The AI model race just turned into a price war — and OpenAI is playing to win. According to a report by The Information, OpenAI is offering steep discounts to pull some of Anthropic's biggest customers across, and it's working. Replit moved more than $100 million a year in AI spending, Rippling flipped most of its coding workload, and DataCamp says its costs fell by as much as 80 to 90 percent. None of the companies have publicly confirmed the exact financial terms, so treat the numbers as reported, not official — but the direction of travel is hard to miss.
The $100 million switch
Replit, the coding platform, shifted most of its AI spending from Anthropic to OpenAI after OpenAI offered a discount of more than 30 percent on a contract worth over $100 million a year, The Information reported. Anthropic had reportedly declined to extend Replit's previous discount without a higher spending commitment. Replit CEO Amjad Masad later softened the headline, saying OpenAI is now Replit Agent's preferred partner based on performance, economics and partnership — and seven
Claude models remain listed in Replit's catalog for customers who choose them. Still, the default changed, and defaults decide where most usage goes.
Rippling's flip tells the same story
At Rippling, the workforce software company, OpenAI's share of engineers' coding work jumped from 16 percent in May to 52 percent, while Anthropic's fell from 57 percent to 18 percent, according to CEO Parker Conrad. That's a swing of more than 30 points in a few months, inside a product where model quality is supposed to matter most. DataCamp's CEO said OpenAI's pricing cut his company's costs by 80 to 90 percent. One US bank reportedly doubled its ChatGPT seats to 500 employees at no extra charge, paying under $7 per user against the usual $20-plus rate.
Why price is winning now
This is the part that should worry every AI lab. Coding is the use case where model differences are meant to be most visible — yet Replit's reported trigger was a withdrawn discount, not a benchmark defeat. That suggests that, wrapped in good tooling, retries and scaffolding, today's frontier models are close enough for many businesses that a 30 percent price gap outweighs any capability gap. OpenAI has also cut its own costs with its more efficient Sol and Luna model lines, giving it room to discount. Meanwhile an investor memo cited in the same reporting says Anthropic's growth slowed from the second quarter to the third — awkward timing, with the company reportedly nearing a public listing.
What it means for everyone else
For users, a price war is good news in the short term: cheaper AI, better deals, and vendors fighting for your business. For the labs, it's a squeeze — margins thin out, and the moat around any single model gets thinner as customers learn they can switch. The open question is whether OpenAI can keep discounting without burning through its lead, and whether Anthropic answers with its own price cuts or with features customers can't get elsewhere. Either way, the AI market just got a lot more competitive — and the winner may be decided by the invoice, not the
benchmark.
Comments
Post a Comment