Google DeepMind and HeyGen to Co-Host Builder Event in LA
Google's official developer account @googleaidevs announced on June 4, 2026 (local time) the opening of registration for an LA builder event co-hosted by Google DeepMind and AI video-generation company HeyGen. Titled "Google DeepMind × HeyGen Hyperframes: Building with Multimodal AI in LA," the event is scheduled for the evening of Thursday, June 11, and targets developers and founders working at the intersection of AI agents, creative tooling, and multimodal apps through demos, conversations, and networking. Attendees can sign up via the Luma registration page, and lightning demo slots—short presentation windows for participants to showcase their work—are also open. The announcement post was echoed in HeyGen's own post with the same link.
The venue is in Los Angeles (listed near 12130 Millennium Dr.), and the program consists of demos, conversations, and participant-led lightning demos. Concrete figures such as pricing or benchmarks have not been disclosed at this announcement stage, and details are to be confirmed on the registration page.
The collaboration draws attention because it pairs the research-driven Google DeepMind with the product-driven HeyGen, two parties with distinct strengths working together directly. Google DeepMind has led advanced research including the Gemini series and the world model Genie. HeyGen, known for AI avatar video generation, was named to the Forbes AI 50 for 2026. HeyGen recently released the open-source video rendering framework HyperFrames, and this event aims to foster a practical community around "multimodal + agents." It can also be seen as a developer-focused follow-up after Google I/O 2026 in May.
HyperFrames is a framework in which AI agents such as Claude Code write HTML/CSS/JS to render video—an HTML-native approach contrasted with the React-based Remotion. Released under the Apache 2.0 license on GitHub, it can be installed with npx skills add heygen-com/hyperframes and is said to enable deterministic MP4 rendering. In the AI video space, competitors include Runway, Pika, Synthesia, and OpenAI's Sora line, but a direct co-hosting by HeyGen and DeepMind is rare.
As it is just after the announcement, large-scale reactions remain limited, but on X one user said the overlap of multimodal and agents is what they most want to see, and that they were curious whether there would be a full-pipeline live demo, while others expressed that seeing both companies demo agents and multimodal tools in the same room was worth a trip to LA. Within the technical community, use cases of HyperFrames—agents writing HTML to produce MP4, with deterministic rendering as a strength—are also growing. Hands-on user testimonials, including criticism, remain scarce, so a fuller assessment will come after the event. The latest details can be confirmed on the registration page.