Skip to main content

oct 15 / san franciscowestin st. francissingle track

This is the AISS flagship event.

One full day. One room. No filler.

The people building with AI and the people who own the risk, on the same stage: agent identity, attacks at machine speed, AI-powered defense, and where the money is (or should be) going.

Save your seat

our-partners/

AgentCloak Arcjet AWS Cognition CrewAI Galileo AI Guidepoint HumanLayer NVIDIA Oleria Qodo RunSybil SpecterOps
founding partners AI Engineer Snyk
founding sponsor Keycard
Arcade Atlassian Casco Cohesity Docker Google Harvey JPMC Okta Orca RBC Skyflow Anthropic Arize Keycard Microsoft Opaque Systems

Interested in sponsoring an AI Security Summit edition? Partner with us at the San Francisco flagship event or one of our satellite regional editions, and put your work in front of people who evaluate on technical merit. Email us to register your interest: partners@aisecuritysummit.com

room.audience()

who-should-attend/

The people who own AI risk at the org level: setting budgets, signing off on vendors, and writing the governance playbook in real time.

  • CISOs, CTOs, VPs of Security and Heads of AI
  • Leaders who set risk budgets and choose vendors
  • The people who answer to the board when something goes wrong

topics we'll cover

  1. Trusting what the AI software factory ships
  2. Agent identity, auth, and where to start on agent security
  3. Attacks at machine speed, and what the research shows
  4. Open source and open models in AI-powered defense
  5. How enterprise leaders, investors, and CISOs are weighing AI risk

register.now()

save-your-seat/

in-person/

register/

Registration open

One room, sized for the people who own AI risk. Register to claim your seat at the Westin St. Francis.

What's included

  1. A full day in one room: keynotes, fireside chats, panels, and the people doing the work, up close
  2. Zero vendor fluff: published research and trade-offs from the people who made them

loading registration form

agenda.schedule()

agenda/

The hard calls on AI risk: budgets, vendors, and deciding what your org defends vs. what it accepts.

speakers.featured()

We book speakers who've done the work: made the calls on AI risk at scale, owned the outcomes, and can show what happened next. Our bar is high. More speakers are announced as they're confirmed.

october_15_2026

  • 09:00 – 09:30Breakfast & Registration
  • 09:30 – 10:0030 min
    keynote

    The AI Hurricane Is Here: Cut Through the Opera. Build for the Storm.

    • Snyk
  • 10:00 – 10:2525 min
    keynote

    How to Build Trusted Delivery by the Software Factory

    • Tessl
  • 10:25 – 10:5025 min
    fireside chat

    Open Source at the Inflection Point: A Fireside Chat with Anthropic on AI-Powered Defense

    • Anthropic
    • Snyk
  • 10:50 – 11:00Break
  • 11:00 – 11:2525 min
    panel

    How Enterprise Leaders Are Sequencing AI Investment and Weighing Risk

    • Telus Digital
    • Snyk
  • 11:25 – 11:5025 min
    keynote

    The Criticality of Open Models Has Never Been More Relevant

    • Faisal Tameesh NVIDIA
  • 11:50 – 12:1525 min
    panel

    Investor Panel: What Investors Are Actually Seeing in AI Security

    • boldstart ventures
    • BlackRock
    • Ken Mac
  • 12:15 – 13:15Lunch
  • 13:15 – 13:5035 min
    panel

    AAuth: Moving Beyond OAuth

    • Keycard
    • Keycard
    • hello.co
    • Insecure Agents
  • 13:50 – 14:2030 min
    fireside chat

    Cyber Autonomy: When Attacks Move at Machine Speed

    • Carnegie Mellon
    • Max Baader
  • 14:20 – 14:4525 min
    panel

    The Last Mile of Agentic AI: Securing Where Agents Actually Touch Your Data

    • Cyera
    • Jason Trunk Island
    • Chris Suen
  • 14:45 – 15:00Break
  • 15:00 – 15:2525 min
    keynote

    The Agent Baseline: Where Should You Start?

    • Keycard
    • Docker
  • 15:25 – 15:5025 min
    talk

    The Attack That Only Works Sometimes: Reproducing Probabilistic Prompt Injection

    • Arize AI
  • 15:50 – 16:3040 min
    panel

    CISO Debate

    speakers to be announced

  • 16:30 – 17:301 hr
    networking

    Happy Hour

    Networking before the official AISS after-party.

  • 17:30 – 20:303 hr
    networking

    Official AISS after-party: AAuth Night

    111 Minna Gallery, 111 Minna St, San Francisco, CA 94105. Free. Separate registration on Luma.

venue.location()

venue/

venue.details()

Westin St. Francis

335 Powell St, San Francisco, CA 94102

directions

A San Francisco landmark on Union Square, open since 1904. The event spaces have the bones to match — grand ballrooms with carved ceilings and gilt pillars in the historic Landmark Building, and floor-to-ceiling views of the Bay Bridge and Golden Gate from the 32nd floor. The kind of venue where the room does some of the work for you.

after.hours()

official-after-party/

Official AISS after-party

AAuth Night Moving Beyond OAuth

The summit's official after-party, co-hosted with Insecure Agents. After happy hour, head to 111 Minna Gallery for live demos of AAuth, the agent auth protocol from OAuth's author Dick Hardt, lightning talks, and drinks.

Registration is free and separate from your summit pass.

AAuth Night: Moving Beyond OAuth, from The AI Security Summit. Thursday, October 15, San Francisco, CA, 5:30 PM to 8:30 PM PDT.
date
Thursday, October 15, 2026
time
5:30 to 8:30 PM PDT
hosts
AI Security Summit with Insecure Agents, powered by Keycard
entry
Free. Separate registration on Luma.
Register on Luma

mission.why()

the-mission/

The best AI security knowledge right now lives inside companies. It's in incident reports that nobody publishes, in architecture decisions that never leave the engineering org, in red team findings that stay in internal wikis. AISS exists to get that knowledge on stage.

San Francisco is where it all started. AISS launched here in 2025 and sold out, the city has been its flagship ever since. Satellite editions take AISS to other cities around the globe: Copenhagen and Sydney follow on November 25.

We select talks the way peer review works: show us what you found, how you tested it, and what happened once it hit production.

There's no place for platitudes, no panels where everyone agrees: the tough conversations are welcome here.

That's the deal: practitioners share what they've found, built, broken, or defended. Leaders share the real-life trade-offs behind the decisions they've made. In a room of the people who own AI risk, that level of honesty is worth it.

75+ speakers across five events
500+ attendees builders and defenders
45+ sessions talks, workshops, panels

latest-videos/

workshop

AI Security Summit 2025 - SF - Sizzle Reel

View all videos

register.now()

secure-the-thing-
everyone-is-shipping/

Save your seat