Nvidia's GPU Technology Conference runs Monday through Thursday in San Jose, with CEO Jensen Huang delivering his keynote at the SAP Center on Monday at 11 a.m. PT. For consumer gamers hoping to see new gaming graphics cards, the advice is simple: look elsewhere.
Nvidia has announced it will not release new gaming GPUs in 2026, focusing instead on chips for AI data centres rather than consumer graphics cards. This marks a decisive strategic turn. The world's most valuable company is choosing to build data centre dominance over consumer market reach, a signal that the real money and influence lie in enterprise compute infrastructure, not desktop gaming rigs.
The conference reflects a broader transformation in Nvidia's business. In the latest quarter, Nvidia generated data centre revenue of over $62 billion, up 75% from a year earlier. That revenue stream now dwarfs consumer segments entirely. Nvidia is expecting over 30,000 attendees arriving from 190 countries for what has become less a developer gathering and more a corporate forum for negotiating AI infrastructure deals.
The centrepiece of GTC is topics ranging from physical AI and AI factories to agentic AI and inference. Agentic AI—systems designed to carry out multi-step tasks autonomously—is receiving particular emphasis. Jensen Huang is apparently quite fond of the agentic framework, reportedly describing it as the "most important software release probably ever". This reflects genuine technical interest rather than marketing; the shift toward autonomous agent systems does require different computational architectures.
Nvidia's Rubin GPU packs up to 288 GB of HBM4 memory and represents a major performance uplift over current Blackwell-generation parts, delivering 5x the dense floating point throughput. Industry observers expect detailed discussion of this architecture, along with hints about Feynman, Nvidia's next-generation GPU scheduled for 2028, with new power and cooling targets likely exceeding a megawatt per rack.
Perhaps more significantly, Nvidia is poised to unveil new details about its agentic-optimised CPUs, with a CPU-only rack likely to appear, as CPUs are becoming the bottleneck in growing out AI and agentic workflows. This is a notable departure. Bank of America predicts the CPU market could more than double, from $27 billion in 2025 to $60 billion by 2030.
On the software side, Nvidia could debut its own platform for AI agents, called NemoClaw, which would allow companies to deploy agents across their systems. Software announcements carry strategic weight because they lock customers into Nvidia's ecosystem. Attendees should also expect clarity on what the company plans to do with its relationship with Groq, the inference company Nvidia reportedly paid to license its technology, with Groq's founder and president agreeing to join Nvidia to help advance and scale that technology.
The conference also signals Nvidia's commitment to physical AI. Since announcing its Isaac GR00T robotics platform nearly two years ago, Nvidia has launched a steady supply of new toolkits, frameworks and hardware development platforms aimed at giving generative AI physical form, with the Omniverse digital twin platform set to make another appearance.
The business case for Nvidia's strategy is sound from a shareholder perspective. Enterprise infrastructure generates higher margins and longer customer lock-in than consumer products. Data centre customers buy complete systems and stick with vendors that manage their technical complexity. Consumer gamers, by contrast, shop aggressively on price and performance. Once Nvidia cedes that ground, winning it back becomes harder. The company is reading the market correctly: the real business opportunity is selling compute capacity to companies building AI systems, not graphics cards to individuals.
Yet the choice reveals important trade-offs. Consumers lose access to next-generation gaming hardware, and Nvidia's cultural identity shifts from a consumer brand to an infrastructure behemoth. That carries risks. Hardware cycles matter; if a competitor eventually offers consumer gaming GPUs that perform well, Nvidia's consumer mindshare erodes. For now, though, the company's dominance in both GPU training and inference makes such competition unlikely.
GTC 2026 will confirm what has become increasingly clear: Nvidia is no longer a gaming company that sells AI chips on the side. It is an infrastructure company that has consigned gaming to secondary status. Whether that proves wise depends on how successfully the company executes its enterprise strategy and how deeply agentic AI penetrates the market. The conference will provide the first real evidence.