Live Markets, Charts & Financial News

Nvidia stock begins recovery after DeepSeek AI frenzy prompted near $600 billion loss

3

NVIDIA (NVDA) shares increased about 9 % on Tuesday, when artificial intelligence maker began to recover from a huge decrease in the previous day, which shaved about $ 600 billion from the market ceiling.

The fall of NVIDIA was defended on Monday by 17 % by investor concerns related to a new and cost -effective artificial intelligence model from the Chinese company Deepseek. Some Wall Street analysts concern that the costs of Dibsic are cheaper, which they claimed to have had trained artificial intelligence models, partly due to the use of fewer artificial intelligence chips, which means that American companies were spending on artificial intelligence infrastructure.

This has led to anxiety among the investment community that the prices of GPU are high (graphics processing unit, or artificial intelligence chip) may be subjected to pressure and this demand can fade on semiconductors.

The decrease in the maximum NVIDIA market was worth $ 589 billion, a one -day loss in the history of the stock market.

Deepseek ads not only NVIDIA but the market in general, with heavy NASDAQ (^IXIC) decreased by 3 %. Chips shares fell across the painting on Monday, but some names began to recover. After a decrease of more than 17 % to start the week, Broadcom (AVGO) increased by 2.6 % on Tuesday.

NVIDIA did not express the concern about the DEPSEEK, as it described the R1 as “excellent progress of artificial intelligence” in a statement on Monday.

Jensen Huang speaks at Nvidia Keynote in Michelob Ultra Arena in Las Vegas, NV, on January 6, 2025. · Deecee Carter/Mediapunch/Mediapunch/IPX

The Wall Street analysts continued to think about the marketing of the market fed on Tuesday, expressing doubts about the low costs of Depsic to train artificial intelligence models and the effects of artificial intelligence shares.

JPMorgan Harlan Sur and Citi Kristen Danley analyst said in separate notes to investors that because Deepseek used a process called “distillation”-and in other words, he relied on the Lama Open-Meeta model (Meta) to develop its-low model. The spending on the start of the Chinese operation (less than 6 billion dollars to train the last V3) did not include its entire costs.

“We believe it is very important to verify these costs before extracting conclusions,” wrote a wall.

Danley added: “Deepseek relies on benefiting from cloud service providers (META) and artificial intelligence is still in its cradle, we tend to the argument for the continuous strong growth in spending artificial intelligence,” Danley added.

However, Deepseek is clearly “it is clear that he cannot access the largest amount of accounts that managed to hyperacton and somehow managed to develop a very competitive model,” wrote Raymond James Srini Pajjuri in an investor note on Monday.

Stockstory aims to help individual investors overcome the market.
Stockstory aims to help individual investors overcome the market.

Laura Praton is a reporter for Yahoo’s funding. Follow it on Bluesky @laurabratton.bsky.social. Email it on laura.bratton@yahoink.com.

Click here to get the latest securities market news and in -depth analysis, including events that move shares

Read the latest financial and commercial news from Yahoo Financing

Comments are closed, but trackbacks and pingbacks are open.