Nvidia is stepping up to ease the jitters in the market surrounding the Chinese AI startup DeepSeek. They’re pointing out that even though DeepSeek has made some impressive strides, it still heavily relies on “a significant number of NVIDIA GPUs” to make everything work. This comes after Nvidia experienced a jaw-dropping loss of over $600 billion in market value just on Monday alone. While they admit that DeepSeek’s advancements are “excellent,” they want to make it clear that the technology isn’t a standalone miracle—it needs their hardware to truly function.
In a statement, Nvidia explained that DeepSeek’s innovations showcase how new models can be developed using something they call “test time scaling,” which cleverly uses common tools and is compliant with export controls. But they made a point to stress that to get those AI models running, you still need a solid number of their powerful GPUs and top-notch networking capabilities. This statement followed the buzz around DeepSeek launching an AI model that’s reportedly matching the performance of major US tech players while using fewer chips. The news sent Nvidia plummeting, marking the worst drop in their history and sending shockwaves through the tech sector.
Nvidia is framing DeepSeek’s breakthrough within familiar territory, describing it as a prime example of Test Time Scaling, while noting that traditional scaling methods, like pre-training and post-training, are still very much in play alongside this new technique. Their move to calm the market follows some analysts raising alarms about potential risks to the US’s lead in AI. Goldman Sachs even hinted at possible “spillover effects” from any tech sector mishaps impacting the bigger market picture. While Nvidia shares did stabilize a bit in the afternoon, they were still on track for their roughest day since March 2020, when the pandemic sent markets into a tailspin.
