Investing.com - Last weekend, a new name in the artificial intelligence arms race was becoming increasingly ubiquitous: DeepSeek.
Once a relatively unknown Chinese start-up, DeepSeek said its free AI model could match the performance of OpenAI's ChatGPT despite operating on less-advanced chips and costing around just $6 million to build.
On Monday, the implications of this statement roiled equity markets. Despite doubts swirling around DeepSeek's claims, investors worried that the firm's emergence may call into question the necessity of heavy AI spending plans being pursued by some of Silicon Valley's biggest players.
The tech-heavy Nasdaq Composite and benchmark S&P 500 fell sharply, while Nvidia (NASDAQ: NVDA ), the poster child of a recent boom in enthusiasm around AI, shed a Wall Street record $593 billion off its market value in just that one session.
In a note to clients this week, analysts at Bernstein argued that, although many questions surround the DeepSeek saga, they believe the dramatic stock market reactions were "in at least some cases [...] likely incorrect or overblown".
Instead, they emphasized concerns that a prevailing belief among some investors that "bigger spending means better results" could be "incorrect". Big Tech names have been pouring massive investments into AI, arguing that the spending is necessary to ensure a strong position in the ongoing race to develop -- and ultimately monetize -- the technology.
"[W]e have worried about for a while that the arms race we have been watching may be ending and the end result may not be as wonderful as many have believed," the analysts led by Mark Moerdler wrote.
Still, some of the tech industry's leading executives, including the bosses of software giant Microsoft (NASDAQ: MSFT ) and Facebook-owner Meta Platforms (NASDAQ: META ), moved to defend AI spending during calls with investors this week.
Microsoft CEO Satya Nadella said costs associated with AI are starting to come down for the company, which has banked greatly on an early investment in OpenAI to boost its AI capabilities. He added that, as the technology becomes "more efficient and accesible", Microsoft "will see exponentially more demand".
Mark Zuckerberg, CEO of Facebook-owner Meta Platforms, also took an optimistic approach to DeepSeek, saying the success of its models bolsters his belief that open source AI is the correct strategy. Meta has previously said it plans to spend $65 billion in 2025 to power its AI ambitions.
Yet, following the rise of DeepSeek, the Bernstein analysts said long-held assumptions that U.S. tech titans like Microsoft and Meta have a substantial lead in AI over their foreign counterparts "simply may not be true".
"This question brings up a different issue and that is — can you create a differentiated moated business when you are training models on data that is generally available to everyone," the analysts said.