2025 marked a breakout year for Artificial Intelligence development on PC, with PC class small language models improving accuracy by nearly 2x over 2024 and narrowing the gap with frontier cloud based large language models. Artificial Intelligence PC developer tools such as Ollama, ComfyUI, llama.cpp and Unsloth have matured, their popularity has doubled year over year and the number of users downloading PC class models grew tenfold from 2024. This momentum is driving a shift toward running increasingly capable generative Artificial Intelligence workloads locally on consumer hardware instead of relying solely on cloud services.
These advances are paving the way for generative Artificial Intelligence to gain broad adoption among everyday PC creators, gamers and productivity focused users. In that context, Nvidia is using CES to introduce a series of Artificial Intelligence focused upgrades for GeForce RTX, Nvidia RTX Pro and Nvidia DGX Spark systems that are designed to provide the performance and memory headroom developers need to deploy generative Artificial Intelligence models directly on PCs. The company is especially emphasizing accelerated 4K video generation, enabled by enhancements to tools like LTX 2 and ComfyUI that integrate more tightly with RTX hardware.
The announced upgrades are positioned as a foundation for a new wave of on device generative Artificial Intelligence applications, from richer creator workflows to interactive gaming experiences and faster content pipelines. By coupling improved small language models with better optimized developer tools and RTX accelerated compute, Nvidia aims to make high resolution Artificial Intelligence video generation and other demanding workloads feasible on a wide range of PCs. The focus on GeForce RTX, RTX Pro and DGX Spark indicates a strategy that spans from enthusiast desktops and workstations to higher end Artificial Intelligence development platforms, all tuned to capitalize on the rapid growth in PC class model downloads and usage.
