Flagship eventSep 12, 2026 · 2–7 PMSan Francisco100 seats · application only

AI Infra Signal: From Research to Production.

The AI stack is still moving — from the models to the systems that train them and the infrastructure that keeps them running. This is a 100-person, application-only room where founders, researchers, enterprise leaders, and infra investors debate what is still unresolved — then carry those conversations into two technical tracks, a salon, and a 90-minute demo floor.

AI Infra Signal key visual

Curated, not crowded.

100

Application-only attendees, reviewed for fit

11

Talks — 3 keynotes + 8 lightning across 2 tracks

90 min

Demo floor & networking

4

Sponsor slots max — no sponsor walls

Confirmed guests

Lightning Speakers.

Tianyi Zhang

AI Researcher @ Together AI

Harshit Joshi

CS PhD @ Stanford SAIL

Yuexing Hao

Founder @ MatrAIx AIResearcher @ MicrosoftPostdoc @ MIT

Xiaomin Li

Senior Research Scientist @ Microsoft

Biwei Huang

Founder @ Aether AIAssistant Professor @ UC San Diego

Yash More

Member of Technical Staff @ Cohere

Dexter Horthy

Founder @ HumanLayer (YC F24)

More coming soon

View on Luma

Every seat reviewed for fit.

No general AI-enthusiast crowd — every conversation at your booth is with a qualified prospect, partner, hire, or investor.

Founders & CTOs

Teams building AI infra, agents, inference, training, data, and security.

Frontier researchers

Researchers and engineers from frontier labs working on AI systems.

Enterprise infra leaders

Deploying AI at scale — with real budgets and real constraints.

Infra investors

Pre-seed to Series B infrastructure VCs and technical angels.

Stage time and floor time.

2:00 PM
Doors, coffee, booth preview
3:00 PM
Opening
3:10 PM
Technical panel
3:40 PM
Keynote block
4:20 PM
Lightning talks I — AI infra & data
4:45 PM
Lightning talks II — research
5:10 PM
Technical salon
5:45 PM
Demo floor, food & drinks — until 7:00 PM

Why join the room.

For the people building, researching, deploying, and backing AI infrastructure.

Open questions

The talks begin where polished demos usually stop.

Speakers bring the architecture decisions, reliability constraints, and unresolved problems they are still working through — not a finished success story.

Full-stack context

See how one decision travels from research to production.

Compute and training, inference and agents, evals and reliability, data and security — discussed as one connected system rather than separate tracks.

The room

Meet the people building and deciding at the infrastructure layer.

Founders, frontier researchers, enterprise infrastructure leaders, and investors — 100 application-only seats reviewed for fit.

Floor time

The useful conversation continues after the talk.

A technical salon and 90-minute demo floor give you time to challenge an idea, inspect a product, and continue the discussion face to face.

We’ve done this before — at scale.

12–20

Roundtable Chat

Closed-door, frontier AI researchers

75–100

AI Salon Night

Deep tech discussion, curated list

100

GTM Series

Founders with technical backgrounds

1,000

Soulscape 2026

3-day summit at KQED · 2,000+ applicants

Roundtable chat · 12–20, closed-door
Roundtable chat · 12–20, closed-door
GTM series · founder panel
GTM series · founder panel
Demo night · live pitches
Demo night · live pitches
Soulscape 2026 · KQED main stage
Soulscape 2026 · KQED main stage
Soulscape 2026 · 1,000 attendees
Soulscape 2026 · 1,000 attendees
Soulscape 2026 · closing night at KQED
Soulscape 2026 · closing night at KQED

Past partners

Bay Area Founders Club · Beta Fund · GMI Cloud · Inference AI and more

Interested in partnering?

Contact the OpenStages team directly.

partnership@openstagesai.com