In a dramatic turn of events, U.S. tech stocks experienced their largest single-day loss in Wall Street history on Monday, coinciding with the rise of Chinese startup DeepSeek's latest advancements in artificial intelligence.
DeepSeek, based in Hangzhou, unveiled its new Janus-Pro-7B image generator, which outperforms OpenAI's DALL-E 3 in various benchmarks. Additionally, the company's DeepSeek assistant app surpassed ChatGPT in U.S. App Store downloads, showcasing China's growing prowess in the AI sector.
However, shortly after these announcements, DeepSeek announced a temporary halt on new user registrations due to a recent cyberattack. This news sent shockwaves through the U.S. tech market, leading to a sharp selloff.
Analysts pointed to several factors exacerbating the downturn, including high tech valuations, concerns over inflated infrastructure spending, and signs that U.S. chip export controls are failing to impede Chinese advancements in AI. Nvidia was particularly affected, its stock plummeting nearly 17 percent and wiping out almost $600 billion in market value. Other major players like Oracle and Broadcom also saw significant declines, with the Nasdaq falling by 3.1 percent and the Philadelphia Semiconductor Index dropping 9.2 percent—the worst since March 2020.
The global impact was swift, with Japan's SoftBank and Europe's ASML among those hit by the market turbulence.
Brian Jacobsen of Annex Wealth Management commented, \"If it's true that DeepSeek is the proverbial 'better mousetrap,' that could disrupt the entire AI narrative that has helped drive the markets over the last two years. It could mean less demand for chips, less need for a massive build-out of power production to fuel the models, and less need for large-scale data centers.\"
Model Momentum Meets Market Mania
DeepSeek's dual release of the Janus-Pro-7B and the DeepSeek-R1 models has intensified existing market uncertainties. The Janus-Pro-7B, an upgrade from December's image model, leverages 72 million synthetic images to achieve significant advancements in multimodal understanding and text-to-image capabilities. Meanwhile, the DeepSeek-V3-powered assistant's rapid rise in the U.S. highlights China's competitive edge in consumer AI.
Notably, the V3 model was developed using only $6 million worth of Nvidia's restricted H800 chips—a fraction of the industry's usual investment—and the R1 model's API is priced at around one-thirtieth of OpenAI's offerings.
Venture capitalists like Marc Andreessen have likened DeepSeek's emergence to a \"Sputnik moment,\" drawing parallels to the Soviet satellite launch that ignited the space race.
Strategic Fault Lines
The stock market selloff has exposed deep divisions regarding the future of AI. Daniel Morgan, senior portfolio manager at Synovus Trust Company, argued that the reaction was exaggerated, noting that DeepSeek's AI models are geared towards mobile phones and PCs rather than the highly profitable data centers.
Despite some optimism, the situation raises significant questions about U.S. competitiveness in the AI arena, a concern recently echoed by former President Trump, who described DeepSeek as a \"wake-up call.\"
Founded in 2023 by quant trading veteran Liang Wenfeng, DeepSeek has rapidly ascended from obscurity to prominence in Silicon Valley. Its open-source strategy and efficient chip usage have made it a prime example of how Chinese companies are navigating and circumventing international sanctions to drive innovation.
Reference(s):
cgtn.com
