Outshining NVIDIA! Dual CPU Giants in the U.S. Stock Market Reach New Highs as AI Attributes Gain Market Recognition

①Taiwan Semiconductor stated: “CPUs are becoming increasingly important in AI data centers.” ②The head of AMD’s data center business noted that there are no signs of a slowdown or halt in CPU demand in the short term. ③A TrendForce research report indicated that the CPU-to-GPU ratio in data centers is expected to evolve to between 1:1 and 1:2 in the era of intelligent AI agents.

According to the STAR Market Daily on April 17, “CPUs are becoming increasingly important in artificial intelligence data centers.” This statement was made during the Q1 earnings call held yesterday. $Taiwan Semiconductor (TSM.US)$ as stated.

The company stated that, due to previously being unable to identify the specific use of sold CPUs, it had not yet included CPU sales revenue in its statistics for AI HPC (high-performance computing) income, but may consider doing so in the future.

It is not just Taiwan Semiconductor; the capital market is also repricing CPUs. Last night, the two leading CPU stocks in the U.S. market surged simultaneously. $Advanced Micro Devices (AMD.US)$ Advanced Micro Devices closed up 7.8%, hitting its highest closing price since October 29, 2025.

$Intel (INTC.US)$ Intel closed up 5.48%, continuing to set new all-time highs.

It is worth noting that the cumulative gains for these two stocks this month have reached 36.78% and 55.22%, respectively, indicating a clear upward trend.

Meanwhile, NVIDIA, once regarded as the “king of computing power,” $NVIDIA (NVDA.US)$ has seen a cumulative increase of only 13.73% this month. Vanda Research noted in its latest report that retail investors seem to be selling off shares of NVIDIA, a former “darling of artificial intelligence.”

Behind the divergence in stock prices lies the tangible demand for CPUs driven by AI. AMD CEO Lisa Su emphasized more than once in 2026 that CPUs have been completely sold out. She projected that the annual growth rate of the data center business could exceed 60% over the next three to five years. Forrest Norrod, head of AMD’s data center business, went so far as to state that there are no signs of any slowdown or cessation in demand in the short term.

A TrendForce research report noted that the CPU-to-GPU ratio in AI data centers is approximately 1:4 to 1:8, but in the era of intelligent AI agents, this ratio is expected to evolve to between 1:1 and 1:2.

This year, due to persistent supply-demand imbalances, AMD and Intel announced price increases in March and April, respectively. Since the end of February 2026, the average price of CPUs has risen by 10% to 15%, with delivery times extending from one to two weeks to eight to twelve weeks, and in extreme cases, potentially lasting up to six months.

The industry’s bright prospects have even attracted other chip giants: in March this year, NVIDIA and Arm successively entered the CPU market. The former unveiled the Vera CPU at the GTC conference, featuring 88 self-developed Olympus cores, with deliveries planned to begin in the second half of 2026. The latter launched the AGI CPU, aiming to provide core computing power support for the next generation of AI infrastructure.

Huaxin Securities pointed out that the AI Agent era is reevaluating the value of CPUs. Artificial intelligence is transitioning from large language models (LLMs) that answer questions to autonomous agents, a paradigm shift that fundamentally reconstructs the supply and demand landscape of computing power. CPUs, once considered the taken-for-granted foundational infrastructure, have now become a frontline critical factor determining AI response speed and costs. In the Agent era, CPUs will need to handle three core tasks:

The sandbox VM serves as the ‘digital body’ of the Agent, executing file I/O, code interpretation, process isolation, and other resource-intensive tasks.

Supports high concurrency and long online characteristics (sessions transitioning from minutes to hours, with exponential growth in concurrency demands).

In a storage-compute separation architecture, it acts as the ‘second GPU memory,’ handling retrieval, filtering, and data forwarding.

From an investment perspective, Caitong Securities believes that benefiting from continuous updates in server CPUs, there is potential for a gradual increase in PCIe generations and channel counts compatible with server CPUs, driving incremental demand across related industrial chains. Recommended companies to watch include: Foxconn Interconnect Technology (PCIe slot supplier), Montage Technology (PCIe Retimer chip supplier), and Wantong Development (acquirer of PCIe Switch chip supplier Shudo Technology).

038.pngLooking to pick stocks or diagnose stock performance? Want to know the opportunities and risks in your portfolio? For all investment-related questions,just ask Futubull AI!

Editor/Jayden



Source link

Visited 1 times, 1 visit(s) today

Related Article

0
Would love your thoughts, please comment.x
()
x