OpenAI makes money from ChatGPT subscriptions (consumer, team and enterprise plans), usage-based API fees paid by developers and companies, and, since early 2026, advertising and commerce on the free tiers of ChatGPT. Its CFO, Sarah Friar, wrote in January 2026 that annualized revenue ran at more than $20 billion in 2025, up from $2 billion in 2023, and that revenue has tracked the computing capacity OpenAI could bring online. OpenAI is private; these figures are the company’s own and are not audited public accounts.
ChatGPT runs a tiered plan system: Free, Go, Plus and Pro for individuals, and Business and Enterprise plans priced per user for organisations.
02
Usage-based API
Developers and enterprises pay for the tokens they use through the API, so spend rises with production workloads. OpenAI also added usage-based pricing inside workplace plans.
03
Ads and commerce on the free tiers
Ads appear only on the Free and Go plans; Plus, Pro and Enterprise stay ad-free. Commerce features let people buy inside ChatGPT, with payments run by partners such as Stripe.
04
Compute bought, not owned
OpenAI says it keeps the balance sheet light by partnering rather than owning data centres, and commits capital in tranches against demand.
What changed in 2025-26
Revenue tracked compute
In a January 2026 post, Friar said compute grew from 0.2 GW in 2023 to 0.6 GW in 2024 and about 1.9 GW in 2025, while annualized revenue went from $2 billion to $6 billion to more than $20 billion over the same years.
OpenAI said on 21 October 2026 that more than 100,000 advertisers had run campaigns in the just over eight months since ads launched, and extended ChatGPT Ads to 89 countries. It says conversations stay private from advertisers and it does not sell customer data.
Enterprise spend widened
At DevDay on 29 September 2026 OpenAI launched a Marketplace where eligible enterprise customers can apply part of their existing OpenAI commitment to approved partner software, starting with 32 partners. A forward-deployed engineering team, founded in early 2025, works inside large customers.
Under an amended Microsoft agreement of 27 April 2026, OpenAI can serve all its products on any cloud, and its revenue-share payments to Microsoft continue through 2030 at the same percentage, subject to a total cap. On 21 October 2026 OpenAI said its models and Codex are coming to Google Cloud.
OpenAI's revenue lines in practice
ChatGPT subscriptions
Who paysIndividuals (Go, Plus, Pro) and organisations (Business, Enterprise), per user per month
Who paysMore than 100,000 advertisers; ads on Free and Go only
SourceOpenAI, 21 October 2026
Commerce
Who paysMerchants selling through Instant Checkout in ChatGPT
SourceStripe, 29 September 2025
Enterprise Marketplace
Who paysEnterprises applying OpenAI commitments to 32 partner products
SourceOpenAI DevDay, 29 September 2026
What it means for you
For marketers: treat ChatGPT as a paid channel that reaches people while they compare options; ads run only on the Free and Go tiers.
For merchants: check whether your payment provider supports checkout inside ChatGPT, because buying is moving into the conversation.
For enterprise buyers: an OpenAI spending commitment can now cover some partner software through the Marketplace, which changes how you budget for AI tools.
For investors: read OpenAI's revenue claims as company statements; the audited numbers that touch OpenAI sit in Microsoft's filings.
For builders on the API: price your product knowing your costs scale with tokens, and that OpenAI products now run on clouds other than Azure.
The Business Engineer lens
OpenAI’s model is a loop in which compute capacity sets the ceiling on revenue, so its growth depends on financing ever larger compute commitments. The Business Engineer view is that ads and commerce on the free tier are what let a billion-user product pay for that compute without charging most of its users.
OpenAI's CFO said annualized revenue was more than $20 billion in 2025, up from $6 billion in 2024 and $2 billion in 2023. OpenAI is private, so these are company figures, not audited public accounts.
Does ChatGPT have ads?
Yes, on the Free and Go plans. OpenAI said on 21 October 2026 that more than 100,000 advertisers had used the platform, which is available in 89 countries.
How much does OpenAI pay Microsoft?
OpenAI pays Microsoft a revenue share that continues through 2030, subject to a total cap. Microsoft recorded $24.1 billion of revenue from OpenAI, including revenue-sharing payments, in its fiscal year to 30 June 2026.
