In a surprising twist, Nvidia $NVDA, a leading figure in the artificial intelligence surge, experienced a significant decline, with its shares tumbling over 15% on Monday. This drop resulted in a massive $500 billion loss in market value, sending shockwaves through the tech-centric Nasdaq Composite and leaving investors baffled by the abrupt change in the AI industry.
The cause of this market disruption? A Chinese startup named DeepSeek, which made headlines by announcing a suite of open-source models purported to rival the performance of industry leaders like OpenAI's ChatGPT, but at a greatly reduced cost. This unforeseen advancement has complicated the strategies of major US tech firms heavily invested in AI infrastructure.
As the situation unfolds, the key question remains whether the significant capital expenditure growth anticipated for 2026 can sustain its pace in a market previously deemed highly efficient. Analysts from US Tiger Research are issuing warnings, suggesting that Nvidia's stock price may remain under pressure until greater clarity emerges.
The introduction of DeepSeek's premier model, DeepSeek R1, is creating quite a stir in the tech sector. Although Bernstein analysts commend the innovative models, noting their capability to compete with offerings from OpenAI and Meta Platforms, they contend that the market's initial reaction could be somewhat exaggerated.
Bernstein's analysts highlight that even if DeepSeek has managed to cut the costs of achieving similar model performance by a tenth, the cost trends of models are already rising by approximately that rate on a yearly basis. They argue such breakthroughs as DeepSeek are crucial for the AI industry to continually innovate and maximize existing hardware.
However, a fascinating debate is unfolding regarding the true computing power utilized by DeepSeek. Alexandr Wang, the billionaire CEO of Scale AI, estimates it involves about 50,000 NVIDIA Hopper GPUs, as opposed to the 10,000 A100 NVIDIA chips initially reported. Tesla CEO Elon Musk shares Wang's assessment, adding credibility to this claim.
The mystery deepens given that DeepSeek cannot disclose their actual chip usage due to U.S. export restrictions on AI chips. This adds another layer of complexity to the evolving situation.
In spite of market anxieties, some industry experts remain optimistic about the long-term outlook for AI infrastructure. Bernstein analysts assert that any new computing capacity is likely to be absorbed by the growing appetite for AI, suggesting that the momentum in AI infrastructure is far from dwindling and might just be gaining steam.
As the dust settles in this unpredictable AI landscape, one thing is certain: the environment is changing more rapidly than anyone predicted. For investors, tech enthusiasts, and industry stakeholders, the message is clear – brace yourself, as the AI revolution is only beginning and promises to be an exhilarating journey.
To keep pace in this rapidly evolving market, consider exploring strategies for trading and copy trading at Limex.
