OpenAI discounts target Anthropic’s enterprise AI customers

OpenAI is reportedly offering unusually large discounts to win more business from Anthropic’s biggest customers. According to The Decoder, citing reporting from The Information, the strategy is already moving spending at companies including Rippling, Replit and DataCamp.

For teams choosing between OpenAI and Anthropic, the report is a reminder that enterprise AI pricing is becoming more negotiable, especially when usage is large enough to matter. The practical question is no longer just which model performs best, but how much leverage buyers have when they can credibly shift workloads between providers.

What OpenAI is reportedly offering

The reported discounts vary by customer, but they are significant. The Decoder says Replit moved after OpenAI offered more than 30% off a contract valued at more than $100 million a year. DataCamp CEO Jonathan Cornelissen said OpenAI’s pricing reduced his company’s costs by 80% to 90%.

The report also describes a US bank with more than $7 billion in assets that received twice as many ChatGPT seats, expanding to 500 employees, without paying more. According to the report, the bank paid under $7 per user rather than the usual $20 or more.

OpenAI appears to be using two levers at once: lower commercial pricing and, in some cases, more seats for the same outlay. For buyers, that distinction matters. A lower per-seat price, discounted API usage, and added seats can look similar in a proposal but affect budgets and adoption plans differently.

Anthropic workloads are in play

The clearest example in the report is Rippling. CEO Parker Conrad said the company raised OpenAI’s share of its coding workload from 16% to 52%, while Anthropic’s share dropped from 57% to 18%.

That kind of shift is important because many AI buyers now split work across more than one provider. A team may use one model family for coding, another for writing, and another for support or internal search. If pricing changes sharply, workloads that once seemed locked to one vendor can move.

For readers comparing OpenAI with Anthropic’s Claude, the takeaway is not that one provider has permanently won a category. It is that procurement has become part of the product decision. If a company can test the same workflow across multiple models, it may also be able to negotiate harder.

Replit is a notable case because it is both a developer platform and a company whose own AI usage can be substantial. The Decoder reports that Replit switched after OpenAI’s discounted offer. Emonarc previously covered related enterprise movement in Replit CEO says the company is no longer working with Anthropic.

Why prices are under pressure

The Decoder reports that OpenAI is partly accepting weaker margins in order to gain share. But the outlet also says OpenAI has achieved real cost reductions through its Sol and Luna model lines, helped by more efficient models and infrastructure.

That second point is important for businesses. A discount that depends only on a vendor absorbing losses may not last. A discount supported by cheaper infrastructure or more efficient models may be easier to sustain. The source does not say which customer discounts fall into which bucket, so buyers should treat the headline savings as a starting point for questions rather than a guarantee of long-term pricing.

The report also says Anthropic’s growth slowed from the second quarter to the third quarter, according to an investor memo. Both OpenAI and Anthropic are also facing pricing pressure from China, where models from Deepseek and Kimi are described as offering comparable performance at far lower cost. We have also tracked related competitive pressure in Moonshot’s Kimi K3.1 Pro narrows gap with Claude on AI benchmarks.

What buyers should ask before switching

For marketers, agencies, software teams and small businesses moving into heavier AI use, the lesson is straightforward: do not assume list pricing is the real market price once usage grows.

A few questions are worth asking before signing or renewing:

• Is the discount on seats, API usage, committed spend, or a bundle?
• How long does the lower price apply?
• Can existing workflows be tested on more than one provider before committing?
• What happens if usage grows faster than expected?
• Does the deal reduce actual cost, or mainly increase the number of included seats?

The larger pattern is a price war for enterprise AI adoption. OpenAI’s reported offers may help customers lower costs now, but they also raise the stakes for vendor choice. The best-positioned teams will be those that can benchmark their own workloads, keep alternatives viable, and avoid building every process around a single provider’s temporary commercial terms.