The complete explainer, cleaned up and reformatted for mobile.
Last Updated: April 2026 โ Enhanced with AI business impact analysis
This remarkable growthstory is built on three primary revenue streams, strategic partnerships, and a controversial shift from its founding principlesโa transformation that has redefined the AI industry while raising fundamental questions about the future of artificial intelligence development.
API Revenue: The Engine of Growth
OpenAI’s API is priced by usage: developers and enterprises pay for the tokens they use, so spend grows with production workloads, according to OpenAI CFO Sarah Friar (January 2026). At DevDay on 29 September 2026 OpenAI also launched a Marketplace where eligible enterprise customers can apply part of their OpenAI commitment to approved partner software.
Consumer Subscriptions: The Steady Foundation
ChatGPT โ as explored in the intelligence factory race between AI labs โ is sold in tiers: Free, Go, Plus and Pro for individuals. Ads appear only on the Free and Go plans; Plus, Pro and Enterprise stay ad-free (OpenAI, 21 October 2026).
The subscription offering has expanded beyond simple chat access to include advanced features like custom GPTs, higher usage limits, priority access during peak times, and early access to new models.
Enterprise Tier: Premium Services for Corporate Giants
For organisations, ChatGPT Business and ChatGPT Enterprise are priced per user, with monthly or annual contracts and the option to buy credits for more access (ChatGPT pricing page, October 2026).
The For-Profit Transition: Redefining the Mission
OpenAI moved from a non-profit research organization to a “capped-profit” structure in 2019. In the recapitalization of 28 October 2025, the for-profit became OpenAI Group PBC, a public benefit corporation controlled by the nonprofit OpenAI Foundation; Microsoft’s stake was valued at about $135 billion, roughly 27 percent on an as-converted diluted basis, after the deal.
This transition has enabled OpenAI to raise unprecedented amounts of capital while creating tensions with its founding mission to ensure AGI benefits all humanity. Critics argue that the profit motive has shifted priorities toward commercializable applications rather than safety research and democratic access to AI capabilities.
Microsoft Partnership: A Double-Edged Alliance
Microsoft’s funding commitments of $13.0 billion ($11.9 billion funded as of 30 June 2026, per its fiscal 2026 10-K) have provided OpenAI with essential capital and computational resources while creating a complex interdependence. This arrangement has provided OpenAI with access to Azure’s massive computational infrastructure โ as explored in the economics of AI compute infrastructure โ , essential for training and running increasingly sophisticated models.
However, the partnership also creates strategic tensions. Microsoft competes directly with many of OpenAI’s API customers while simultaneously providing the infrastructure that powers their competitors’ AI capabilities. The arrangement gives Microsoft significant influence over OpenAI’s strategic direction, raising questions about the company’s independence.
The Compute Cost Challenge
Revenue has tracked compute. OpenAI’s CFO wrote in January 2026 that compute grew from 0.2 GW in 2023 to 0.6 GW in 2024 and about 1.9 GW in 2025, and that more compute would have meant faster adoption and monetization. OpenAI says it partners rather than owns capacity and commits capital in tranches against demand; in October 2025 it contracted to buy an incremental $250 billion of Azure services from Microsoft.
This creates a fundamental tension in OpenAI’s businessmodel: the need to generate sufficient revenue to fund increasingly expensive model development while maintaining competitive pricing and accessibility. As models become more sophisticated, compute costs continue to escalate, requiring ever-larger revenue streams to maintain profitability.
Mission vs. Profit: The Central Tension
OpenAI’s remarkable commercial success has come at the cost of increasing tension between its stated mission and market realities. While the company continues to invest in safety research and maintains some free access to its technologies, the pressure to generate returns for investors and fund expensive AI development has shifted focus toward profitable applications.
This evolution reflects the broader challenge facing AI development: balancing the enormous costs and commercial pressures of advancing AI capabilities with the goal of ensuring these technologies benefit humanity broadly rather than concentrating power and capability among a few well-funded organizations.
THE BUSINESS ENGINEER
Complete Business Model Analysis
How the world’s most valuable companies actually make money โ with proprietary frameworks and competitive maps.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
