OpenAI’s Anthropic discount push targets big AI customers

OpenAI is reportedly offering major discounts to win business from Anthropic, with some large customers moving more of their AI spending toward OpenAI. The Decoder, citing The Information, reports that the push is already affecting companies using AI for coding and workplace productivity.

The story matters for creators, freelancers and small businesses because enterprise price cuts often signal where the broader AI market is headed. If big customers can push down costs for ChatGPT and competing systems, smaller teams may soon face a more price-sensitive choice between OpenAI and Claude, not just a quality comparison.

What OpenAI is reportedly offering

According to The Decoder’s summary of The Information’s reporting, OpenAI is using aggressive pricing to pull major customers away from Anthropic. The article names Rippling, Replit and DataCamp as companies shifting more AI spending to OpenAI.

The largest reported cuts vary by customer:

• Replit moved after OpenAI offered more than 30 percent off a contract valued at more than $100 million a year, according to the report.
• DataCamp CEO Jonathan Cornelissen said OpenAI’s pricing lowered his company’s costs by 80 to 90 percent.
• A U.S. bank with more than $7 billion in assets reportedly had its ChatGPT seats doubled to 500 employees at no additional cost, bringing its price below $7 per user rather than the usual $20 or more.

Rippling CEO Parker Conrad also said the company changed the balance of its coding workload. OpenAI’s share rose from 16 percent to 52 percent, while Anthropic’s share fell from 57 percent to 18 percent, according to the report.

For OpenAI, this looks like both a pricing and market-share strategy. The Decoder says OpenAI is partly absorbing lower margins to win customers, while also benefiting from cheaper delivery through its Sol and Luna model lines, which the company has made more efficient through model and infrastructure improvements.

Why coding teams are a key battleground

The reported customer list is notable because coding-heavy AI usage can become expensive quickly. Replit is also a developer platform in Emonarc’s directory, and its reported shift follows earlier signs that the relationship between Replit and Anthropic had changed. We previously covered Replit CEO says the company is no longer working with Anthropic.

For software teams, the practical question is not simply which model is best in a single prompt. It is how the model performs across repeated coding tasks, how easy it is to route work between vendors, and what the total monthly bill looks like once usage grows.

That matters for small agencies and solo developers too. Even if these specific discounts are for large customers, they could reset expectations in procurement conversations. Teams that rely heavily on AI for code generation, debugging or internal tools may increasingly ask vendors to justify price differences with clear workflow benefits.

Pressure is building on Anthropic and rivals

The Decoder reports that Anthropic’s growth slowed from the second quarter to the third quarter, citing an investor memo. The report does not say the slowdown was caused solely by OpenAI’s discounting, but the timing shows why price competition is becoming harder to ignore.

Anthropic is not the only company under pressure. The report also points to China-based models from Deepseek and Kimi, which it says offer comparable performance at much lower cost. Emonarc recently covered related pricing pressure in coding with Deepseek V4.1 Flash 1017 targets cheaper coding agents, and Kimi’s progress in Moonshot’s Kimi K3.1 Pro narrows gap with Claude on AI benchmarks.

For buyers, that means the AI assistant market is becoming less about choosing a single default provider and more about workload allocation. A company might prefer one model for writing, another for coding and another for low-cost automation. The more interchangeable the models become for everyday tasks, the more pricing can influence usage.

What smaller teams should watch

The clearest takeaway is that AI budgets are becoming negotiable at the top end of the market. Smaller businesses may not receive the same terms as Rippling, Replit or DataCamp, but they can still learn from the pattern: track actual usage, compare outputs across providers and avoid building every workflow around one model unless the benefits are obvious.

For teams choosing between ChatGPT and Claude, cheaper access can change the spreadsheet but not the full decision. A lower per-seat or contract price is useful only if the tool fits the work: coding, research, drafting, support or internal operations. The reported discounts make cost a bigger part of the conversation, but they do not remove the need to test quality on real tasks.

If OpenAI keeps using pricing as a lever, Anthropic may face more pressure to defend high-value accounts. For users, that could mean better offers, more bundled seats or faster movement toward efficient model options. The risk is that headline discounts may be temporary or limited to large contracts, so smaller buyers should treat them as a signal, not a guarantee.