OpenAI — The World's Most Famous AI Research Lab

OpenAI was founded in 2015 as a nonprofit organization with a stated goal of ensuring that artificial general intelligence benefits humanity, before transforming over time into a commercial machine valued today at over $850 billion. The distance between these two realities tells the company’s entire story.

Founding and Founders

The company launched in San Francisco in December 2015 by a group of technologists and investors, most notably Sam Altman, Greg Brockman, Ilya Sutskever, and Elon Musk. The founding idea was clear: create an open research lab confronting the commercial path major technology companies had taken in AI. Initially, Musk and several investors poured over $1 billion, but Musk left the board in 2018 under circumstances still disputed today.

In 2019, a fundamental structural transformation occurred when the company created a capped‑profit subsidiary, opening the door to Microsoft’s $13.75 billion investment. That was the beginning of the most impactful partnership in AI history.

Core Products

ChatGPT launched in November 2022, and was essentially a research experiment that turned into a cultural phenomenon. It reached 100 million users in just two months — a figure no consumer application had achieved at that speed. By February 2026, weekly active users exceeded 900 million, while paying subscribers with monthly fees surpassed 50 million subscribers.

GPT models evolved from GPT‑3 in 2020 to GPT‑5 then GPT‑5.4 by April 2026. Each generation brought distinctive capabilities: GPT‑4 established the multimodal reasoning standard, while GPT‑5 introduced new reasoning capabilities. GPT‑5.4 today powers agentic workflows and processes over 15 billion input tokens per minute via APIs.

DALL‑E, Whisper, and Sora complete the product ecosystem: DALL‑E for image generation, Whisper for speech‑to‑text, and Sora for video generation from text. Codex evolved from a code‑suggestion tool into an independent software agent capable of reading massive codebases and executing hour‑long tasks.

Numbers and Funding

OpenAI’s financial trajectory is hard to believe if you don’t read the numbers in sequence. Revenue rose from $2 billion in 2023 to $6 billion in 2024 then exceeded $20 billion by the end of 2025, recording continuous triple‑digit annual growth. By February 2026, annualized revenue reached $25 billion.

On the funding front, the company closed a $122‑billion funding round in April 2026 led by SoftBank, Andreessen Horowitz, TPG, and others, following a $110‑billion round in February 2026 with participation from Amazon and NVIDIA. The current valuation exceeds $850 billion, making it the world’s most valuable private company.

Financial burdens, conversely, are massive: inference costs reached $8.4 billion in 2025 and are expected to rise to $14 billion in 2026. The company won’t reach positive cash flow before 2030.

Project Stargate and Infrastructure

In January 2026, OpenAI announced Project Stargate to build $500 billion of infrastructure over three years in partnership with Oracle, SoftBank, and other investors. The first site in Abilene, Texas, houses NVIDIA GB200 processors. The overall ambition is to reach 10 gigawatts of computing power.

Diversifying from dependence on Microsoft as a sole computing supplier became a strategic priority, as the company struck deals with AMD for alternative chips and the right to acquire up to a 10% stake.

Structural Transformations and Internal Tensions

In October 2025, OpenAI completed a foundational restructuring that transformed it into a public‑benefit corporation, while retaining the renamed nonprofit OpenAI Foundation with a 26% stake. Attorneys general in California and Delaware approved the new structure after lengthy negotiations.

On the leadership front, a number of founders and executives left the company during 2024 and 2025, including chief scientist Ilya Sutskever and several leaders from safety and research teams. These collective departures raised serious questions about strategic direction and safety priorities.

Competition and Challenges

Main competitors in 2026 are Anthropic, Google DeepMind, and Meta, along with open‑source Chinese models. ChatGPT holds consumer leadership by a clear margin: 6× web visits compared to the nearest competitor, and 4× user time spent. However, Anthropic strongly competes for enterprise clients.

The relationship with Microsoft grew more complex in 2026, as OpenAI entered direct competition with Microsoft‑owned GitHub through the Codex agent. The renegotiated agreement in October 2025 obligates OpenAI to pay 20% of its total revenue to Microsoft through 2032.

2026–2027 Outlook

The stated priority for 2026 is “practical adoption,” meaning narrowing the gap between what AI can do and how companies actually use it. Health, science, and enterprises are the target sectors for expansion.

Personal AI devices are scheduled for launch in 2026 after completing the acquisition of Jony Ive’s design studio. The OpenAI software‑collaboration platform will put the company in direct confrontation with GitHub and GitLab. And as enterprise‑sector revenue approaches consumer‑revenue levels, OpenAI is shaping into an integrated platform, not just a consumer product.

The question that remains open: Can a company burning $17 billion annually, competing with free open‑source models, and undergoing repeated internal struggles maintain its technical leadership and reach profitability at the same time? This equation hasn’t been settled yet.

---

Explore More

Want to learn more about the latest models mentioned here? Visit our Top AI Models List for a comprehensive comparison, or browse the Latest AI Tools to boost your productivity.