Nvidia is developing a system-on-chip (SoC) for Windows PCs, with Dell and Lenovo among the OEMs planning to build notebooks and desktops around the processor later this year. The chip would be based on the GB10, which Nvidia developed with MediaTek and launched in October 2025 , the Wall Street Journal reported , citing people familiar with the matter. The GB10 currently powers Linux-based AI workstations from Dell, Lenovo, Asus, MSI, and Gigabyte, priced between $3,000 and $4,000 and aimed at machine learning researchers and AI model developers. None of those systems runs Windows. Nvidia has previously explored Arm-based chip designs for Windows PCs, and the GB10 represents its most concrete step yet toward that goal. The engineering gap The GB10 was built for a specific purpose: running large AI models in a developer workstation environment. It pairs a MediaTek-designed CPU tile featuring 20 Arm cores with an Nvidia Blackwell GPU tile, delivering up to one petaFLOP of AI performance at FP4. At 140 watts under full load, roughly three times the thermal budget of a high-end business laptop, it was never designed for standard notebook form factors. A PC-grade derivative would need to ship at significantly lower compute configurations, said Rishi Padhi, principal analyst at Gartner. “I would expect Nvidia’s product line to come with lower compute power to be more efficient on power consumption and also the heat generated, sufficient to be cooled by a laptop-based cooling system,” he said. Nvidia would also need to leverage a unified memory architecture to bring power consumption within range of competing platforms, “similar to how Apple achieves power efficiency gains through a unified memory architecture,” Padhi added. Should Nvidia bring a power-efficient, Windows-ready derivative to market, analysts said it would pose a direct competitive challenge to the vendors that currently dominate PC silicon. Threat to incumbent chipmakers Nvidia’s entry into integrated PC silicon poses a direct competitive challenge to Qualcomm, Intel, and AMD, according to analysts. The sharpest near-term pressure falls on Qualcomm. Windows-on-Arm laptops have historically traded strong battery life for weaker graphics performance, a trade-off Nvidia’s architecture is designed to eliminate. “By fusing the performance compute of a Blackwell GPU directly onto an Arm die, Nvidia essentially nullifies the traditional trade-off associated with Windows-on-Arm machines,” Padhi said. “This will directly impact the share of Qualcomm’s offerings in AI PCs.” For Intel and AMD, the threat operates differently, the analyst said. High-end mobile processors from both vendors are routinely paired with discrete Nvidia GPUs in premium laptops. “With Nvidia slated to launch a capable SoC of its own, it can prompt OEMs to consider purchasing a single integrated chip from Nvidia that offers better battery life and equivalent graphical performance to an x86 CPU plus Nvidia GPU combo,” Padhi said. “The economic and engineering incentives heavily favor the unified system-on-chip.” Shreeya Deshpande, senior analyst at Everest Group, said the move signals a broader strategic shift. “Nvidia’s reported move into Arm-based PC SoCs marks a strategic shift from being primarily a GPU supplier to competing at the core platform level in Windows laptops,” she said. “If successful, it could materially intensify competition in the Windows-on-Arm ecosystem and increase pressure on incumbent PC silicon providers to further differentiate on AI performance and efficiency.” The stacks are significant. AI PCs are on track to account for more than 50% of all PC shipments by 2026, according to Gartner. Beyond the competitive market dynamics, analysts said Nvidia’s entry into PC silicon carries a distinct implication for enterprises already running Nvidia infrastructure in the data center.