DeepSeek fries AI, chip-making shares with Nasdaq 100 futures down 4%, Nvidia futures fall by over 10%

The selloff reflected in the Nasdaq 100 futures too, taking the index down between 3-4% in pre-market trading, possibly indicating at a massive wealth erosion in global equities across US and Europe.

The global AI and semiconductor stock boom is set for a shakeup, as evident by a more than 10% selloff in Nvidia shares during pre-market trading on Nasdaq on January 27, spooked by Chinese AI startup DeepSeek's emerging as a cost-effective player running of less-advanced chips that US peers.

The selloff reflected in the Nasdaq 100 futures too, taking the index down between 4-5% in pre-market trading, possibly indicating at a massive wealth erosion in global equities across US and Europe.

With investors looking to take some profit off leading tech sector names, CBoE VIX, the measure of volatility, has spurted sharply to hit the highest level since December 20, 2024.

Read More: Moneycontrol Pro Panorama | The AI Slingshot

On the Euronext exchange in Amsterdam, shares of global chip-component making monopoly ASML were trading more than 10% lower in mid-day trade. This is the biggest selloff on ASML shares since October 2024. On the German Xtra index, run by the Frankfurt Stock Exchange, shares of Siemens Energy AG were lower by more than 20%, as companies related to electrification too were seen under pressure. Shares of SoftBank - the entity funding as much $500 billion in AI infrastructure - fell 8% in Japan this morning. SoftBank is also the leading shareholder of chip designer company ARM.

Read More: DeepSeek R1 Explained - China’s ambitious answer to ChatGPT and Gemini

Venture capitalist Chamath Palihapitiya wrote on his X handle that this would be a moment when investors of AI companies would question the amount of funding that seems to be going into developing the LLMs and making the chips that power it, with Nvidia poised to bear the biggest brunt of the selling.

"There will be volatility in the stock market as capital markets absorb all of this information and re-price the values of the Mag7. Tesla is the least exposed, the rest are exposed as a direct function of the amount of CapEx they have publicly announced. Nvidia is the most at risk for obvious reasons. That said, markets will love it if Meta, Microsoft, Google etc can win WITHOUT having to spend $50-80B PER YEAR," wrote Chamath Palihapitiya.

Read More: The importance of being DeepSeek

DeepSeek's AI model is seen as a very strong competitor to OpenAI and Meta Platforms. For Wall Street investors, this may mean the stratospheric valuations of AI-related companies could come under question, as also the massive funding and investment plans lined up to develop the AI models.

Palihapitiya added that the innovation from China shows how 'asleep' the US AI ecosystem has for the past 15 years. "We’ve been running towards the big money/shiny object spending programs where we (USA) have thrown hundreds of billions of dollars at a problem vs thinking through the problem more cleverly and using resource constraints as an enabler," Palihapitiya said.

One CNBC report said analysts at Citi had 'inquiries around cost of compute'. Marc Andreessen, the general partner of Andreessen Horowitz posted that Deepseek R1 is 'AI's Sputnik moment'.

Bloomberg News quoted Vey-Sern Ling, managing director at Union Bancaire Privee, saying, "It can potentially derail the investment case for the entire AI supply chain, which is driven by high spending from a small handful of hyperscalers."

"It's a case of a crowded trade, and now DeepSeek is giving a reason for investors and traders to unwind," Reuters quoted Wong Kok Hoong, head of equity sales trading at Maybank.

DeepSeek's latest updates came just days after US President announced the Stargate AI Project, calling it the 'largest AI infrastructure in the history of AI', backed by some of the biggest names in tech and business. Stargate aims to invest $500 billion over the next four years to create world-class AI infrastructure in the United States.

"The idea that the most cutting-edge technologies in America, like Nvida and ChatGPT, are the most superior globally, there's concern that this perspective might start to change," Reuters reported quoting Masahiro Ichikawa, chief market strategist at Sumitomo Mitsui DS Asset Management.

DeepSeek researchers said the DeepSeek-V3 model used Nvidia's H800 chips for training, which were not top-of-the-line processing chips, and cost them less than $6 million.

The resulting turmoil is meant that even bitcoin has fallen below the $100,000 mark.

This is being updated.