Tech Stocks, Cryptocurrencies Pull Back as OpenAI Competitor DeepSeek Spooks Market

A premarket tech stock selloff extended into the cryptocurrency market on Monday (January 27) ahead of a data-packed week that includes interest rate announcements from the US Federal Reserve and Bank of Canada.

The selloff began after DeepSeek, a Chinese rival to OpenAI, became the top free app in the Apple Store over the weekend. DeepSeek-R1, which was released on January 10, can reportedly perform reasoning tasks just as well as OpenAI’s o1, but costs less and is a partially open system that allows researchers to study it.

Also last week, ByteDance released Doubao-1.5-pro, an upgrade to its flagship artificial intelligence (AI) model, claiming it outperforms OpenAI's o1 in AIME, a benchmark test that measures how well AI models understand and respond to complex instructions. ByteDance is the owner of popular social media app TikTok.

Reuters later reported that outages were affecting DeepSeek as its popularity skyrocketed.

For its part, OpenAI unveiled its AI agent, Operator, on January 23, launching in the US for ChatGPT Pro users; it has no immediate plans to expand the release to Plus, Team or Enterprise customers. Operator is trained to interact with graphical interfaces online, allowing it to “see” web browsers and perform tasks such as making reservations. It can reportedly self-correct if it makes a mistake and will hand control back to the human user if it runs into challenges.

Elsewhere, Meta Platforms (NASDAQ:META) CEO Mark Zuckerberg posted a projection of his company's AI spending targets, indicating that Meta plans to spend between US$60 billion and US$65 billion on AI in 2025.

The funds will primarily be used to build a data center so large that it “covers a significant part of Manhattan.” The data center will power an AI engineer, which will contribute to the company’s research and development efforts.

With the release of DeepSeek, the level of spending and investment in AI by western countries is suddenly facing higher levels of scrutiny, with a cheaper alternative readily available and popular with consumers.

“We still don’t know the details and nothing has been 100 percent confirmed in regards to the claims, but if there truly has been a breakthrough in the cost to train models from US$100 million+ to this alleged US$6 million number this is actually very positive for productivity and AI end users as cost is obviously much lower meaning lower cost of access,” Reuters quotes Jon Withaar, a senior portfolio manager at Pictet Asset Management, as saying.

Researchers for the small Hangzhou startup behind DeepSeek wrote in a paper last month that the DeepSeek-V3 model was trained using NVIDIA's (NASDAQ:NVDA) H800 chips, which were initially developed as a reduced-capability product to get around US restrictions on sales to China. Training the model reportedly cost less than US$6 million.

US sanctions subsequently banned the chips. If DeepSeek is able achieve these results with mid-level chips, it challenges the narrative that advanced computing abilities are necessary.

Don't forget to follow us @INN_Technology for real-time news updates!

Securities Disclosure: I, Meagen Seatter, hold no direct investment interest in any company mentioned in this article.

Editorial Disclosure: The Investing News Network does not guarantee the accuracy or thoroughness of the information reported in the interviews it conducts. The opinions expressed in these interviews do not reflect the opinions of the Investing News Network and do not constitute investment advice. All readers are encouraged to perform their own due diligence.