Fireworks AI hosted Forge 2026 on November 3 at Pier 27 along San Francisco’s Embarcadero, bringing together technology executives, AI developers and software industry leaders. The one-day San Francisco AI conference featured NVIDIA CEO Jensen Huang and David Heinemeier Hansson, with sessions focused on open AI models, specialized AI systems and enterprise applications. A waterfront fireworks display accompanied the event’s closing activities.
Key Takeaways
- Fireworks AI organized Forge 2026 on November 3, 2026, at Pier 27 in San Francisco
- NVIDIA CEO Jensen Huang participated in a discussion with Fireworks AI CEO and co-founder Lin Qiao
- David Heinemeier Hansson and Hugging Face co-founder Thomas Wolf were included in the conference program
- Sessions covered open AI models, model training, inference, enterprise deployment and AI development costs
- An approximately 10-minute fireworks display took place along the Embarcadero waterfront
Fireworks AI Hosts Forge 2026 at San Francisco’s Pier 27
Fireworks AI held Forge 2026 at Pier 27 on Tuesday, November 3, bringing technology professionals to San Francisco for discussions about developing and deploying artificial intelligence systems. The conference centered on open models and how organizations can adapt them to their own data, technical requirements and business applications.
The program included keynote presentations, executive discussions, technical sessions and workshops. Topics ranged from model customization and inference infrastructure to AI agents and the operational requirements of enterprise deployment.
Fireworks AI CEO and co-founder Lin Qiao opened the conference with a keynote addressing specialized intelligence. The event’s broader agenda explored how developers can train, fine-tune and operate AI systems designed for particular applications rather than relying exclusively on general-purpose models.
The conference also offered livestream access to selected sessions, allowing participants outside San Francisco to follow portions of the program. In-person registration had sold out ahead of the event.
Jensen Huang and Technology Leaders Join Conference Program
NVIDIA President and CEO Jensen Huang participated in a morning fireside conversation with Fireworks AI CEO Lin Qiao. The discussion formed part of the conference’s opening program, connecting the event’s emphasis on open models with the computing infrastructure used to develop and deploy artificial intelligence systems.
NVIDIA supplies graphics processing units and computing platforms widely used for AI training and inference. Its involvement in the Bay Area’s AI sector also extends to venture investment. The company’s investment arm, NVentures, participated in Series A financing for San Francisco AI startup Reactor, which develops infrastructure for real-time AI applications involving video, simulated environments and robotics.
David Heinemeier Hansson, Basecamp’s chief technology officer and creator of Ruby on Rails, was another featured participant. The conference agenda paired Hansson with Thomas Wolf, co-founder and chief science officer of Hugging Face, for a discussion about artificial intelligence and software development.
Microsoft Executive Vice President Jay Parikh also appeared on the program, alongside representatives from companies including AMD, Meta, Replit and Salesforce.
The speaker lineup brought together professionals working across computing hardware, software engineering, AI infrastructure and enterprise technology. Their participation reflected the conference’s emphasis on both the technical development of AI systems and their use in business environments.
Open AI Models and Enterprise Applications Shape Conference Sessions
Open artificial intelligence models and specialized AI applications were central to Forge 2026. The conference examined how organizations can adapt models to specific requirements while addressing the computing, evaluation and operational challenges involved in deployment.
Open-weight models give developers access to model parameters that can support customization, subject to applicable licenses. Businesses can use these systems for tasks such as document processing, internal knowledge retrieval, software assistance and customer-facing applications.
The Forge 2026 agenda included sessions on fine-tuning, model routing, inference engineering, evaluation pipelines and AI spending. These subjects are relevant to organizations selecting models, assessing their performance and managing the resources required to operate AI applications.
Several presentations addressed enterprise implementation directly. The program included discussions involving PayPal’s approach to AI benchmarks and governance, UiPath’s enterprise automation applications and Moody’s use of specialized AI systems.
San Francisco-based Together AI has also outlined enterprise AI infrastructure plans involving IBM Cloud and NVIDIA hardware. The company’s announced agreement is intended to expand computing capacity for open-model inference services, with the dedicated infrastructure expected to become available in the first quarter of 2027.
At Forge 2026, the emphasis on model training, inference and enterprise deployment addressed practical considerations for AI developers. Organizations implementing these technologies must evaluate how models perform on specific tasks, how they integrate with existing systems and what resources are required to maintain them.
Workshops provided an additional technical component, with scheduled activities covering model routing and reinforcement learning environments for AI agents.
Together, the sessions connected open-model development with the requirements of operating AI applications in production environments.
Fireworks AI Organizes Waterfront Fireworks Display
Fireworks AI also organized an approximately 10-minute fireworks display along San Francisco’s Embarcadero waterfront as part of Forge 2026.
The display began shortly before 7 p.m. on November 3 near Pier 27, where the conference had taken place throughout the day. It featured illuminated patterns, including smiley-face designs, visible above the waterfront.
The conference’s published agenda included fireworks as part of its closing reception, extending the event beyond its scheduled presentations and technical sessions.
Pier 27’s location along the Embarcadero placed the display near public waterfront areas, making it visible beyond the conference venue. The activity attracted attention in the surrounding area while connecting the evening display with the technology gathering.
The event concluded a day of discussions involving executives, software developers and AI infrastructure specialists. Forge 2026 combined technical presentations about open models and enterprise AI applications with a public-facing waterfront display in San Francisco.
Frequently Asked Questions
What was Forge 2026 in San Francisco?
Forge 2026 was a one-day artificial intelligence conference organized by Fireworks AI on November 3, 2026. It focused on open AI models, specialized intelligence and enterprise AI applications.
Where was Forge 2026 held?
The conference took place at Pier 27 along the Embarcadero waterfront in San Francisco, California.
Who were the featured speakers at Forge 2026?
Featured participants included NVIDIA CEO Jensen Huang, Fireworks AI CEO Lin Qiao, David Heinemeier Hansson, Hugging Face co-founder Thomas Wolf and Microsoft Executive Vice President Jay Parikh.
What topics did the Fireworks AI conference cover?
Conference sessions addressed open-model development, fine-tuning, AI inference, model routing, evaluation, enterprise deployment and the costs associated with operating AI systems.
Was there a fireworks display during Forge 2026?
Yes. An approximately 10-minute fireworks display took place along San Francisco’s Embarcadero waterfront on November 3 as part of the conference’s closing activities.
