DA Davidson: The market is seriously "mispricing" the AI cycle, with Microsoft more deserving of the "AI crown" than Google

According to the Zhitong Finance APP, Gil Luria, head of technology research at D.A. Davidson, stated on Monday that the technology sector is being distorted by extreme valuation "misalignment." The pricing of some semiconductor and hyperscale cloud service stocks seems to have already reflected an AI boom that will last until 2030, while other stocks are being priced by the market as if the AI cycle is peaking right now.

"Is the AI cycle peaking now, or will it continue until 2030? This is a very difficult question to answer," Luria said. Rather than trying to determine the turning point of the cycle, he focuses on identifying areas where the market has created significant valuation discrepancies among comparable companies.

The valuation gap between memory chips and CPUs is the most pronounced. Semiconductor equipment manufacturers and optical device companies are being priced by the market as if AI infrastructure construction will last for years, while the stock price trends of Micron (MU.US) and even Nvidia (NVDA.US) seem to suggest that the AI cycle is about to end.

The data comparison is striking: Micron currently has a price-to-earnings ratio of only 8 to 9 times, while CPU-related stocks have P/E ratios as high as 40 to 50 times. Luria believes that this pricing logic is fundamentally flawed.

"If the AI cycle really extends to 2030, Micron's current valuation could have nearly four times the upside potential," he said. "The strategic importance of memory chips in the current market has surpassed that of CPUs, and the competitive landscape in the memory market is far less intense than in the CPU market. Therefore, this is a significant valuation divergence, and we believe it represents the biggest investment opportunity right now."

He also pointed out that Intel (INTC.US) and Cerebras (CBRS.US) can only support their current valuations if the AI cycle continues for five years—if the cycle turns now, the fair value of both companies will be far below their current trading prices.

Similar valuation differentiation is also evident among hyperscale cloud service providers. The market is lenient towards Google's and Amazon's AI capital expenditures, while Microsoft is penalized for having similar levels of capital expenditure.

Ironically, Luria noted that Microsoft's backlog of AI computing orders is 50% higher than Google's, indicating that its actual sales of AI infrastructure far exceed those of Google. However, Microsoft has just experienced its worst monthly performance in nearly two decades, while Alphabet and Amazon recorded gains during the same period.

"The pendulum has swung too far in favor of portraying Google as the AI winner and Microsoft as the loser," Luria said. "In fact, both are winners."

A year ago, the situation was completely opposite—Google had a P/E ratio of about 18 times, while Microsoft's was as high as 30 times. Now, the market has overcorrected.

Luria also refuted the notion that "Microsoft's software business faces an existential threat from AI." He believes that Microsoft is being punished for its overly aggressive AI spending, while simultaneously being discounted by the market as a "traditional software company that will be disrupted by AI."

"We all know that five years from now, we will still be using Outlook, Teams, Word, and PowerPoint—AI agents will also use these tools," he pointed out, emphasizing the strong user stickiness of Microsoft's enterprise software. "Yet the current stock price trend of Microsoft seems to suggest it is a software company that will be eliminated by AI." ”