The diverging fortunes within the AI stock market are intensifying, prompting a market analyst to advise a shift away from cash-burning tech giants and toward semiconductor companies.

Ben Reitzes, head of technology research at Melius Research, is recommending investors steer clear of major cloud service providers such as Alphabet (GOOG.US, GOOGL.US), Meta Platforms, Inc. (META.US), and Amazon.com (AMZN.US), citing their inability to generate substantial free cash flow. "I remain negative on the hyperscale data center operators for one simple reason: they cannot produce meaningful cash flow," Reitzes stated in an interview. "Who cares? Buy the chip companies."

This sentiment comes after a brutal sell-off on Thursday that wiped nearly $800 billion from the combined market capitalization of the "Magnificent Seven" stocks. The Nasdaq 100 index fell 1.9%, with Alphabet shares plunging 7% and Tesla Motors (TSLA.US) crashing roughly 15%, marking their worst single-day performances in over a year. The market's harsh reaction followed their earnings reports, which revealed massive capital expenditure plans and a sharp turn to negative free cash flow, drawing intense scrutiny from investors.

In stark contrast, chip stocks like NVIDIA (NVDA.US) and Micron Technology (MU.US) moved higher. Reitzes argues that investors should focus less on the rising capital expenditure numbers and more on the margin pressure these investments create. "The increase in capital expenditure was largely in line with expectations. I believe this is a positive signal for the semiconductor industry," he explained.

Reitzes elaborated that Alphabet's recent margin weakness is directly attributable to cloud computing costs, data center expansion, and accelerated capital spending. He warns that the "margin story" for these hyperscale operators is just beginning and will soon pressure Meta, Amazon, and Microsoft (MSFT.US) as they report their own earnings. Despite investor anxiety and falling stock prices, Reitzes believes these tech giants are too deeply committed to retreat from their massive AI infrastructure investments, as the risk of stopping is now too great.

Reitzes highlighted Meta as the hyperscaler he is watching most closely, expressing concern over its frequently shifting AI strategy. "Their strategy seems to change on a daily basis, or perhaps no one knows what it is," he said. He hopes these companies can quickly convert their enormous capital investments into clear, tangible revenue streams.

Where the AI Rally is Splitting: Chipmakers Up, Cloud Giants Down

The AI market in 2024 is exhibiting an extreme structural divergence. "Pick-and-shovel" chip suppliers, who hold the core production capacity, are enjoying a strong rally. Meanwhile, the tech giants—the "shovel users" who are aggressively purchasing computing power and building AI infrastructure—are being sold off by the market. Year-to-date, the Roundhill Magnificent Seven ETF (MAGS) is down 4%, while the Philadelphia Semiconductor Index (SOX) has surged 74% over the same period.

This internal split within the AI theme is drawing widespread attention. JPMorgan strategists have warned that the current pattern of "strong chips, weak clouds" is highly reminiscent of the final stages of the 1990s internet bubble. In a recent report, strategist Jason Hunter cautioned that if the hyperscale cloud giants fail to break through key technical resistance levels, while the semiconductor sector also remains below its own critical resistance, "what begins as a rotation within the AI theme could evolve into a more concerning, systemic breakdown."