Nvidia Unveils DeepSeek R1 AI Model, API Access Coming Soon

Nvidia Corporation (NASDAQ: NVDA) announced that China’s DeepSeek R1 AI model is now available as a preview on its enterprise software platform. Developers will soon gain access to its API, Nvidia confirmed in a statement.

DeepSeek R1 is accessible through Nvidia Inference Microservice (NIM), enabling developers to run AI applications locally while leveraging Nvidia’s remote hardware. Nvidia highlighted that the DeepSeek-R1 NIM microservice delivers up to 3,872 tokens per second on a single NVIDIA HGX H200 system. API support will be introduced as a downloadable NIM service soon.

DeepSeek R1 has rapidly gained traction, demonstrating capabilities comparable to larger AI models like OpenAI’s Chat-GPT, despite running on older hardware and a significantly lower budget. This efficiency has sparked concerns over AI infrastructure investments, leading to a $600 billion drop in Nvidia’s market valuation this week.

Investors fear that companies may pivot towards leaner, cost-effective AI models, reducing the need for expansive data center infrastructure, which has been a major driver of Nvidia’s growth. The chipmaker has dominated the AI market with its high-performance processors, seeing its valuation soar amid rising AI adoption.

With DeepSeek R1’s release, the industry faces a potential shift towards efficiency over high-end infrastructure, raising questions about future demand for Nvidia’s cutting-edge chips.