NVIDIA GTC – A "pep rally" for "shocked" investors

At the recently concluded NVIDIA GTC conference, despite Jensen Huang showcasing powerful next-generation chips and various new products like personal "supercomputers," investors seemed unconvinced, with the company's stock price plummeting 4% after Huang's keynote speech, marking a nearly 12% drop since the beginning of the year.

Behind this phenomenon, multiple concerns about NVIDIA are emerging in the market. On one hand, Trump's tariff policy may impact its overseas business, the rapid rise of DeepSeek companies could threaten the demand for high-end AI chips, and on the other hand, many tech giant clients are turning to self-developed chips. These three major threats fundamentally challenge NVIDIA's long-standing market monopoly.

Crisis Behind Record Attendance

This year, NVIDIA's GTC conference attracted a record 25,000 attendees to the San Jose Convention Center and surrounding buildings. Many workshops, speeches, and panel discussions were packed, forcing people to stand against the walls or sit on the floor, enduring the organizers' loud and stern instructions to queue.

At the conference, Jensen Huang showcased a dazzling array of new products, including the Blackwell B200 super GPU chip, a personal supercomputer called Jetson, and an adorable robot named Blue.

Huang claimed these products would "revolutionize the AI computing experience," particularly noting that the Blackwell chip offers over five times the performance improvement compared to the previous generation H100.

"The more you buy, the more you save," Huang said during his keynote on Tuesday, "In fact, it's even better than that. Now, the more you buy, the more you earn."

However, these flashy technological achievements did not soothe investors' nerves. Despite the excitement and enthusiasm at the conference, NVIDIA's stock price fell about 4% on the trading day following Huang's speech. The market sent a clear signal of skepticism regarding NVIDIA's future growth prospects.

Trump's Tariff Policy - A Stumbling Block for NVIDIA's Global Expansion

The primary threat NVIDIA currently faces comes from the political arena. According to CCTV News, President Trump previously stated that he would impose widespread reciprocal tariffs and additional tariffs on specific industries starting April 2.

At the GTC conference, NVIDIA's management attempted to alleviate market concerns about tariffs, with Huang claiming that tariffs would not cause "significant damage" in the short term. However, he did not commit to NVIDIA being shielded from long-term economic impacts—regardless of how they ultimately manifest.

NVIDIA has clearly received the Trump administration's "America First" message, with Huang promising to invest hundreds of billions in manufacturing in the U.S. at the GTC. While this will help the company diversify its supply chain, it also represents a significant cost for NVIDIA, whose trillion-dollar valuation relies on healthy profit margins

The Threat of DeepSeek

What worries investors even more is the rise of ultra-high cost-performance models like DeepSeek, which poses a threat to the demand for high-end AI chips. The widespread adoption of DeepSeek may mean that the AI industry does not necessarily need to hoard NVIDIA's expensive chips.

Jensen Huang once again tried to alleviate investors' concerns at the GTC conference, stating that the worries about the DeepSeek inference model R1 are completely misplaced, and that such software will require fewer chips and more powerful servers in the future.

Huang believes that the demand for computing remains extremely high. Artificial intelligence (AI) inference has driven up the demand for computing. New AI models that can produce more complex answers will only increase the demand for computing infrastructure.

The Trend of Customers Developing Their Own Chips: Reducing Dependency

The third threat comes from within NVIDIA's largest customer base.

From Meta and Google to Amazon and Microsoft, almost all major tech giants are actively developing their own AI chips in an attempt to reduce their reliance on NVIDIA. Once they succeed, it is almost certain that this will weaken NVIDIA's monopoly in the AI chip market.

According to a report by tech media The Information, Amazon Web Services (AWS) is adopting a price-cutting strategy to persuade customers to switch to servers powered by self-developed AI chips, abandoning the more expensive NVIDIA chips.

The report stated that a long-term AWS customer revealed that AWS recently pitched them servers using Trainium chips. These servers have computing power similar to those with NVIDIA's H100 chips, but cost only a quarter of the latter