Some of the heaviest hitters in the tech sector may be asking themselves that question following the brass-knuckled debut of China’s DeepSeek AI agent, which hosed down the stock market with gallons of red ink.
💵💰Don’t miss the move: Subscribe to TheStreet’s free daily newsletter 💰💵
DeepSeek claimed that its model was developed at a fraction of the cost of those of its rivals, which analysts said raised questions about the future of America’s AI dominance and the scale of investments U.S. tech firms are planning.
Companies are expected to shell out an estimated $1 trillion to build data centers and other AI infrastructure.
Earlier this month, President Donald Trump announced the creation of a new company, Stargate, to bulk up U.S. artificial-intelligence infrastructure with plans to inject up to $500 billion into the project in the coming years.
Trump said the “release of DeepSeek AI from a Chinese company should be a wakeup call for our industries that we need to be laser-focused on competing.”
AI-chip monsoon Nvidia (NVDA) got the wind knocked out of its monstrous sails on Jan. 27, as its shares dropped a stunning 17%.
And Advanced Micro Devices (AMD) , which saw its shares drop about 18% in 2024, lost nearly 7% when DeepSeek came to town.
The company has struggled to meet investors’ expectations for its efforts to develop and sell chips for AI applications.
Lisa Su, chairwoman and CEO of Advanced Micro Devices, during Computex in Taipei. Photographer: Annabelle Chih/Bloomberg via Getty ImagesBloomberg/Getty Images
AMD, scheduled to report quarterly results on Feb. 4, beat Wall Street’s third-quarter earnings and revenue expectations in October.
However, the company’s fourth-quarter revenue estimate fell short of Wall Street’s expectations.
Related: Analyst unveils AMD stock price target on AI and server update
CEO Lisa Su suggested that chip supplies could be tight this year, but added that “but we’ve also planned for significant growth going into 2025.”
AMD addressed DeepSeek in a Jan. 29 post on X, declaring that “deploying DeepSeek R1 distilled ‘reasoning’ models on AMD Ryzen AI processors and Radeon graphics cards is incredibly easy and available now through @lmstudio. See how to quickly get them up and running!”
Some of the market’s more bullish experts aren’t worried about what DeepSeek might mean for AI stocks and the broader tech sector. In fact, they say it could end up as a favorable indicator for tech stocks.
Stephen “Sarge” Guilfoyle spoke bluntly about the DeepSeek disaster in his recent column on TheStreet Pro.
“OK. So, Monday was a lousy day. That said, it wasn’t a really, really, really lousy day,” he said.
“The ‘ugly stick’ was out and about, as we had said it would be. Parts of the market recovered, though, well before the closing bells had rung their last down at Wall Street and up at Times Square.”
The veteran trader, whose career dates back to Wall Street in the 1980s, noted that U.S. equities did not take a broad beatdown and “winners beat losers at the NYSE by a rough 4 to 3.”
Guilfoyle noted that DeepSeek’s quality and accuracy have not been verified.
“We really have no idea if the expense that DeepSeek claimed it took to train and develop the model is accurate, nor do we know if the number of and type of Nvidia chips claimed to have been used are accurate,” he said.
Guilfoyle noted there is still a legitimate possibility that much more expensive U.S. models are considerably superior when put to a high-level test.
“It may be true that the gap between models is less than once thought and that elite level U.S.-designed chips, from Nvidia in particular, are grossly overpriced,” he said. “Then again, take nothing for granted and trust not a soul. Until we know more.”
Guilfoyle said he’d decided to start rebuilding his long positions in AMD and Nvidia, telling readers that they “are now my 16th and 17th largest allocations, so I am not yet overly exposed.”
AMD shares were up nearly 2% to $116.32 at last check.
More 2025 stock market forecasts
Baird analysts said they, too, would be buyers of Nvidia and AMD on the current stock corrections.
Nvidia stands out as a less expensive stock than Broadcom (AVGO) , while AMD offers the most value at current levels, the investment firm said.
Baird said the launch of DeepSeek “catalyzes concerns” around rising power and implementation complexities in artificial intelligence systems.
However, AI demand remains very strong, the firm said.
Baird noted that the rise of graphics processing units in graphics since the late 1990s has been hardware-driven, and it is “very skeptical on performance workarounds” using simpler GPU architectures or pooling consumer central-processing-unit resources.
While uncertainties about 2026 capital expenditures remain, AI stocks “cannot ignore strong numbers for the rest of this year,” Baird contended.
Melius Research downgraded AMD to hold from buy with a price target of $129, down from $160.
The downgrade, however relates not to DeepSeek but rather to a more cautious take on the x86 server, the dominant architecture for personal computers and servers, and on personal computers more broadly, over the long term for AMD, firm said.
Melius said Nvidia would increasingly “come for” AMD in both markets with its Arm-based central processing units optimized for accelerated PCs.
Also, the firm said it was more and more concerned that custom CPUs and Nvidia CPUs will cannibalize even more of the x86 server market over the long term even though AMD’s current chip, Turin, is doing well.
Wedbush analysts cut their price target on AMD to $150 from $200, but maintained their outperform rating.
“We are moderating our outlook for AMD’s 2025 results with reduced expectations around AI GPU sales, albeit partially offset by increased PC/server CPU expectations,” the firm said.
Wedbush said AMD should make further inroads into the corporate market in 2025, helped by Dell’s new AMD-based business PCs. The Turin launch should yield further gains in 2025.
Related: Veteran fund manager issues dire S&P 500 warning for 2025