OpenAI

OpenAI is an artificial intelligence research organization based in San Francisco, California, founded in December 2015. It was created around a stated mission: to ensure that artificial general intelligence (AGI) —systems able to match or surpass people at most tasks— benefits all of humanity.

Its founders include Sam Altman, Elon Musk, Greg Brockman and Ilya Sutskever, alongside other researchers and engineers; Musk stepped away a few years later. The project began with the belief that advanced AI should be developed openly and aligned with human values, rather than left in the hands of a handful of governments or corporations. Since then, as documented by encyclopedias such as Wikipedia and Britannica, OpenAI has grown from a little-known lab into a central player in the field, especially after the launch of ChatGPT.

Mission and structure

OpenAI was established as a nonprofit. In 2019 it adopted a hybrid «capped-profit» model: a company whose returns to investors were bounded by a cap, controlled by a nonprofit parent. That same year it forged a strategic partnership with Microsoft, which provided investment and the computing power of its Azure cloud to train the models.

The structure kept evolving. In 2025, as reported by the business press (CNBC, Time, Bloomberg), OpenAI completed a reorganization: its commercial arm became a Public Benefit Corporation, bound by its charter to pursue the mission, while control remained with the nonprofit, now called the OpenAI Foundation. The design aims to reconcile the funding needed for research with the original premise of serving the public interest.

From GPT to ChatGPT, DALL·E, Whisper and Sora

The organization is best known for the GPT (Generative Pre-trained Transformer) family, language models trained on vast amounts of text. GPT-2 arrived in 2019, GPT-3 in 2020 and GPT-4 in 2023, with steady gains in capability and versatility. In November 2022, OpenAI released ChatGPT, a conversational interface that brought these models to a mass audience and accelerated the adoption of generative AI.

Its catalog reaches beyond text. DALL·E generates images from written descriptions; Whisper transcribes and recognizes speech; and Sora, unveiled in 2024, produces video from text prompts. Together they reflect a bet on multimodal systems able to work across language, image, sound and video.

Its role in the ecosystem

OpenAI has become a reference point in the debate over AI: it sets the pace of research, offers its models through an API and a suite of products, and takes an active part in discussions about the safety and governance of these technologies. Its trajectory —from nonprofit lab to an organization with a powerful commercial arm backed by Microsoft— captures much of the tension and opportunity that define contemporary AI.