Nvidia's stock gets swept up in software selloff, but this analyst says that makes no sense

By Britney Nguyen

The 'indiscriminate' chip selloff is reminiscent of DeepSeek fears that were 'proved unfounded,' a Bank of America analyst says

Shares of Nvidia and other AI-related chip stocks fell on Wednesday.

The recent free fall in software stocks has spilled over to the chip sector, and a Bank of America analyst says the selling doesn't make sense.

The "indiscriminate" volatility around chip stocks "seems internally inconsistent and suggests another DeepSeek-like overblown selloff," Bank of America analyst Vivek Arya said in a Tuesday note.

Shares of chip makers including Nvidia (NVDA) and Broadcom (AVGO) each fell more than 3% on Wednesday. Memory and storage stocks such as Micron Technology (MU) and Sandisk (SNDK), which have seen massive gains due to artificial-intelligence-driven demand, slid even more.

The extent of the selloff reflects "internally inconsistent" ideas, according to Arya. The fall in AI-related chip stocks suggests that investments into AI are "deteriorating to the point" that returns on investment will be weak and further growth is not viable, he said in his report. At the same time, the fall in software stocks implies that the market believes "AI adoption will be so pervasive and productivity-enhancing" that software-based businesses and applications are destined for extinction.

"Both outcomes cannot occur at once," Arya said, writing of a resemblance to fears about DeepSeek, "which ultimately proved unfounded." The release of the Chinese AI startup's reasoning model in January 2025 roiled the tech trade as investors worried that if DeepSeek could develop a competitive AI model cheaply, other companies would also be able to do that without requiring so many expensive and powerful chips.

See more: Why AMD's stock is diving toward its worst day in years after earnings

But that market hiccup was followed by more spending on AI and accelerating growth of AI tokens, or the pieces of information processed by AI models, Arya said. In 2025, cloud capital expenditures grew an estimated 69% annually - more than double original forecasts for 20% to 30% growth, he added.

While current AI models are impressive, Arya said it will likely take several years for them to prove their worth from a productivity standpoint. In the meantime, he doesn't see AI investments slowing down, and he expects them to continue even after AI models achieve major productivity breakthroughs, as they'll still need to improve and keep users engaged.

"While we cannot opine on the ultimate shape of the software industry, we believe the chip industry has been and will continue to be positively levered to the AI buildout," Arya said.

Don't miss: Nvidia and Oracle are flashing similar warning signs about the AI trade

Enterprise adoption of AI is still in the early stages, Arya said, and investments are now also coming from sovereign states that want to deploy and control AI initiatives.

Meanwhile, Arya said chip-sector valuations already bake in caution over possible slowdowns in spending or drops in earnings expectations "that we believe may not materialize."

-Britney Nguyen

This content was created by MarketWatch, which is operated by Dow Jones & Co. MarketWatch is published independently from Dow Jones Newswires and The Wall Street Journal.

(END) Dow Jones Newswires

02-04-26 2047ET