DeepSeek AI, a low cost Chinese artificial intelligence (AI) startup based in Hangzhou, has shaken U.S. tech stocks and global markets, causing chip manufacturer Nvidia’s value to plummet by up to $593 billion overnight—a historic 17 percent single-day decline for the Silicon Valley star.
As per LSEG data, this marks the highest one-day market capitalization loss for a Wall Street stock ever recorded.
The upheaval caused by AI has affected Jensen Huang’s Nvidia the most severely, yet the situation is grim for everyone — the S&P 500 fell, and the Nasdaq declined by more than 3 percent.
Concerns over the expensive US pursuits in AI technology, in contrast to the seemingly similar Chinese option—which is accessible as open source and produced at a significantly lower cost—were primarily responsible for dragging down the markets.
ALSO READ: Who is this Indian-origin leader aiming to succeed Justin Trudeau.
We examine how DeepSeek’s AI disruption created a $593 billion shockwave for Nvidia, impacted global tech stocks, and unsettled stock markets globally.
DeepSeek AI
As reported by Reuters, Chinese start-up DeepSeek introduced a free AI assistant last week, asserting that it utilizes less data and is significantly cheaper than other current options, including OpenAI’s ChatGPT, Google’s Gemini, and others.
As of January 27, DeepSeek AI surpassed ChatGPT in downloads from the US iOS Apple app store and exceeded it in various areas, such as the UK, Australia, Canada, China, and Singapore.
What Makes China’s DeepSeek a threat to US AI Companies?
DeepSeek’s models, such as DeepSeek-V3 and DeepSeek-R1, are created by a startup in Hangzhou, predominantly owned by Liang Wenfeng, who is a co-founder of the quantitative hedge fund High-Flyer.
The company gained prominence last month when several benchmarks indicated that its V3 large language model (LLM) surpassed those of numerous well-known US tech firms, even though it was created at a significantly lower expense.
DeepSeek’s R1 language model, which replicates elements of human reasoning, also matched and exceeded OpenAI’s latest o1 model in several benchmarks.
Additionally, in a paper released last month, DeepSeek researchers noted that the V3 model utilized Nvidia H800 chips for training, costing under $6 million—an insignificant amount when contrasted with the billions that AI leaders such as Microsoft, Meta, and OpenAI are set to invest this year alone.
