI is preparing to compete on price.

The pricing review isn't happening in a vacuum. Anthropic has executed one of the fastest enterprise land grabs in tech history, driven almost entirely by Claude Code, its autonomous coding agent.

Decrypt reports that Anthropic's annualized run rate surged from $9 billion at the end of 2025 to $47 billion by May 2026 — a 422% jump in five months. For the first time, more companies tracked by the Ramp AI Index are paying for Anthropic than for OpenAI.

Q2 2026 marks Anthropic's first profitable quarter, a milestone OpenAI has yet to reach. The company posted a -122% adjusted operating margin in Q1 2026, meaning it lost $1.22 for every dollar of revenue it brought in.

The market share numbers tell the same story. ChatGPT's slice of global generative AI web traffic fell from 77.6% in May 2025 to 53.7% by April 2026,.¹ While still dominant, the trend line is unmistakable — and it's accelerating.

OpenAI has since made Codex, its own coding tool, a company priority. But as Decrypt notes: "It's playing catch up." The company that defined the current AI era with ChatGPT is now chasing Anthropic on the product that drives the most enterprise revenue.

While OpenAI and Anthropic debate how much to charge, open‑source inference providers are already serving Chinese models like DeepSeek V4, GLM, and Kimi at roughly one‑thirteenth the cost of closed‑model alternatives — with comparable performance on coding benchmarks.

"Chinese labs open source frontier‑grade models," Delphi Ventures' Tommy Shaughnessy.¹ "The model is the single biggest cost an inference provider has, and they get it for free."

As long as China's AI labs stay open‑source — and so far, most appear committed to that approach — the floor on intelligence pricing keeps falling toward zero. Any margin recovery at OpenAI or Anthropic becomes "a math problem with no clean solution," Decrypt noted.

The enterprise AI spending frenzy — nicknamed "tokenmaxxing" in Silicon Valley — is fueling both demand and anxiety. Uber's CTO burned through the company's entire 2026 AI budget by April. Some JP Morgan employees are spending more on AI tools than their own salary,.¹

JP Morgan analysts published a note this month titled "AI Bills Are Out of Control." Palantir CEO Alex Karp compared the spending pattern to a ¹ last week — an unusually blunt assessment from someone whose company sells AI software.

The tension is structural. The $20/month flat‑fee consumer plans were always priced below what heavy usage actually costs — loss leaders designed to drive adoption. But once a real business needs AI at scale, it moves to metered API pricing and burns through credits far faster than flat fees ever suggested.

Both OpenAI and Anthropic filed confidentially for IPOs this month, per the.² The SpaceX IPO — which saw shares surge 23% on their first day of trading to a $2 trillion valuation — has set an impossibly high bar and created a template that both AI companies are expected to follow.

But SpaceX has something OpenAI and Anthropic don't: profits. Aggressive discounting could pressure already‑thin margins right when public market investors will be scrutinizing unit economics. It's a classic pre‑IPO dilemma: grow market share now at the expense of profitability, or protect margins and risk ceding ground to competitors.

For developers and builders integrating AI into their products, the coming price war is a net positive — at least in the short term:

• Cheaper API calls are coming. Whether through official price cuts or competitive pressure, the cost of serving AI features is likely heading down across all major providers.
• Multi‑model strategies make more sense than ever. With DeepSeek V4 offering comparable coding performance at a fraction of the price, and Anthropic and OpenAI both fighting for enterprise loyalty, routing different types of prompts to different models based on cost and capability is no longer optional — it's table stakes.
• Lock‑in risk is real. As Fortune reported from Brainstorm Tech, companies that drop AI into existing processes without redesigning workflows end up with inefficiencies that "get weaponized at scale" by AI agents. The builders who abstract their model layer now will have options; the ones who hard‑code to one provider will be stuck negotiating renewal pricing.
• Tokenmaxxing isn't sustainable. If your AI spend is growing faster than your revenue, you're not alone — but the JP Morgan note is a warning shot. Builders who can demonstrate clear ROI on their AI usage will survive a pricing shakeout. Those burning tokens for the sake of it won't.