CPU, memory, optical communication Wall Street interprets AI stock's epic frenzy: the market is eager to chase the "bottleneck"!

Over the past few weeks, AI-driven enthusiasm in the U.S. stock market has sparked widespread discussion. Analysis shows that the semiconductor-related stocks that have driven the S&P 500 and Nasdaq to record highs share a common trait: they are all positioned at critical bottlenecks in the artificial intelligence (AI) industry chain.

GPU and CPU

In the past, what was hard to obtain was NVIDIA’s graphics processing units (GPUs)—the main engine behind the AI boom.

But with the development of agent-based AI, market demand for central processing units (CPUs) is surging, as AI agents rely on CPUs to perform tasks and generate outputs. Unlike chatbots that only respond to prompts, AI agents can autonomously handle tasks for hours.

“Over the past few quarters, we’ve seen a new demand driver—agent-based AI workloads, which are better suited for server CPUs rather than GPUs,” said John Vinh, an analyst at KeyBanc Capital Markets, last week.

This is a major positive for CPU manufacturers like Intel and AMD, whose stock prices are currently hovering near all-time highs. The demand for CPUs is so strong that NVIDIA also launched its own Vera CPU in March, aiming to carve out a share in the data center CPU market.

“Market investors—whether hedge funds, retail investors, or others—love chasing bottlenecks,” said Angelo Zino, an analyst at CFRA, last week.

Memory Chips

Another increasingly prominent bottleneck is storage chips in AI infrastructure.

As memory demand soars, Micron Technology’s stock hit a record high last week. The world’s largest storage chip manufacturer, Samsung Electronics, joined the trillion-dollar market cap club last week. SK Hynix’s stock price is also at a historic high.

Tech giants like Meta, Microsoft, and Apple are all discussing rising memory costs.

“Memory manufacturers have been able to sign long-term agreements with hyperscale cloud providers, and we believe this will support a reevaluation of storage stocks’ valuations in the future,” Vinh said.

Analysts say that the industry boom in large-scale production of high-bandwidth memory (HBM) for AI chips is reshaping the entire semiconductor supply chain. Storage chip stocks like SanDisk have also surged, with gains exceeding 400% this year.

Optical Communications

The third bottleneck exists in optical communication. The industry is shifting from traditional electrical signal transmission to transmitting data within chip infrastructure using photons.

Last week, NVIDIA announced a partnership with Corning. The chipmaker also invested in Coherent and Lumentum. The stocks of these optical communication companies are all at historic highs.

Bullish investors point out that the AI-driven cycle could extend this tech bull market. The industry has just begun exploring robotics and autonomous systems, which are expected to become the next major drivers of AI.

Zino from CFRA noted that these bottlenecks are likely to persist at least until 2027, or even longer, especially in memory.

Wall Street’s Overall Optimism but Growing Warnings

Wall Street generally expects the U.S. stock market to continue rising this year, driven by strong earnings growth.

Last week, Ahmed Riesgo, CIO of Insigneo, said he sees growth opportunities in companies like Alphabet, thanks to investments in TPU chips, cloud services growth, and Gemini AI products.

UBS analysts also stated last week that technology stocks remain a way to continue investing in the market.

“Investors can also expand their scope to include tech giants across the AI value chain, covering enabling, intelligent, and application layers, including semiconductor and chip manufacturing equipment, power and resources, and infrastructure,” UBS analysts wrote.

On Friday, Royal Bank of Canada Capital Markets raised its year-end target for the S&P 500 from 7,750 to 7,900 points, citing strong earnings growth and the continued strength of AI-related industries.

However, some Wall Street professionals have issued warnings about the AI-driven frenzy in U.S. stocks. Michael Burry, the investor who accurately predicted the 2008 US subprime mortgage crisis and the protagonist of the movie “The Big Short,” warned last week that the current market enthusiasm for AI resembles the final stages before the dot-com bubble burst in 2000.

Long-term market bull Jim Paulsen, chief investment strategist at Leuthold Group, also warned that the recent rally in tech stocks might not last much longer, citing “four major bearish signals,” including corporate cash flow depletion.

(Original source: Caixin)