Open Source x Sovereign AI: SF Tech Week
Every organization is asking the same questions. How do we use AI without giving away our IP? How do we keep costs predictable as usage grows? How much of our AI should we own? This three-day track brings enterprise leaders, builders, and researchers together to work through the answers with open-weight models and sovereign AI. Day 1 is about owning your AI stack. Day 2 is about keeping your IP private. Day 3 brings both together in an open unconference and ends with demo night.
October 6-8, 2026 · San Francisco · Part of SF Tech Week
Hosted by Apollo Innovation Commons. Day 1 host: Outcast Ventures. Headline sponsor: TrustedRouter.
WHO SHOULD COME
Enterprise leaders: CTOs, CIOs, and AI leads deciding what to build, buy, or own
Security, legal, and compliance teams who protect IP and sign off on deployments
Finance and operations leaders watching AI spend and looking for predictable costs
Builders and researchers creating open models and privacy tools
HOW THE DAYS WORK
Builders' curiosity about best practices with open-weight models is at an all-time high. Each day is designed for expert discussions, builder feedback, demo tables, and connective networking.
Doors and coffee at 8:00 AM on Day 1, and 8:30 AM at The Fold on Days 2 and 3
Panels, talks, and working sessions where practitioners share what they chose, what it cost, and what they would do differently
Lunch roundtables every day, with demo tables and a builder mixer on Day 1
Happy hour every evening: until 7:00 PM on Day 1, until 8:00 PM on Day 2, and Day 3 closes with Demo Night and the afterparty at Temple San Francisco
DAY 1 · TUE OCT 6 · OWNING YOUR AI STACK
The strategic case for sovereign AI: when open-weight models make sense, how to verify what your AI is doing, and how to build trust infrastructure for humans and agents. Hosted by Outcast Ventures at the Old GitHub office in SoMa. Curated by Tricia Wang of the Advanced AI Society.
8:00-10:00 AM Doors, coffee and check-in
10:00-10:30 AM Opening with Apollo Innovation Commons, TrustedRouter, Outcast Ventures and Tricia Wang (Advanced AI Society)
10:30-11:10 AM Panel: Open Verification: Expanding Open Source Ethos Into Agentic AI, with Joseph Perla (TrustedRouter), Michelle Finneran Dennedy (Abaxx Technologies) and Mo Sadek (Alice), moderated by Tricia Wang
11:30 AM-12:10 PM Talks: Embedded Governance for Fiduciary AIs, Richard Whitt (GliaNet Alliance) · Who Does Your AI Remember For? Sovereign Memory for People and Their Agents, Junius Ho (Affinidi)
12:30-1:30 PM Lunch + Roundtables
2:00-2:40 PM Getting to Yes with Security, Legal, and Compliance: Sovereignty as a Control Issue, Irina Marinescu (Continuum Partners)
2:40-3:00 PM Lightning talks: Who Owns What We Know?, Dr. Joshua Shinavier (Cognisee) · Whose GPU Is This, Anyway?, Craig Thayer (DC3 Network)
3:00-4:00 PM Builder Mixer + Demo Tables
4:00-4:40 PM Talk + live demo: Deterministic Laws for Probabilistic Minds, Dylan Hobbs (Vouched)
4:40-5:30 PM Panel: The Agentic World: Agentic Memory in Open Source, curated by Learning Layer Labs, with Devinder Sodhi (Learning Layer Labs), Vishakha Gupta (ApertureData), Aditya Shrivastava (HydraDB), and Rayyan Zahid (Immersive Commons)
5:30-7:00 PM Open Source Happy Hour + Demos
6:00-6:40 PM Panel: Building Cooperation and Sovereignty into the Tech Stack, with Athena Aktipis (Cooperative Futures Institute, Arizona State University), moderated by Judy Zhou (Apollo Innovation Commons), Madisen Taylor (AI Commons) and SJ Klein (AI Commons, Public AI Network)
DAY 2 · WED OCT 7 · KEEPING YOUR IP PRIVATE
How organizations put AI to work on their most sensitive material, from trade secrets and product plans to customer and regulated data, and the privacy protections and governance that let security, legal, and compliance teams say yes. TrustedRouter headlines.
