Blab AI
SIGGRAPH 2026 Highlights Neural Rendering, GPU Innovation, and Generative AI in Sponsored Keynotes
SIGGRAPH 2026, the 53rd annual conference on computer graphics and interactive techniques, opened on July 19 in Los Angeles and will run through July 23. The event, organized by the Association for Computing Machinery (ACM) and hosted at the Los Angeles Convention Center, is the world’s leading forum for researchers, artists, and technologists to showcase new graphics and AI technologies. The 2026 program includes three sponsored keynotes from NVIDIA, Bolt Graphics, and Tripo AI that focus on neural rendering, next‑generation GPU design, and full‑stack generative 3D AI.
NVIDIA’s keynote, scheduled for Monday, July 20 at 3:45 p.m. PDT, will be delivered by Neil Ashton, Ming‑Yu Liu, and Edward Liu. The speakers will discuss advances in neural rendering, world models, and AI‑driven simulation. Ashton, an engineer at NVIDIA Research, will cover AI physics and agentic AI for computational engineering. Liu, an IEEE fellow leading the NVIDIA Cosmos Lab, will present work on world foundation models and generative AI for physical simulation. Liu, director of applied deep learning research, will highlight NVIDIA’s Deep Learning Super Sampling (DLSS) technology and other real‑time neural rendering research.
Bolt Graphics will present its Zeus GPU on Tuesday, July 21 at 2:30 p.m. PDT. Founder, CEO, and CTO Darwesh Singh will explain how the company is rethinking GPU architecture to support real‑time path tracing and physics simulation. The keynote will cover the move away from traditional rasterization toward a design that prioritizes efficiency and scalability for photorealistic rendering and simulation workloads. Bolt’s approach combines scalar, vector, and specialized processors to deliver high‑performance real‑time path tracing.
The third sponsored keynote, scheduled for Wednesday, July 22 at 2:30 p.m. PDT, will feature Dr. Yanpei Cao and a team of researchers from Tripo AI. The talk will focus on the shift from static appearance generation to interactive 3D intelligence. Dr. Cao, Tripo AI’s chief scientist, will describe how the company is building a full‑stack generative ecosystem that supports gaming, 3D printing, and robotics simulation. The presentation will cover large‑scale architectures, multimodal integration, and the development of functional, agentic systems that can populate dynamic environments.
Conference Chair Chris Redmann said the event continues to be “the premier stage for unveiling the ideas and technologies that define the future of computer graphics and interactive techniques.” He added that the sponsored keynotes reflect the rapid evolution of AI‑driven content creation and the collaboration between industry and research.
SIGGRAPH has been held annually since 1974 and is organized by ACM SIGGRAPH, a special interest group within the ACM. The conference has historically been a launchpad for new graphics hardware, APIs, and research breakthroughs. The 2026 program follows that tradition by presenting cutting‑edge GPU architecture, neural rendering, and generative AI that are already influencing game development, architectural visualization, and robotics.
For creators and developers, the NVIDIA and Bolt Graphics sessions provide insight into how neural rendering and path‑tracing hardware can improve visual fidelity and simulation accuracy. The Tripo AI keynote offers a view of how generative AI can accelerate content creation across multiple industries. Together, the keynotes illustrate how hardware and software innovations are converging to enable more realistic, interactive digital worlds.
The conference is open to the global community of researchers, artists, and technologists. Registration is available through the official site s2026.siggraph.org. Attendees can expect to see demonstrations of NVIDIA’s DLSS 5, Bolt’s Zeus GPU, and Tripo AI’s generative pipelines, as well as discussions on the future of AI‑driven graphics and simulation.
The keynotes underscore the collaborative nature of the graphics and AI ecosystem, but questions remain about how quickly these technologies will be adopted in production pipelines and how they will impact existing hardware ecosystems. The 2026 conference will likely provide further clarification as the industry moves toward more integrated, AI‑powered graphics workflows.
NVIDIA’s keynote, scheduled for Monday, July 20 at 3:45 p.m. PDT, will be delivered by Neil Ashton, Ming‑Yu Liu, and Edward Liu. The speakers will discuss advances in neural rendering, world models, and AI‑driven simulation. Ashton, an engineer at NVIDIA Research, will cover AI physics and agentic AI for computational engineering. Liu, an IEEE fellow leading the NVIDIA Cosmos Lab, will present work on world foundation models and generative AI for physical simulation. Liu, director of applied deep learning research, will highlight NVIDIA’s Deep Learning Super Sampling (DLSS) technology and other real‑time neural rendering research.
Bolt Graphics will present its Zeus GPU on Tuesday, July 21 at 2:30 p.m. PDT. Founder, CEO, and CTO Darwesh Singh will explain how the company is rethinking GPU architecture to support real‑time path tracing and physics simulation. The keynote will cover the move away from traditional rasterization toward a design that prioritizes efficiency and scalability for photorealistic rendering and simulation workloads. Bolt’s approach combines scalar, vector, and specialized processors to deliver high‑performance real‑time path tracing.
The third sponsored keynote, scheduled for Wednesday, July 22 at 2:30 p.m. PDT, will feature Dr. Yanpei Cao and a team of researchers from Tripo AI. The talk will focus on the shift from static appearance generation to interactive 3D intelligence. Dr. Cao, Tripo AI’s chief scientist, will describe how the company is building a full‑stack generative ecosystem that supports gaming, 3D printing, and robotics simulation. The presentation will cover large‑scale architectures, multimodal integration, and the development of functional, agentic systems that can populate dynamic environments.
Conference Chair Chris Redmann said the event continues to be “the premier stage for unveiling the ideas and technologies that define the future of computer graphics and interactive techniques.” He added that the sponsored keynotes reflect the rapid evolution of AI‑driven content creation and the collaboration between industry and research.
SIGGRAPH has been held annually since 1974 and is organized by ACM SIGGRAPH, a special interest group within the ACM. The conference has historically been a launchpad for new graphics hardware, APIs, and research breakthroughs. The 2026 program follows that tradition by presenting cutting‑edge GPU architecture, neural rendering, and generative AI that are already influencing game development, architectural visualization, and robotics.
For creators and developers, the NVIDIA and Bolt Graphics sessions provide insight into how neural rendering and path‑tracing hardware can improve visual fidelity and simulation accuracy. The Tripo AI keynote offers a view of how generative AI can accelerate content creation across multiple industries. Together, the keynotes illustrate how hardware and software innovations are converging to enable more realistic, interactive digital worlds.
The conference is open to the global community of researchers, artists, and technologists. Registration is available through the official site s2026.siggraph.org. Attendees can expect to see demonstrations of NVIDIA’s DLSS 5, Bolt’s Zeus GPU, and Tripo AI’s generative pipelines, as well as discussions on the future of AI‑driven graphics and simulation.
The keynotes underscore the collaborative nature of the graphics and AI ecosystem, but questions remain about how quickly these technologies will be adopted in production pipelines and how they will impact existing hardware ecosystems. The 2026 conference will likely provide further clarification as the industry moves toward more integrated, AI‑powered graphics workflows.
