A look at OpenAI's business model, as the company works to more than double its yearly run rate by 2025; ChatGPT Enterprise now has 300+ paying customers
The AI startup is one of the fastest-growing companies in history, but questions remain about the long-term viability of its revenue stream X: @madhumita29 and @jbrowder1 X: Madhumita Murgia / @madhumita29 : For today's @ft Big Read, we looked at OpenAI's business - how it's growing, who is seeing utility from it and what its future looks like. With @GeorgeNHammond Joshua Browder / @jbrowder1 : Fascinating A.I. story in the @FT today on how API pricing impacts DoNotPay. In bill negotiations, it is an arms race between our AI and the large company AI. When our agent needs a boost, we switch to a more expensive model mid-conversation! https://www.ft.com/... [image]
Context & Ripple Effects
OpenAI entered this period after reporting a sharp rise in annualized revenue, with the company having recently passed $1.6B in annualized revenue. The business-model question is whether that growth can become durable across both direct subscriptions and API usage.
Enterprise adoption is still early by customer count, while API buyers such as DoNotPay expose how model pricing directly shapes the economics of downstream AI products. Later reporting of more than 1M paid corporate ChatGPT users suggests the enterprise channel became an important test of that proposition.
First-order effects
- ChatGPT Enterprise’s 300-plus paying customers give OpenAI a direct business-sales base alongside consumer and API revenue as it pursues a much higher run rate.
- API pricing becomes an immediate operating constraint for customers such as DoNotPay, which may switch to higher-cost models when a task requires more capability.
Second-order effects
- AI application builders face pressure to route tasks across models and control usage, because their service margins can move with OpenAI’s API prices and model-selection needs.
- Enterprise buyers gain a clearer packaged route to deploy OpenAI, while rival model providers have an incentive to compete on capability, pricing, and predictable operating costs.
Third-order effects
- The sector is moving toward a two-layer AI market: vendors monetizing enterprise seats and usage, and application companies managing inference cost as a core cost of goods sold.
- Whether rapid revenue growth proves durable will depend on whether recurring customer value rises faster than the cost of serving increasingly capable models; this report identifies that tension but does not resolve it.
The trend: Generative-AI vendors are shifting from initial adoption toward proving that enterprise subscriptions and API usage can support sustainable unit economics.