OpenAI zeroes in on enterprise market with cheaper AI

Sarah Friar, OpenAI's chief financial officer, said Monday that the company is focusing on specialized sectors including chip design, life sciences, and financial services. In parallel, OpenAI is lowering prices to counter cost-competitive open-source alternatives as demand for AI in the enterprise accelerates.

Speaking at Goldman Sachs' Communacopia + Technology Conference in San Francisco, Friar observed that businesses are increasingly looking for AI tools purpose-built for specific jobs, rather than broad, general-purpose systems.

She added that OpenAI is piloting outcome-based pricing models, moving away from usage-based fees, as corporate clients press for more clearly defined returns on their AI investments.

The move comes amid intensifying competition from Chinese open-weight models and rivals such as Anthropic, with enterprise customers scrutinizing the cost-effectiveness of their AI expenditures more closely than before.

Friar highlighted OpenAI's own work on its 'Jalapeno' chip as a demonstration of the technology's value, noting that the chip's design was completed and submitted for manufacturing—a phase referred to as tape-out—in just nine months using OpenAI's models.

Open-source and open-weight systems are generally seen as a more economical choice compared with OpenAI's and Anthropic's frontier models.

Friar said OpenAI reduced the price of Luna, its budget-tier model, by 80%, which helped drive usage up roughly tenfold. She also observed that Codex, the company's coding assistant, has expanded to 25 million users.

Enterprise revenue climbed 32% between June and July, Friar said, exceeding the 20% growth in total annualized revenue recorded across the company over the same interval.

By the middle of the year, enterprise and consumer revenue had reached near parity, according to Friar, putting OpenAI ahead of its target to fully balance the two business lines by year-end.

Friar contended that running Luna can be more cost-effective than deploying Chinese open-source models through cloud providers. "If you're deploying Luna and compare that to (Z.ai's) GLM 5.3, for example, on a cloud layer, we are cheaper," she said.