How Does OpenAI Make Money? The ChatGPT and API Business Model

Table of contents

Open Table of contents

• What OpenAI is
• How does OpenAI make money?
  • 1. ChatGPT subscriptions (the biggest line)
  • 2. The API (usage-based, for developers)
  • 3. Enterprise contracts
  • 4. The Microsoft partnership
  • 5. Newer and adjacent products
• Is OpenAI profitable?
• How this compares to Anthropic
• OpenAI Revenue Model — 2026 FAQ
  • What is OpenAI’s biggest source of revenue?
  • How does OpenAI’s API make money?
  • Is OpenAI publicly traded? Can I buy OpenAI stock?
  • How does the Microsoft partnership make OpenAI money?
  • Does OpenAI make money from free ChatGPT users?
• The shorter version

What OpenAI is

OpenAI is the AI research company behind ChatGPT and the GPT family of models, plus products like the Sora video model, image generation, and the Codex coding agent. Founded in 2015, it reached mainstream fame when ChatGPT launched in late 2022 and became one of the fastest-growing consumer products in history.

Its structure is unusual: it began as a nonprofit and built a capped-profit/for-profit arm to raise the enormous capital that training frontier models requires. It’s not publicly traded, and it has a deep, multi-year partnership with Microsoft that provides compute, distribution, and capital. The product, like every AI lab, is intelligence-as-a-service — sold across consumer, developer, and enterprise channels.

How does OpenAI make money?

1. ChatGPT subscriptions (the biggest line)

This is what makes OpenAI different from its peers. ChatGPT is free to use, with paid tiers that convert a slice of its huge user base into recurring revenue:

• ChatGPT Plus — a flat monthly fee for access to the best models, higher limits, and premium features. The mass-market tier.
• ChatGPT Pro — a higher-priced tier for power users wanting maximum usage and the most capable model settings.
• ChatGPT Team — per-seat plans for small businesses, with shared workspaces and admin tools.
• ChatGPT Enterprise — for large organizations: advanced security, compliance, SSO, larger context, and usage guarantees.
• ChatGPT Edu — a version tailored to universities and schools.

Because ChatGPT reaches hundreds of millions of weekly users, even a low single-digit conversion rate to paid produces an enormous subscription business. This consumer scale is OpenAI’s defining advantage, and subscriptions are reportedly its largest revenue source.

2. The API (usage-based, for developers)

Developers and companies build OpenAI’s models into their own products and pay per token — per chunk of text (or image, or audio) processed. Pricing scales with model capability: the flagship reasoning models cost more per token than the smaller, faster, cheaper ones, and output is priced higher than input.

The API turns every company building on GPT into a metered customer whose bill grows with their own usage. It’s the same compounding dynamic every AI lab relies on: a startup that embeds OpenAI and scales to millions of users generates more API revenue every month with no new contract.

3. Enterprise contracts

Beyond self-serve API and Team plans, OpenAI signs large, custom deals with big companies — bulk usage, dedicated capacity, custom support, and security/compliance commitments. These are recurring, expand over time, and are sticky once a company builds critical workflows on top of the models. This enterprise motion sits alongside the consumer business and is a major growth area.

4. The Microsoft partnership

Microsoft is OpenAI’s largest strategic partner. The relationship works on several axes:

• Compute — Microsoft’s Azure cloud provides much of the infrastructure OpenAI trains and serves models on.
• Distribution — OpenAI’s models are offered through Microsoft’s platforms (Azure’s AI services, Copilot products), putting GPT in front of Microsoft’s gigantic enterprise customer base.
• Revenue sharing — the two companies share revenue under their commercial agreement, and Microsoft has invested heavily in OpenAI.

This partnership is part capital, part go-to-market: it gives OpenAI reach into enterprises it would take years to sell to directly.

5. Newer and adjacent products

OpenAI keeps expanding the surface it can monetize:

• Codex — its agentic coding tool, monetized through subscriptions and API usage (and a driver of heavy token consumption).
• Sora — video generation, offered within paid tiers and as a product in its own right.
• Image generation and other modalities — bundled into subscriptions and metered via the API.
• A developer/agent ecosystem — custom GPTs, an agents platform, and tools that let businesses build on top of OpenAI’s models.

Each of these is another wrapper around the same core asset, aimed at capturing more of what users and developers are willing to pay for.

Is OpenAI profitable?

OpenAI is private and doesn’t publish audited financials. The widely reported picture: revenue is very large and growing fast, but so are costs — training frontier models and serving hundreds of millions of users consumes staggering amounts of compute. Like its peers, OpenAI is in a heavy-investment phase where the priority is growth and capability, not near-term profit. The bet is that scale plus rising enterprise adoption eventually outpaces compute costs.

How this compares to Anthropic

The building blocks are similar — consumer subscriptions, a usage-based API, enterprise deals, coding tools — but the emphasis differs. OpenAI’s defining edge is consumer scale (ChatGPT) and its Microsoft partnership; Anthropic leans harder into the developer/enterprise API and is backed by Amazon and Google. For the other side of the comparison, see how Anthropic makes money.

OpenAI Revenue Model — 2026 FAQ

What is OpenAI’s biggest source of revenue?

ChatGPT subscriptions. Because ChatGPT reaches hundreds of millions of users, its paid tiers (Plus, Pro, Team, Enterprise, Edu) make up OpenAI’s single largest revenue line — an unusual profile for an AI lab, most of which earn more from APIs and enterprise than from consumers.

How does OpenAI’s API make money?

Developers pay per token to use OpenAI’s models in their own apps — per chunk of text, image, or audio processed. More capable models cost more per token, and output is priced higher than input. Revenue grows automatically as customers’ own usage grows.

Is OpenAI publicly traded? Can I buy OpenAI stock?

No. OpenAI is privately held and its shares are not available on public exchanges. Most people cannot buy in directly. Microsoft holds a major stake through its partnership, but that’s not the same as OpenAI being public.

How does the Microsoft partnership make OpenAI money?

Microsoft provides Azure compute, distributes OpenAI’s models through its products and cloud to a huge enterprise base, and the two share revenue under their commercial agreement. Microsoft has also invested heavily in OpenAI. It’s both a funding source and a distribution channel.

Does OpenAI make money from free ChatGPT users?

Not directly — the free tier is a funnel. Revenue comes when free users upgrade to Plus or Pro, when businesses buy Team or Enterprise seats, and when developers build on the API. The free product’s role is reach; the paid tiers and API convert it.

Related reading: How does Anthropic make money · How does SpaceX make money · The beginner’s guide to AI agents

---

The shorter version

OpenAI turns ChatGPT’s enormous user base into subscription revenue (Plus, Pro, Team, Enterprise), meters developers per token through its API, signs large enterprise deals, and leans on Microsoft for compute, distribution, and shared revenue. Its defining trait is consumer scale — most AI labs monetize developers first; OpenAI built a consumer phenomenon and a business behind it.