The news of the development of the DeepSeek AI model—reportedly significantly more capital and energy-efficient than other competitors—sent riptides through markets. AI chip leader Nvidia alone shed 17% of its market value on Monday (nearly $600 million), only to bounce back 9% on Tuesday.
That volatility has advisor investment committees and CIOs grappling with whether DeepSeek represents a paradigm shift in AI that requires tweaking allocations or merely reintroduces some volatility in a disruption-prone sector that was bound to happen at some point.
Carson Group, for example, has a neutral position on tech and has been advising clients to diversify even before the DeepSeek tumult.
“The reality is that tech (and tech-adjacent) companies make up a significant portion of the broad S&P 500 index. So even before the DeepSeek news, there was a good case to be made for diversifying within equities, into areas like mid/small cap stocks and sectors outside tech, to reduce concentration risk,” said Sonu Varghese, global macro strategist at Carson Group. “We wrote in our 2025 outlook that we expect the bull market to broaden out this year based on the fundamentals of the economy and policy opportunities. In fact, the tech sector underperformed the S&P 500 in 2024, especially in the second half.”
Kristian Kerr, head of macro strategy at LPL Financial, added that a key question is whether diffusion models like DeepSeek represent a paradigm shift and a challenge to U.S. tech company dominance in AI.
“The bearish thesis for large language model developers and related semiconductor companies hinges on the belief that diffusion models represent a paradigm shift,” Kerr said. “By significantly reducing training and inference costs, DS could disincentivize massive capital expenditures and undermine the competitive advantage of companies like OpenAI and Google.”
However, if the efficiency advancements and the fact that DeepSeek released the model as open source help drive efficiency across the AI market, that could ultimately be a net positive, Kerr said.
“The market is going to grapple with whether or not this is a major fundamental shift and the potential end of ‘U.S. exceptionalism.’” Kerr said. “This abrupt re-evaluation challenges a narrative that has dominated the market for nearly two years. If the bearish case is right, it is a potentially significant shift, particularly given the recent surge in capital expenditures by leading tech companies, which has further contributed to the prevailing narrative of U.S. dominance. This sudden questioning of that narrative demands serious consideration, but it likely won’t be resolved in a few hours of trading.”
Brent Coggins, chief investment officer at Triad Wealth Partners, assessed that most of the damage has been explicitly focused on AI-connected companies rather than tech overall.
“Markets were 'deep seeking' a catalyst to blow some of the frothiness out of this AI-driven rally,” Coggins said. “What's reassuring is that initially, the damage, although severe, appears to be highly localized to the AI industry. We don't believe this is the ‘AI winter’ some have called for, just the beginning of a potential recalibration of market dominance and competitive advantages related to the technology.”
Additionally, because DeepSeek is a Chinese firm, security concerns could blunt the potential impact, according to Jake Miller, co-founder and chief solutions officer at Opto Investments, a private markets investment specialist.
“Are the training timelines and estimates to be trusted? Do we have independent confirmation of the actual number and type of chip used?” Miller said. “Early reactions to cost and logic abilities, especially in math and coding, have been very positive, but given security concerns for a Chinese firm, it’s unlikely we will see enterprise adoption anytime soon, especially in the current climate.”
Regarding long-term implications for AI investment overall, Miller still sees value in what he calls “boring AI,” such as applying machine learning to improve efficiencies in industries like healthcare, logistics, legal, construction and energy, which have been slow to adopt the tech.
“This ecosystem offers multiple points of entry—beyond just backing the next generative AI unicorn,” Miller said.
