What to Expect at AMD Advancing AI 2026

AMD Advancing AI 2026: What to Expect in San Francisco

Advancing AI 2026, the annual AMD gathering for the global AI community, takes place July 22–23 in San Francisco. It brings developers, customers, and channel partners together to experience the latest in AI infrastructure, architecture, and development, with dedicated programming for each audience. Here is what the agenda holds for enterprise decision-makers and developers alike.

How the two days are shaped

The event is built around the keynote, expo, and 100+ sessions including: breakouts, tech talks, workshops, meet-the-experts, and theater sessions.

Wednesday, July 22 opens with a Luminary session before moving into the three concurrent session formats in the morning. The afternoon features the Expo, designed to foster meaningful engagement for developers, IT professionals, and researchers followed by a welcome reception to close the day. Thursday, July 23 kicks off with a Keynote from Dr. Lisa Su, followed by more tech talks, certifications, breakouts, and the Expo, wrapping with a closing reception.

Enterprise and infrastructure leaders can look forward to main stage sessions and breakouts on AI strategy, infrastructure direction, and market trends, while Developers will experience hands-on workshops and labs built around real implementation work. Between sessions, both audiences share the same hallways, partner showcases, and peer conversations, which deliver some of the most valuable exchanges of the event.

For enterprise and infrastructure leaders

Sessions for infrastructure and operations leaders focus on outcomes, not just silicon. Attendees will hear from a variety of AMD executives, industry leaders from Dell, Gartner, Google, Lenovo, Nutanix, TensorWave, and more, as they address the questions enterprise teams work through every day: how to deploy AI infrastructure and agent computers for agentic AI, inference, and sovereign AI, the true total cost of ownership behind AI infrastructure, the power and density constraints that can break a business case, and how to scale from clusters to full "AI factories" where network efficiency matters as much as raw compute.

The Expo delivers a second layer of value for this audience. AMD and partner booths are organized by solution area, so you can move directly to the workloads that matter to you, see solutions side by side, and discuss production use cases with the teams that built them. Expo Theater Sessions and The Cube add live, stage-side content on the floor, and Welcome and Closing Receptions create room to connect with customers and partners further along the same path. The show floor and hallway conversations are sure to be as engaging and informative as the sessions themselves.

Plan your agenda in advance here.

For developers and practitioners

Developers who develop, optimize, and deploy AI models and applications get a dedicated track. Expect workshops and technical deep dives spanning AI inference, training, deployment, optimization, open tooling, and production-scale infrastructure — with direct access to the engineers building it.

The workshops are implementation-focused and span experience levels. Build a multi-modal agent from scratch in Build Your OpenClaw Agent with Multi-Modal Models, architect a hybrid system that routes work between a Ryzen™ AI PC and AMD Instinct™ GPUs in Building Hybrid Multi-Agent Systems from Client to Cloud, or run vision-and-language models for robotics on a Ryzen™ AI iGPU in Enabling Physical AI on AMD.  

Pre-registered attendees will receive priority entry. Walk-ins are welcome beginning 5 minutes before each session, on a first-come, first-served basis, until the room reaches capacity. 

Developers can also earn a credential on-site. The four-hour ROCm Certification workshop pairs three hours of hands-on training with a one-hour exam and qualifies toward Associate Level in the AMD ROCm™ Certified Developer Program. Guided labs cover ROCm fundamentals, AI and HPC development, PyTorch, HIP programming, CUDA porting, and more.

Between sessions, the Developer Zone keeps the momentum going. Drop into labs (no scheduling required), visit focused stations about AI agents, model fine-tuning, and hybrid and physical AI, catch lightning talks with open-source contributors, and connect with other developers in the lounge. 

Throughout the event, AMD luminaries and industry pioneers including leaders from the Linux Foundation, PyTorch Foundation, the vLLM and SGLang projects, and Stanford research labs will share production lessons from the front lines of AI. 

Explore the full lineup and build your agenda on the developer sessions page. 

Building what’s next in AI — together

Advancing AI puts the people shaping AI in one place: AMD experts, ecosystem partners, and peers already running at scale. Decision-makers will leave with architecture guidance and cost clarity. Developers will leave with new skills ready to apply and a community to build alongside. For both audiences, the event offers a direct view of where AI compute is heading and the tools and connections to act on it.

We can’t wait to see you there!