Founder narrative locked before the week.
Deliverables (3)
- ▸Pre-event
- ▸Account map
- ▸Intro engineering
Two days with the people shaping frontier AI, from the teams building next-generation chips to the labs training open and frontier models and the engineers running them in production. FORKOFF is a community partner. Use FORKOFF community code COMM20-E0C05EA for 20% off an Expo Pass.
Most teams buy a ticket and hope the right inference buyer walks past the expo floor. Some post once from the venue and go quiet. FORKOFF runs SGLang Summit week as an operating cycle that keeps working after the doors close.
Speaker affiliations as listed by the organiser for SGLang Summit 2026.
Inference, model and hardware buyers concentrated in one two-day San Francisco window, inside a builder community that judges a product by whether it runs.
Inference engineers, model labs, AI hardware teams and AI infrastructure founders in one venue.
Serving, hardware and model-platform decisions get discussed in the same two-day window.
The summit grows from an open-source inference project, so practitioner credibility moves fast.
Talks and demos keep circulating for weeks when they are captured and routed after the event.
A main stage split across two themed days, breakout rooms, a hands-on lab, and a GPU Expo floor, all inside Fort Mason Center on the San Francisco waterfront.
Nov 12 to 13, 2026
Fort Mason Center, 2 Marina Blvd, San Francisco. Main stage, breakout rooms, a hands-on lab, the GPU Expo and exhibitor booths, plus an Open Stage and community meetups between sessions.
2,000-plus engineers, researchers, builders and leaders, according to the organiser. The room spans inference and serving engineers, model labs, AI hardware teams and founders shipping on top of them.
Day one leads with infrastructure and models; day two turns to agents and applications.
Models, reinforcement learning, AI hardware, cloud infrastructure, generative media, evaluation, local AI, robotics and world models, the production stack, and recommendation systems.
Founder narrative locked before the week.
Deliverables (3)
Daily founder posts from Fort Mason across both days, timed around the main-stage themes that matter to the product..
Deliverables (3)
Founder interviews and demo moments recorded during the week feed the clipping pipeline, cut for short-form across every platform..
Deliverables (3)
The clip cycle runs for about 30 days after the summit, and qualified intros convert into follow-up calls recorded at the T+30 close..
Deliverables (3)
Narrative before the week, cadence across the two days, and a clip cycle afterwards. Each phase closes with a verified entry.
Sample cadence for an on-site activation around the two summit days.
| Day | Date | Motion | FORKOFF owns |
|---|---|---|---|
| Wed (D-1) | Nov 11, 2026 | Land in SF · last-mile narrative cadence · intro confirmations | Intro routing + post drafting |
| Thu | Nov 12 | Summit day 1 · infrastructure, models and AI hardware on the main stage | Daily post + on-site capture |
| Fri | Nov 13 | Summit day 2 · agents, applications and physical intelligence | Daily post + founder interview capture |
| Sat (D+1) | Nov 14 | Founder recap thread · first clips cut | Thread craft + clip pipeline |
| Following weeks | Nov 15 onward | Clip cycle across X, TikTok, Reels and Shorts | Distribution sequencing + weekly record |
A pre-event narrative and an intro plan across the inference and AI infrastructure accounts attending SGLang Summit, delivered before the week starts. Selective on fit.
Every tier carries the verified record. Tier scales scope, not the approach.
by application
By application
By application
SGLang Summit 2026 is a two-day conference by LMSYS, the team behind the open-source SGLang inference framework for large language models. It brings the whole AI stack to one stage: the teams building next-generation chips, the labs training open and frontier models, the engineers running them in production, and the founders putting them to work. The organiser expects 2,000-plus engineers, researchers, builders and leaders.
SGLang Summit 2026 runs November 12 and 13, 2026, at Fort Mason Center, 2 Marina Blvd, San Francisco, CA. Registration is through the organiser's Luma page at luma.com/sglang-summit, and the full program and FAQ live at sglang.io/summit.
Yes. FORKOFF is a community partner of SGLang Summit 2026. Use FORKOFF community code COMM20-E0C05EA for 20% off an Expo Pass at checkout on the Luma registration page. The organiser also offers a separate developer and education rate on the Expo Pass for developers, students and researchers.
The organiser lists Lip-Bu Tan (Intel), Lilian Weng (OpenAI), Soumith Chintala (Thinking Machines Lab) and Aravind Srinivas (Perplexity) among the speakers, with more to be announced. Main-stage sessions cover the SGLang ecosystem, open models, recursive self-improvement, AI hardware, AI agents, emerging AI labs, AI applications, and physical intelligence.
The Expo Pass covers the breakout rooms, the hands-on lab and the expo floor, where the GPU Expo and exhibitor booths run live demos. The Full Pass adds the main-stage talks. At the time of writing the organiser showed the Full Pass as sold out, so check the Luma page for current availability.
FORKOFF runs event distribution for AI infrastructure and developer-tool founders, and SGLang Summit is one of the densest rooms of inference, model and AI hardware builders in the calendar. As a community partner we help builders reach the event, and for teams that want the week to keep working we run an activation: pre-event narrative, on-site cadence, and a post-event clip cycle.
Yes. An activation is a pre-event narrative and intro plan across the inference and AI infrastructure accounts attending, on-site cadence across the two days, and a post-event clip cycle that keeps the week's talks circulating for about 30 days. Scope is set by application, from a single pre-event narrative to a full on-site week.
The AI-side playbook for inference, serving and AI infrastructure teams beyond one event week.
How FORKOFF runs event weeks end to end, from pre-event narrative to the post-event clip cycle.
How talks and demos captured at SGLang Summit become a month of short-form distribution.
The June San Francisco builder week for agentic AI and AI infrastructure founders.
Every event week on the FORKOFF roster, so a founder can sequence the year.
The San Francisco GTM playbook and on-ground partner network beyond summit week.
As Featured In
Full press shelf







