OpenAI is preparing to open its doors to developers, partners and the wider public next month as DevDay 2025 aims to showcase the company's latest technical advances and commercial strategy. The in-person gathering at Fort Mason in San Francisco, which the company says will host more than 1,500 attendees, culminates with a one-hour keynote from CEO Sam Altman beginning at 10 a.m. Pacific on October 6. That keynote will be broadcast live on OpenAI’s YouTube page.

The company billed the gathering as its "biggest event yet," and the agenda underscores why: Altman is expected to unveil "announcements, live demos, and a vision of how developers are reshaping the future with AI," language the company used to describe the keynote. The program also includes multiple executive presentations and a fireside conversation between Altman and longtime Apple designer Jony Ive, a pairing that signals attention to product design and user experience as AI moves into mainstream applications.

DevDay has become a platform for major product moves. At its inaugural DevDay in 2023, OpenAI introduced GPT-4 Turbo and outlined plans for a marketplace of AI agents the company described as the GPT Store. That event reshaped expectations about how quickly new models would be released and how third-party developers could monetize AI-powered services. This year, developers are watching for upgrades to model capabilities, new tools for agent creation, SDKs, and commercial terms that affect how startups and enterprises build on OpenAI’s stack.

The live-demo format that OpenAI favors can provide immediate, compelling examples of new technology, but it also raises familiar questions about robustness and reproducibility. Live demos historically create moments of excitement while testing the limits of systems under variable conditions. Observers say the real measure of progress will be documentation, APIs and developer tooling that allow teams to integrate advances into products reliably and safely.

Beyond technical details, DevDay 2025 arrives at a fraught moment for the industry. Governments and regulators worldwide are scrutinizing generative AI’s societal impacts, from misinformation to labor disruption. OpenAI’s announcements will be parsed not just for new features but for safety guardrails, transparency measures and how access is controlled. The existence of a centralized marketplace for AI agents—if expanded—would amplify questions about platform power, moderation and economic opportunity for smaller developers.

For those who cannot attend in person, the livestream offers a clear alternative. The one-hour keynote begins at 10 a.m. Pacific (1 p.m. Eastern) on October 6 and will be available on OpenAI’s official YouTube channel; the company has indicated additional sessions and materials will be posted after the event. Developers and observers are advised to watch for subsequent blog posts and API notes that typically follow big product announcements and provide the technical specifics necessary to evaluate and adopt new capabilities.

As OpenAI stages another high-profile gathering, the industry will be watching to see whether the company prioritizes incremental performance gains, new forms of developer commerce, or governance features designed to address the technology’s growing societal footprint. The announcements could set the tone for AI development in the year ahead.