OpenAI is the San Francisco research company whose chatbot, ChatGPT, introduced generative AI to the general public in November 2022. A decade after its founding as a nonprofit lab, it is simultaneously one of the world’s most valuable private companies and one of its most scrutinised.
Its mission statement has remained constant — to ensure artificial general intelligence benefits all of humanity — while almost everything about its structure, funding and product line has been rebuilt around that sentence. Understanding OpenAI means holding both facts at once.
From nonprofit lab to hybrid giant
OpenAI was founded in 2015 by a group including Sam Altman and Elon Musk, organised as a nonprofit dedicated to open, safety-conscious research. The sums required to train frontier models broke that model within four years: in 2019 it created a capped-profit subsidiary to take outside investment, most consequentially from Microsoft.
The structure kept evolving under commercial pressure. After a failed board coup against Altman in November 2023 exposed the fragility of nonprofit control over a commercial rocket ship, OpenAI restructured in 2025: the commercial arm became a public benefit corporation, with the nonprofit — renamed the OpenAI Foundation — holding a major stake and board role.
What OpenAI sells
ChatGPT is the consumer face, sold in tiers from free to Pro. The API is the industrial one: developers rent GPT models per token to power their own products, and it remains a pillar of the business. Around those two sit an expanding platform — custom GPTs, agent capabilities, deep research and image generation.
The Microsoft relationship threads through everything. Microsoft is OpenAI’s largest backer and cloud partner, embeds its models across Copilot products, and simultaneously competes with it for enterprise customers — a partnership both companies periodically renegotiate as their interests shift.
The pressures: cost, law and safety
OpenAI’s economics are dominated by compute. Training and serving frontier models consumes data centres at a scale that commits the company to infrastructure spending measured in the hundreds of billions, funded by investors betting the revenue catches up.
Legal pressure is the second front. Beyond copyright suits from publishers and authors, eight state attorneys general sued the company in October 2026, alleging ChatGPT was sold as safe while endangering children. OpenAI disputes the claims, but the cases mark a shift: AI safety marketing is now litigated as consumer protection.
The third pressure is the mission itself. Every restructuring invites the charge that the nonprofit purpose is being diluted by commercial gravity; every safety delay invites the opposite charge from competitors. OpenAI’s governance is best read as a permanent negotiation between those two accusations.
People, rivals and the ecosystem
OpenAI’s gravitational pull shows in the talent market. Its researchers are recruited like athletes, its alumni have founded or joined nearly every serious rival — Anthropic most famously — and its internal debates about safety and speed have repeatedly become public arguments about the industry’s direction.
Around the company sits an ecosystem of developers building on its API, startups whose products are thin layers over its models, and enterprises standardising on its stack. That ecosystem is a moat and a vulnerability at once: it multiplies OpenAI’s reach, and it means a pricing or policy change in San Francisco ripples through thousands of businesses by morning.
The rivals are not standing still. Google bundles comparable capability into products billions already use, Anthropic owns growing share of professional and coding work, and open-weight models keep eroding the price of “good enough” intelligence. OpenAI’s lead is real; so is the speed at which this industry has historically punished complacency.
The company’s footprint is now physical as well as digital. Its Stargate programme — a multi-year build-out of giant data centres with partners including Microsoft, Oracle and SoftBank — commits OpenAI to owning the infrastructure layer its rivals rent, and makes it a player in energy and industrial policy, not just software.
Where it stands
OpenAI remains the category’s centre of gravity: the most recognised brand, the largest consumer base, and a research organisation still capable of setting the industry’s agenda with a single launch. Its rivals — Anthropic, Google, and a field of open-weight challengers — define themselves partly against it.
Whether its hybrid structure can carry frontier costs, legal exposure and a public-benefit mission at the same time is the defining corporate question of the AI era. ChatGPT’s success bought OpenAI the right to face that question at full scale.
0 Comments