Nvidia is touting the performance of DeepSeek’s open source AI models on its just-launched RTX 50-series GPUs, claiming that they can “run the DeepSeek family of distilled models faster than anything on the PC market.” But this announcement from Nvidia might be somewhat missing the point.
This week, Nvidia’s market cap suffered the single biggest one-day market cap loss for a US company ever, a loss widely attributed to DeepSeek. DeepSeek said that its new R1 reasoning model didn’t require powerful Nvidia hardware to achieve comparable performance to OpenAI’s o1 model, letting the Chinese company train it at a significantly lower cost. What DeepSeek accomplished with R1 appears to show that Nvidia’s best chips may not be strictly needed to make strides in AI, which could affect the company’s fortunes in the future.
That said, DeepSeek did train its models using Nvidia GPUs, merely weaker ones (H800) that the US government allows Nvidia to export to China. And today’s blog post from Nvidia wants to show that its new 50-series RTX GPUs can be useful for R1 inference – or what an AI model actually generates – saying that the GPUs are built on the “same NVIDIA Blackwell GPU architecture that fuels world-leading AI innovation in the data center” and that “RTX fully accelerates DeepSeek, offering maximum inference performance on PCs.”
But how DeepSeek did its training is part of what has been such a big deal. (And it’s worth noting that China is getting a less powerful version of the RTX 5090.)
Other tech companies are trying to ride the DeepSeek wave, too. R1 is also now available on AWS, and Microsoft made it available on its Azure AI Foundry platform and GitHub this week. However, Microsoft and OpenAI are reportedly investigating if DeepSeek took OpenAI data, Bloomberg reports.
