.png){kind=logo}
NVIDIA stock edged lower after a Chinese artificial intelligence company, DeepSeek, released a new open-source AI model. The move revived market concerns about rising competition in AI computing and software. Shares pulled back even though NVIDIA recently reported strong quarterly earnings and guidance.
DeepSeek launched its latest open-source large language model soon after Alphabet’s Google rolled out its Gemini 3 system. The Chinese firm already disrupted AI markets in January with a low-cost model that undercut many rivals. Developers can download and adapt DeepSeek’s models from platforms such as Hugging Face and GitHub.
Industry analysts say more efficient open-source models can reduce the computing power needed to train and run large systems. This shift may eventually trim demand growth for the most advanced accelerators, including NVIDIA’s high-end AI chips. Chinese groups such as Baidu, Alibaba, Zhipu AI, MiniMax, and Moonshot AI also continue to release new open models.
Commentators note that Chinese developers now account for a growing share of new downloads for open models. Venture firms also report that many new AI startups pitch products built on Chinese open-source technology stacks. This shift reshapes global open-source dynamics.
US researchers and firms pursue a different response, though one focused on transparency and architecture, rather than scale alone. Projects such as AI2's OLMo 3 and NVIDIA's Nemotron aim to provide fully documented models for enterprise users. Startups like Arcee explore new model designs that balance performance with responsible deployment.
Despite periodic stock pullbacks, NVIDIA continues to report strong demand for its AI data center products. Chief Executive Jensen Huang has pointed to robust interest in the company’s Blackwell and Rubin chip platforms. Large cloud providers such as Amazon, Microsoft, and Google keep increasing capital spending on AI infrastructure.
Analysts also watch developments in China’s AI application programming interface pricing, which trends lower than global averages. Lower prices may accelerate the adoption of open-source AI services and encourage more experimentation.
NVIDIA also deepened ties with leading AI labs, including OpenAI and Anthropic, through multi-billion-dollar investments and supply agreements. These partnerships commit the labs to deploy extensive NVIDIA-powered data center capacity in the coming years. At the same time, Chinese export controls still limit NVIDIA’s sales of advanced Hopper chips into that market.
Investors now monitor how fast open-source models can reduce hardware needs compared with proprietary systems. They also watch whether Chinese open-source adoption slows demand growth for US AI chip makers. For now, NVIDIA remains central to global AI infrastructure, even as the competitive market shifts.
