GMI Summer Signal '26 - The Modern Renaissance - From multimodal to agentic
Multimodal → Agentic AI, live.
What happens when image, video, voice, world models, and agents stop being separate demos — and become one production stack?
At GMI Summer Signal ’26, we’re bringing together the teams building across the multimodal + agentic AI stack for an afternoon of live product demos, technical showcases, launches, and conversations around what is actually working in production today.
Featuring builders from GMI Cloud, MiniMax, Inworld AI, Luma AI, Utopai Studios, Reactor, Tripo AI, Resemble AI, LlamaIndex, and more.
More at Summer Signal '26
The Modern Renaissance
The first Renaissance broke the walls between art, science, and engineering. Leonardo da Vinci showed one mind can paint, dissect, design machines, study motion, and imagine new systems of the world.
Multimodal AI is that convergence returning — vision, language, sound, motion, reasoning, and now action, coming together in agentic systems.
For one day, GMI Cloud is turning the Exploratorium into a digital gallery for the multimodal stack: live models, demos, launches, and workflows showing how these capabilities connect in practice.
Agenda
3:00 PM · Enter the Gallery
The Modern Renaissance Gallery live on screens. Bring your own artifact, or start building one in the room.
Chapter I: Seeing
4:00 PM · Opening Keynote — Alex Yeh, CEO, GMI Cloud
The Multimodal Stack, Rebuilt for Agents
Launching GMI’s new Model-as-a-Service platform and MCP tooling, with a live look at how developers can discover, connect, and run multimodal models as part of agentic workflows.
Chapter II: Speaking
4:30 PM · Inworld AI — Igor Poletaev, CSO
From AI Models to Living Characters
How real-time voice, intelligence, and interaction are reshaping consumer AI experiences.
Chapter III: Moving
From static generation to cinematic performance, continuity, and story
4:45 PM · GMI Cloud — Haomin, Inference PM
What It Takes to Run Frontier Video Models at Scale
A look under the hood at GMI Prime Inference and the infrastructure behind high-performance multimodal generation.
5:00 PM · MiniMax — Victor S., Global Product Marketing Manager
MiniMax H3 Live: The Next Leap in AI Video
A live demo of H3 and what the newest generation of video models can now create.
5:15 PM · Utopai Studios — Jie Yang, CTO
From Generation to Direction: Agentic Filmmaking with PAI 3.0
How AI systems can move beyond creating shots to planning and directing complete cinematic workflows.
Gallery Break
watch the gallery evolve throughout the museum.
Chapter IV: Worlds
From generated media to generated environments that models can understand, simulate, and act inside.
6:00 PM · Reactor — Ahmed Ahres, Head of GTM
The future with real-time video & world models
What changes when generative video becomes fast enough to interact with, steer, and build products around.
6:15 PM · Runway — Jimei Yang, Director of Research
Building real-time interactive videos models: characters, worlds and interfaces.
How real-time generative video is evolving from passive outputs into interactive systems — with controllable characters, persistent worlds, and new interfaces for creation
6:30 PM · Tripo AI
From Image to Interactive 3D World
A live look at turning visual inputs into generated 3D assets and environments.
6:45 PM · Resemble AI — Tedi Papajorgji, CTO
How We Build Detection Models for a Threat That Never Stops Growing
How Resemble built a multimodal detection system that keeps up with rapidly evolving generative models — across audio, image, and video at production latency.
Chapter V: From Multimodal to Agentic
Agents step in to orchestrate the stack — choosing models, calling tools, and turning a single brief into a working pipeline.
7:00 PM · GMI Cloud — Mingjun Sun, PM
The Multimodal Renaissance Is Here. Can Your Product Keep Up?
Execution Environments for Multimodal Agents at Scale
7:15 PM · Luma AI — Kyt Janae, Forward Deployed Creative
Demo: Build a World in 15 mins with Luma
A live demonstration of how generative models can move from an idea to a complete world.
7:30 PM · LlamaIndex — Simon Suo, CTO & Co-founder
What Happens After You Generate?
The understanding and agentic layer beneath multimodal systems — how models retrieve context, reason, choose tools, and take action.
Chapter VI: The CONVERGENCE
7:45 PM · Panel: From Hollywood to AI-Native Entertainment — Shu Studio, Real Reel
Steve Bannerman, former Head of International VFX & Head of Post Production, Amazon MGM Studios
Ali Albazaz, CEO, Inkitt
Gray Crawford, Interaction Designer & Artist, Luma Labs
Yujing Qian, VP of Engineering, GMI Cloud
Moderated by Celine Zen, Founder, SHÙ
8:30 PM · SCALE Cohort 2 Reveal & Closing Reception
Who should join
Summer Signal is built for the people shaping the multimodal and agentic era:
Founders & engineers building AI-native products
AI/ML & infrastructure leaders — CTOs, VPs of AI, Heads of ML, Platform & Inference
Enterprise teams deploying multimodal or agentic AI in production
Developers & API builders scaling inference
Creators & studios working at the frontier of AI content
Investors & operators tracking what's next
About GMI Cloud
GMI Cloud is an AI-native inference platform and one of only seven NVIDIA Reference Platform Cloud Partners globally. We provide the GPU infrastructure and a library of 170+ AI models across LLM, video, and audio, optimized for real-time, production multimodal workloads.