8:30-10:00 AM Doors, coffee and check-in
10:00-10:30 AM Opening with Judy Zhou (Apollo Innovation Commons) and Joseph Perla (TrustedRouter)
10:30-11:15 AM Don't miss: Fireside chat: The History and Pitfalls of Open Source, with Bruce Perens (Co-founder, Open Source Initiative) in conversation with Raymond Cheng (Public Goods Foundation)
11:30 AM-12:10 PM Fireside chat: Policy for Sovereignty in the Age of AI, with Sara Ahmadian and Sheila Warren
12:10-12:40 PM Fireside chat: Organizing the Revolution: How the Right Kind of Institutions and Governance Can Lead to Pro-Human AI, with Olga Vox (Alinea Institute) and Brian Behlendorf (Polyfuse)
12:40-1:30 PM Lunch + Roundtables
1:30-2:15 PM Sovereign AI Needs a Sovereign Cloud, with Molly Mackinlay (FilOz), plus showcase demos
2:30-3:10 PM Private by Design: Confidential Computing and Encrypted AI, with Amean Asad (Confidential AI) and Anthony Ronning (Maple AI)
3:30-4:10 PM The Same Fight: Privacy for Payments and Privacy for AI, Brady Forrest (Sentz Global)
4:30-5:10 PM The Commons Is the Moat: Designing Shared Standards That Ship Solutions, with Nathan Argetsinger and Patrick Deegan (Amulet AI)
5:15-6:00 PM Keynote: The Private AI Enterprise, Joseph Perla (TrustedRouter)
6:00-8:00 PM Open Source Happy Hour + Demos
DAY 3 · THU OCT 8 · THE SOVEREIGN ENTERPRISE
Two days in, bringing it together: case studies from organizations running AI on their own terms, a conversation on data sovereignty and policy, an open unconference, and demo night. The unconference board opens at 8:30 AM and anyone can propose a session.
8:30-10:00 AM Coffee + unconference sign-ups: the board opens for case studies, lunch tables and afternoon sessions
10:15-11:30 AM Case Studies: AI on Our Own Terms, with Hang Yin (Phala Cloud), Matthew Wise (Archipelo) and attendee case studies by sign-up
11:45 AM-1:00 PM Data Sovereignty and the Policy Horizon, with Tessa Brown (Germ Network) and breakout tables by sign-up
1:00-2:00 PM Lunch + Open Build, with topic tables by sign-up
2:00-2:30 PM Open Unconference opener: Engineering Loops and Graphs with Local Models, Dustin Dannenhauer (Delegance AI)
2:30-3:00 PM Workshop + demo: Context Tools for Knowledge Graphs, Dylan Kawalec (Tracn)
3:00-5:00 PM Open Unconference: 30-minute sessions by sign-up, opening with a talk by Patrick Deegan (Amulet AI): OpenForest: Data Rights and Trust for Physical AI, followed at 3:20 PM by a panel on Confidential AI with Ansgar Grunseid (Confidential AI) and Joseph Perla (TrustedRouter), moderated by Tessa Brown (Germ Network), and at 3:50 PM by an encore of Joseph Perla's keynote, The Private AI Enterprise (TrustedRouter)
5:00-10:00 PM Demo Night + Afterparty at Temple San Francisco, 540 Howard St: our official happy hour and afterparty with Agents, Atoms, Allocators. RSVP on Partiful
DEMO NIGHT + AFTERPARTY · OCT 8, 5:00-10:00 PM · TEMPLE SAN FRANCISCO
Sovereign AI you can see working. Teams show products and tools that keep data private, bring AI costs down, or run on open models, including projects that came out of the week's sessions. We close the week at Temple San Francisco, 540 Howard St, as the official happy hour and afterparty for Agents, Atoms, Allocators: panels on agents, robotics and capital, then food, drinks and a rooftop DJ set. Doors at 5:00 PM. RSVP on Partiful to get in.
TAKE PART
Space is limited, so register soon and bring the sovereign AI builders, researchers, and operators in your circle. Propose a Day 3 unconference session, apply to demo, or apply for a contributor pass at opensourcesf.com.
Contributor passes are subsidized for people who build or maintain an open model, a privacy tool, or the infrastructure that lets organizations run AI on their own terms.
Registration is free and reviewed. A $200 sponsor ticket is available for anyone who wants to back the track, and full sponsorship packages are at opensourcesf.com.
ABOUT THE HOSTS
Apollo Innovation Commons is a civic innovation hub in San Francisco, organized as a lab of labs across AI, biotech, robotics, climate, and the arts. The work is about widening access to the tools, capital, and shared infrastructure that let people build.
Advanced AI Society is the alliance for verifiable AI. We unite market leaders and deployers with insurers, policymakers, and civil society to standardize, deploy, and scale open verification, an AI security approach purpose-built for the agentic age.
Anchor labs: Learning Layer Labs and MASS Arts. Headline sponsor: TrustedRouter. Day 1 venue courtesy of Outcast Ventures.
Full schedule and speaker updates at opensourcesf.com.