Nvidia’s annual GTC conference, widely regarded as the company’s most important event of the year, is set to begin Monday in San Jose, California. The event will open with a keynote address from CEO Jensen Huang.
The keynote is scheduled for 2 p.m. ET at the SAP Center in San Jose, where Huang will outline Nvidia’s strategic priorities and technological roadmap for the coming year. Developers, industry analysts, and members of the media typically look to the presentation for insights into the company’s latest innovations and future direction.
Huang, well known for delivering presentations in his signature leather jacket, often uses the GTC stage to introduce a series of new products and platform updates. This year’s keynote is expected to follow the same pattern, with multiple announcements likely spanning hardware, software, and artificial intelligence infrastructure.
Over the past several months, Nvidia has been actively expanding its partnerships through a range of agreements with semiconductor and software companies. Observers expect the company to provide further details on how these newly acquired technologies and collaborations will be integrated into Nvidia’s broader ecosystem.
Among the company’s notable moves in the past year was a nonexclusive partnership with chip startup Groq announced in December. The agreement allows Nvidia to utilize Groq’s inference-focused chip technology. Nvidia also brought several key Groq executives into its ranks, including founder Jonathan Ross and president Sunny Madra.
Groq develops processors known as language processing units (LPUs), designed specifically for AI inference—the process of running trained models in real-world applications. According to the company, its chips can execute large language models and other AI workloads as much as ten times more efficiently than conventional GPUs.
As the artificial intelligence industry increasingly shifts its focus from training models to deploying them in practical environments, demand is rising for hardware optimized for inference. Businesses are searching for solutions that can deliver strong performance while keeping operational costs in check. Nvidia has emphasized the efficiency of its GPU architecture, but incorporating Groq’s technology—or unveiling a dedicated inference processor—could further address concerns that specialized chips might challenge the company’s dominance in the AI hardware market.
Another potential highlight at GTC could be the introduction of Nvidia’s long-speculated laptop CPU. Reports from The Verge suggest the company may unveil two Arm-based processors, known as the N1 and N1X, intended for Windows laptops.
These chips are expected to use Arm architecture similar to processors from Qualcomm but with a strong emphasis on gaming performance. Given Nvidia’s reputation in the gaming industry through its widely used GPUs, a gaming-focused CPU could attract significant interest among PC enthusiasts.
Nvidia already supplies the chips used in Nintendo’s Switch and Switch 2 gaming consoles and has participated in various computing platforms in the past, making a move into laptop processors a logical extension. Such a product could also reinforce the company’s relationship with the gaming community as Nvidia continues to prioritize its rapidly growing data center business.
However, even if introduced, laptop CPUs are unlikely to rival the revenue generated by Nvidia’s core data center and GPU segments. In 2025, Nvidia reported $22.5 billion in revenue from its gaming division, compared with $193.5 billion generated by its data center operations.
Beyond potential hardware launches, Huang is also expected to discuss Nvidia’s next-generation AI platforms. This may include additional updates on the upcoming Vera Rubin architecture as well as the more advanced Vera Ultra platform planned for release in the second half of 2027. The company may also share new details regarding the future Feynman GPU architecture scheduled for 2028.
On the software front, Nvidia could introduce a new platform designed for AI agents. According to Wired, the service—reportedly called NemoClaw—would allow organizations to deploy autonomous AI agents across their digital systems.
Nvidia has increasingly emphasized its software capabilities alongside its hardware offerings, promoting open-source AI models and frameworks used in areas ranging from physical simulation to autonomous vehicle development. Expanding further into AI infrastructure with a platform dedicated to agents would align with the growing industry focus on autonomous AI systems.
In addition to these announcements, observers expect a range of updates related to Nvidia’s broader AI strategy. This could include developments in physical AI and robotics—fields that Huang has repeatedly highlighted as representing a potential multi-trillion-dollar market opportunity in the years ahead.
