Top 20 US stock trading volumes on March 31: NVIDIA falls into a bear market territory

On Monday, U.S. stocks had Nvidia ranked #1 by trading volume, falling 1.40% with $30.467 billion in shares traded. Nvidia closed today at $165.17, down 20.2% from its October 29, 2025 closing price of $207.02, marking that the stock has entered a bear market zone.

French AI company Mistral AI said on Monday that it has raised $830 million through its first round of debt financing. The funds are provided by a bank syndicate. The financing will be used to build a data center near Paris, expected to deploy around 13,800 Nvidia GB300 graphics processing units.

Although Nvidia does not disclose pricing for individual chips, analysts estimate that a rack containing 72 of these GPUs costs about $3 million. Based on this figure, the new facility could involve Nvidia hardware worth about $575 million.

Micron Technology, a memory concept stock ranked #2, fell 9.88% with $24.175 billion in shares traded. The stock has already declined more than 20% this month. Recently, Google released its TurboQuant technology, which triggered a broad drop across the global semiconductor sector. The market is concerned that AI’s memory demand elasticity is declining.

According to reports, Google’s TurboQuant technology is designed to reduce AI inference memory usage, which may ease reliance on high-end storage products such as HBM. This drop does not negate AI demand; instead, it reflects a reassessment of the “AI memory scarcity narrative.” Investors have begun to adjust the high premiums they previously placed based on a hardware-stacking logic. Going forward, the semiconductor sector may enter a period of divergence, with different manufacturers being affected to varying degrees. Capital flows may shift from chasing rallies to validation, and short-term volatility is likely to increase.

Meta Platforms ranked #4, rising 2.03% with $12.062 billion in shares traded. Meta is further moving into fossil-fuel-based infrastructure to expand what may be its most electricity-intensive AI projects, highlighting the scale at which AI infrastructure is being built. According to the utility company, Meta has reached an agreement with Entergy to provide funding for an additional seven natural gas power plants related to its Hyperion data center in rural areas of Louisiana. This expands the scale of the three previously approved facilities further, bringing the total to ten power plants. These are expected to provide more than 7 gigawatts of power, making Hyperion Meta’s largest data center project.

The project’s scope appears to extend well beyond generation capacity. Entergy Louisiana said Meta will also fund the construction of about 240 miles of transmission lines, connecting northern and southern Louisiana with Arkansas. At the same time, it will invest in increasing capacity for battery energy storage and nuclear power plants. The data center itself is expected to provide about 5 gigawatts of computing power, and the additional capacity will support a wider campus area. After the news was released, Entergy’s stock price rose as much as 7.2%, hitting a record high. This suggests investors may have responded positively to the visibility of long-term electricity demand driven by AI expansion.

Ranked #5, another memory concept stock, SanDisk, fell 7.04% with $11.947 billion in shares traded. On Monday, U.S. stocks for optical communications and storage concept stocks generally declined.

Ranked #9, Broadcom fell 2.42% with $8.187 billion in shares traded. A Broadcom executive recently said that its manufacturing partner TSMC’s production capacity is nearing its limit, and with a surge in demand for AI chips, the entire industry is facing more widespread supply-chain constraints. “We see that TSMC’s capacity has reached its limit,” Natarajan Ramachandran, Broadcom’s director of product marketing for the physical layer product division, told reporters. “They will increase capacity before 2027, but this has become a bottleneck, or in 2026 it will already be impeding the supply chain,” he added.

Ranked #12, Palantir fell 3.85% with $5.633 billion in shares traded. The stock is down 22.6% year to date.

Ranked #14, TSMC fell 3.13% with $4.972 billion in shares traded. Recently, there were reports that TSMC’s 3nm capacity has reached its limit, leaving it only able to guarantee demand from long-term key customers and cloud AI giants.

Ranked #15, Lumentum fell 6.82% with $4.895 billion in shares traded. On Monday, U.S. stocks in the optical communications and storage concept sectors generally declined.

Ranked #16, Intel fell 4.50% with $3.481 billion in shares traded. Recently, Intel and AMD each notified customers that they would raise prices across their entire CPU lineup starting in March and April, respectively. Multiple people with knowledge of the matter said CPU quotes have been increased multiple times since the beginning of this year, with average price rises of 10% to 15%, and some products seeing even higher increases. At the same time, CPU delivery lead times jumped from the previous one to two weeks to an average of eight to twelve weeks, and in some cases even up to six months.

(Screenshot from Sina Finance app App行情-U.S. Stocks-Stock Market Sectors; swipe left for more data) Download the Sina Finance app

A massive amount of information, precise interpretation—everything is in the Sina Finance app

责任编辑：张俊 SF065