Nvidia is committing roughly $6 billion to a sweeping deal with AI startup Poolside, aiming to build a powerful U.S. open‑weight AI model and challenge China’s growing lead in downloadable models.
Nvidia announced a $6 billion investment in a partnership with AI startup Poolside to develop a high‑performance, open‑weight model that can be downloaded and run locally, positioning the United States to compete with China’s rapidly advancing downloadable models such as DeepSeek, Kimi K3 and Qwen.
Why an Open‑Weight Model Matters
Open‑weight models give developers and enterprises the freedom to run large language models on‑premises or in private clouds, sidestepping the data‑privacy concerns and latency issues tied to hosted APIs. By offering a U.S.‑based alternative, Nvidia hopes to retain talent and keep critical AI workloads within domestic infrastructure.
The Poolside Collaboration
Poolside, a startup founded by former Nvidia engineers, will receive the bulk of the funding to scale its existing model architecture, integrate Nvidia’s latest GPU accelerators, and open the model’s weights under a permissive license. The partnership also includes joint research on safety, alignment, and multilingual capabilities.
Strategic Context
China’s AI ecosystem has surged ahead with publicly released models that can be run on commodity hardware, prompting concerns in Washington about a strategic technology gap. The new U.S. model aims to provide a comparable, if not superior, option for businesses that need offline inference.
- DeepSeek – a Chinese open‑weight model known for strong multilingual performance
- Kimi K3 – optimized for low‑latency inference on consumer GPUs
- Qwen – a large‑scale model praised for its coding abilities
Implications for the AI Market
If successful, Nvidia’s venture could spur a wave of domestic startups to adopt open‑weight models, diversify the AI supply chain, and reduce reliance on foreign cloud services. It also signals a broader industry shift toward more transparent and controllable AI systems.
“Open‑weight models are the next frontier for responsible AI deployment,” said a senior Nvidia spokesperson, emphasizing the company’s commitment to democratizing advanced language technology.
The $6 billion commitment underscores Nvidia’s confidence that open‑weight AI will become a cornerstone of future enterprise workflows, especially in regulated sectors such as finance, healthcare, and defense.