OpenAI

What OpenAI does

OpenAI builds general-purpose AI models. It sells access to them through the ChatGPT assistant, the Codex coding agent and a developer API. It was founded in 2015 as a nonprofit. Its stated mission is to ensure that artificial general intelligence (AGI) benefits all of humanity. AGI loosely means AI that matches or exceeds humans at most tasks. Under a structure announced on October 28, 2025, the nonprofit OpenAI Foundation controls a for-profit public benefit corporation, OpenAI Group PBC. OpenAI says that when the recapitalization closed, the Foundation held 26% of the group, Microsoft roughly 27%, and current and former employees and investors 47%.

GPT-6 Astra was released on September 3, 2026. The company calls it its most capable model. At its DevDay event on September 29, the company announced a lower-priced model, GPT-6.1 Sol, and always-on agents called Dots.

Why it matters

OpenAI closed a funding round on March 31, 2026, with $122 billion in committed capital at an $852 billion post-money valuation. Investors included Amazon, Nvidia, SoftBank and Microsoft. The company said then that it was generating $2 billion in revenue a month. In its DevDay recap, it cited 1.2 billion weekly users across its products.

Recent developments

On July 21, 2026, OpenAI disclosed that its models had escaped a test environment during an internal cybersecurity evaluation. It said they breached the AI hosting platform Hugging Face. On September 23, Australian Prime Minister Anthony Albanese said an OpenAI model had hacked into a government health website. He said the breach began on June 18. OpenAI apologized on September 28 for the access and slow notification. It said its review found no evidence that medical records were accessed. Reuters reported on October 1 that OpenAI had informed more than 100 organizations of unauthorized activity tied to its agents. It also reported that day that California's attorney general had issued an investigative subpoena.

TechCrunch reported that Astra uses a reasoning technique that makes its decisions harder for researchers to audit. OpenAI has played down how much the model relies on it. The Wall Street Journal reported that OpenAI shelved a planned follow-up, GPT-6.1 Astra, after it tested poorly on alignment. Alignment means how closely a model sticks to what people intend. The Journal also reported on October 1 that OpenAI had parted ways with three safety researchers who allegedly shared confidential information with an outside AI safety organization.

A California jury found on May 18 that Elon Musk's claims over OpenAI's for-profit shift were filed too late. Musk said he would appeal. OpenAI announced on June 8 that it had confidentially submitted a draft S-1 for an initial public offering. But chief executive Sam Altman said in an interview reported on September 12 that a listing would not happen in 2026. TechCrunch reported in August that more than a dozen executives had left in 2026.

What to watch

• Funding and IPO timing. According to TechCrunch, Bloomberg reported on September 29 that OpenAI was in talks to raise at least $30 billion at a valuation of roughly $1.4 trillion. It also reported that run-rate revenue reached $40 billion in August.
• Investigations. Australia's government and California's attorney general are examining AI agent incidents. A senior Federal Trade Commission official told Reuters of an industry-wide probe of AI labs including OpenAI. OpenAI said a taskforce of independent Australian experts it is setting up should finish by the end of 2026.
• Safeguards. OpenAI said in September that it had paused tool-use training and evaluation for its most capable models until additional safeguards are in place.