Nvidia’s annual GPU Technology Conference (GTC) returns to San Jose this week, drawing intense attention from investors, developers, and technology leaders eager to see what the AI chip giant plans next. CEO Jensen Huang is expected to deliver a keynote outlining Nvidia’s roadmap for artificial intelligence hardware and software, as well as new partnerships and emerging technologies. With Nvidia at the center of the global AI infrastructure boom, the announcements at GTC 2026 could provide important signals about the future of computing, data centers, and the rapidly evolving AI ecosystem.

AI Hardware Roadmap Takes Center Stage

Nvidia’s keynote presentations have historically served as launchpads for new chips and platform updates, and this year is likely to continue that tradition. Industry watchers expect updates on Nvidia’s next-generation AI computing platforms, including further details about the Vera Rubin architecture and its successor technologies planned later in the decade.

The company may also provide insights into the future Feynman GPU architecture scheduled for release around 2028. These announcements could reinforce Nvidia’s leadership in high-performance computing as the demand for AI infrastructure continues to expand across cloud providers, enterprises, and research institutions.

Inference Chips and the Groq Partnership

One of the most closely watched developments could involve Nvidia’s strategy for AI inference—the process of running trained AI models to generate responses and insights. Over the past year, Nvidia entered a nonexclusive agreement with chip startup Groq, known for its language processing units designed specifically for AI inference workloads.

Groq claims its processors can run large language models significantly more efficiently than traditional GPUs. Nvidia may reveal how it plans to integrate Groq’s technology into its ecosystem or potentially introduce its own dedicated inference-focused hardware to maintain its competitive edge as AI workloads evolve.

Possible Entry Into the Laptop CPU Market

Another anticipated announcement involves Nvidia’s potential expansion into laptop processors. Reports suggest the company is developing Arm-based chips, known as the N1 and N1X, designed to power Windows laptops. These processors would likely combine Nvidia’s graphics expertise with energy-efficient CPU architecture, targeting gaming and AI-enabled personal computers.

While Nvidia already supplies graphics processors for many gaming systems and previously powered devices such as Nintendo’s Switch consoles, a dedicated CPU line would deepen its presence in consumer computing markets.

Software and the Rise of AI Agents

Beyond hardware, Nvidia is increasingly emphasizing its software ecosystem. At GTC 2026, the company could unveil a new platform designed to support AI agents capable of performing complex tasks across enterprise systems. The rumored service, known as NemoClaw, would allow organizations to deploy intelligent agents that automate workflows and interact with digital infrastructure.

Such an offering would align with the broader industry trend toward agentic AI, where systems move beyond chat interfaces to execute real-world tasks such as managing processes, analyzing data, and coordinating operations across software environments.

Looking Ahead to Physical AI and Robotics

Another major theme expected at the conference is the intersection of artificial intelligence with robotics and physical-world applications. Nvidia has repeatedly emphasized what it calls “physical AI,” which involves using advanced computing systems to train robots, autonomous vehicles, and industrial machines.

According to Huang, these technologies could represent a multi-trillion-dollar market opportunity over the coming decades. Announcements at GTC may therefore include new software frameworks, simulation tools, or hardware platforms designed to accelerate the development of intelligent machines.