Nvidia’s stock continued its downward trend on Monday, falling 1.9% to $110.59 in premarket trading. This decline comes even as new data points to robust demand for the company’s artificial intelligence chips.

Investors appear more concerned with broader economic worries than with Nvidia’s strong business fundamentals. The overall market sentiment was negative, with S&P 500 futures down 1.1%.

The chip giant has now fallen 16% year-to-date through March 7. This marks Nvidia’s worst start to a year since 2022, when shares dropped 27% during the same period.

Market uncertainties stemming from President Donald Trump’s tariff policies have weighed heavily on Nvidia stock. Trump’s weekend comments refusing to rule out a U.S. recession this year likely contributed to Monday’s sell-off.

Despite these headwinds, evidence of strong AI chip demand continues to emerge. Taiwan Semiconductor Manufacturing Company (TSMC), which manufactures Nvidia’s AI chips, reported February sales up 43% compared to the same month last year.

In another positive development, Hon Hai Precision Industry, better known internationally as Foxconn, announced that Nvidia helped it develop its own large-language AI model. Nvidia supported Foxconn’s FoxBrain AI model through its Taiwan-based supercomputer and technical consulting services.

Melius Research analyst Ben Reitzes maintained a Buy rating on Nvidia stock in a Monday research note. However, he lowered his two-year price target to $170 from $195.

“At this point, we believe Nvidia and several others in the AI semis and hardware space are on sale and good buys right now,” Reitzes wrote. He cautioned that near-term performance may still be affected by uncertainties around “regulations and geopolitics including tariffs.”

Investors are looking forward to Nvidia’s GPU Technology Conference next week. CEO Jensen Huang will deliver a keynote speech on March 18, which could provide clarity on the company’s future plans.

Reitzes expects Nvidia to present “a clear GPU innovation roadmap” at the conference. This will likely include details on upcoming products like “Blackwell Ultra” and its GB300 system, the “Rubin” GPU planned for 2026, and the Arm-based CPU called “Vera.”

In a major development, Bloomberg reported on March 6 that OpenAI and Oracle are planning to fill a massive new data center in Texas with Nvidia chips. This facility will be part of the $100 billion Stargate infrastructure project.

The data center is expected to house 64,000 of Nvidia’s GB200 semiconductors by the end of 2026. These powerful chips cost between $60,000 and $70,000 each, representing a major revenue opportunity for Nvidia.

According to an OpenAI spokesperson, the startup is collaborating with Oracle on the design and delivery of the Abilene data center. Oracle will be responsible for managing the supercomputer once complete.

The OpenAI-Oracle deal highlights the continued strong demand for Nvidia’s high-end AI products. Even as investor sentiment has cooled on chip stocks, real-world deployments continue to expand.

Market analysts have noted that attention may be shifting from chip stocks to software companies as the AI boom progresses. Some investors now see software firms as the next major beneficiaries of AI advancements.

Despite this potential shift, Nvidia remains the dominant provider of the hardware foundation for AI development. The company’s chips continue to be essential components for training and running large AI models.

Jensen Huang, Nvidia’s CEO, has successfully positioned the company as the key supplier of AI acceleration technology. While stock performance has been volatile in early 2025, the company’s business fundamentals appear to remain strong.