It was a black Monday for Nvidia and tech companies. The emergence of DeepSeek and Janus Pro, the Chinese Artificial Intelligence models, triggered one of the most impactful stock market crashes in recent history. At the center of this controversy are fears that the high spending on next-generation chips, which has been crucial for AI so far, could be threatened by new approaches that optimize resources and reduce dependence on expensive hardware.
Below, we’ll break down everything that happened between DeepSeek and Nvidia, and why the fall of the Santa Clara company dragged down the rest of the tech market.
Why did Nvidia fall, and how much did it really lose?
Monday, January 27, was a black day for Nvidia shareholders. The company lost 16.97% of its value in a single day, equivalent to $589 billion in market capitalization.
This drop far exceeded its previous historical record (a 9% setback in September when it lost nearly $279 billion in value). Never in the history of the United States stock market had such a large reduction been recorded in a single day for a company of this size.
Nvidia’s debacle not only affected the company itself but also spread to the rest of the market, thanks to the significant weight its shares have in the S&P 500 and Nasdaq 100 indices.
While the S&P 500 fell around 1.5% during Monday’s early trading, the Nasdaq 100 plunged nearly 3%. This chain reaction revealed the tech sector’s heavy dependence on expectations surrounding Artificial Intelligence and, in particular, Nvidia’s role as the leading chip provider for training AI models.
ALSO READ: Is DeepSeek under cyber attack? What we know
What is DeepSeek?
DeepSeek is a Chinese Artificial Intelligence startup that has garnered attention in the sector in recent months due to its remarkable progress.
Under the leadership of Liang Wenfeng—an expert in quantitative investment funds—the company recently launched an Artificial Intelligence model that analysts claim is comparable in results to those developed by giants like OpenAI and Meta.
This wouldn’t be newsworthy if it weren’t for the fact that the training cost and infrastructure required for its AI are reportedly much lower than those of its Western competitors.
DeepSeek has managed to top the list of free apps most downloaded on Apple’s App Store, which contributed to increasing media pressure and curiosity about the technology behind its AI. Additionally, its open-source philosophy has raised expectations in the community by allowing developers worldwide to analyze and leverage its models.
ALSO READ: What’s happening in Mexico? Video of Selena Gomez crying is trending on Google
What are investors’ concerns about DeepSeek?
Alarm bells rang when it was revealed that DeepSeek’s approach is based on efficiency and the use of less expensive resources. Wall Street feared a simple equation: if a Chinese AI model, running on lower-powered hardware, achieves similar performance to Western models, why continue paying a fortune for high-end chip-based solutions?
Jefferies, a financial analysis firm, made it clear in its report to investors that this “low-cost approach” could destabilize the current AI business model. Until now, the prevailing idea was that very costly and powerful components—largely supplied by Nvidia—were necessary to keep improving deep learning algorithms.
DeepSeek demonstrates that there could be shortcuts and more efficient training and inference methods, thereby threatening the massive investment in chips and data centers.
What role does Nvidia play in Artificial Intelligence development?
Nvidia has become synonymous with Artificial Intelligence acceleration. Its graphics processing units (GPUs) are indispensable for training large AI models, such as OpenAI’s ChatGPT or Meta’s systems.
This market dominance turned Nvidia into one of the most valuable and admired companies in recent years, with its market value quintupling in just a couple of fiscal years.
The cost of Nvidia hardware is high since its chips are designed to process vast amounts of data with great speed and energy efficiency. Tech giants have spent billions equipping themselves with this hardware, and any threat to this consumption model naturally raises investors’ concerns.
How did Nvidia respond to DeepSeek’s emergence?
Surprisingly, Nvidia was quick to comment on DeepSeek’s advances, describing them as an “excellent breakthrough in artificial intelligence” that complies with U.S. export restrictions. The email statement noted that the Chinese startup’s achievement is a good example of how new AI models can be trained using “Test Time Scaling” and publicly available tools without violating North American export controls.
This response from Nvidia had a dual interpretation: on one hand, the company sought to dispel suspicions that DeepSeek had accessed restricted high-end chips for its training; on the other, it seemed to emphasize that even with more efficient models, the inference phase (i.e., when the AI operates and “reasons” in real-time) still requires a large number of GPUs and robust infrastructure.
What impact does DeepSeek’s rise have on other tech giants?
The pressure generated by DeepSeek extended to giants like Meta, Microsoft, and OpenAI. Seeing Nvidia’s setback, analysts and investors began reevaluating the multimillion-dollar investments these firms are making in AI.
For example, Meta announced on January 24 a significant increase in its capital budget for Artificial Intelligence projects, surpassing $65 billion annually. Similarly, the joint venture Stargate, by OpenAI, SoftBank, and Oracle, plans a mega investment of $100 billion to build data centers in the United States.
However, DeepSeek’s success, which some call “the disruptive model of Chinese AI,” raises the possibility of optimizing hardware and reducing costs.
What role does JANUS play in DeepSeek’s strategy?
While DeepSeek has made headlines for its successful chatbot and its ability to compete with OpenAI’s ChatGPT, the company launched Janus Pro. This project, specializing in generating images and creative content, aims to position itself as an alternative to models like DALL-E or Stable Diffusion, with the same promise of resource efficiency and open code.
Janus hasn’t been the main factor concerning Nvidia investors, but it reinforces the perception that DeepSeek aims to cover various fronts in AI, from language processing to visual content creation.
