Nvidia investors wiped nearly $600 billion from the chip maker’s value on Monday. Nvidia stock is down 15% on a 5-day rout because of a new Chinese AI company called DeepSeek. And it might be just the beginning.
Nassim Taleb, known for forecasting ‘Black Swans’, told Bloomberg News on Tuesday that Nvidia’s performance is just an idea of what investors who blindly invested in AI are going to face.
But American AI companies are fighting back. OpenAI announced on Thursday that it has signed an agreement with the U.S. National Laboratories to “supercharge their scientific research using our latest reasoning models.”
OpenAI said that sharing its technology with the nation’s top scientists “aligns with our mission of building generative AI to benefit humanity, and we believe the U.S. government is a critical partner to achieve this goal.”
OpenAI is also claiming that DeepSeek, the low cost high-performance Chinese genAI, distilled its knowledge out of OpenAI’s models.
The Chinese startup AI model seems to rival ChatGPT’s performance at a fraction of its costs, making markets question whether the amount of money deployed to OpenAI, Nvidia, and its US peers has really been worth it.
Amid the AI community, however, DeepSeek’s impact on US tech stocks was not taken as a surprise. “DeepSeek is a proof that everything was a matter of architecture and modeling optimization” said Dimas Timmers, an entrepreneur leading research and development in human-centered AI.
He claims that OpenAI’s gigantic machine learning architectures are inefficient in the long run and DeepSeek created a way to use reinforcement learning and latent space.
“They did something really stupid but brilliant, it gave the AI time to think. The result is 95% cheaper, just because it optimized the architecture.
On the other hand, investors are taking a different approach. Alastair Rampell, general partner at Andreessen Horowitz, wrote on a FreePress post that by making its model open-source, the Chinese hedge fund behind DeepSeek “has confirmed how counterproductive the Biden administration’s pro-containment, pro-hegemony, anti-open source AI strategy had been.”
Rather than blaming the US companies’ performance, Rampell recalled that former President Joe Biden issued an executive order “which sought to constrain computers under an arbitrary threshold, bar open source as an alleged threat to national security, and effectively allow regulatory capture by the biggest players.”
On LinkedIn, Jean Boivin, Head of the BlackRock Investment Institute, Global Head of Research at BlackRock, noted that “it is much too soon to conclude that capital spending by U.S tech on the AI buildout is overdone.”
Leave a Reply