SGLang Summit 2026 is a two-day conference by LMSYS, the team behind the open-source SGLang inference framework, held November 12 and 13, 2026 at Fort Mason Center in San Francisco. It gathers 2,000-plus engineers, researchers, builders and leaders across AI hardware, open and frontier models, production inference and AI applications. FORKOFF is a community partner, and code COMM20-E0C05EA takes 20% off an Expo Pass.
SGLang Summit grows out of an open-source inference project, so the people in the room are the engineers who run models in production, the labs that train them, and the hardware teams that serve them. That audience checks a product by whether it holds up under a real technical question. Two days at Fort Mason reward a founder who arrives with a plan for who to meet and what to show, then keeps the conversation going in public after the summit ends.
As a community partner, FORKOFF shares a code that takes 20% off an Expo Pass. Enter COMM20-E0C05EA at checkout on the organiser's Luma registration page. The Expo Pass covers the breakout rooms, the hands-on lab and the expo floor. The organiser runs its own developer and education rate as well, and the full program and FAQ are on the SGLang Summit event page at sglang.io/summit.
SGLang Summit runs on the same operating spine as the rest of the FORKOFF calendar. The event management service covers the pre-event narrative, the on-site cadence and the capture that feeds a month of clips. The 2026 events calendar maps every week we run, and for the June builder week in the same city the AI Engineer World's Fair page covers the agentic AI and AI infrastructure room.
Reviewed by the FORKOFF events team, the operators who run these activations on the ground.
Use FORKOFF community code COMM20-E0C05EA for 20% off an Expo Pass. Planning to make the week count for your company? Apply for a FORKOFF activation around SGLang Summit.

A recap of AI, DePIN and RWA Builders, the evening side event FORKOFF hosted with Taiko and io.net during TOKEN2049 week in Singapore: panels, partners, data.

A recap of AI, DePIN and RWA Day, the half-day side event FORKOFF co-hosted with io.net and FailSafe in Seoul during KBW2025: hosts, agenda and real numbers.

A media partner recap of India Blockchain Week 2025 in Bengaluru: the conference, Demo Day, sponsors, the numbers from X and YouTube, and lessons for 2026.