Investors say Japanese stocks that once enjoyed valuation premiums due to artificial intelligence (AI) business exposure now look overly risky.
NVIDIA's (NVDA.US) CEO Jensen Huang's abrupt departure has further intensified market concerns about the entire industry's outlook.
After the U.S. midterm elections, global AI stocks suffered a sell-off. With the leadership change at NVIDIA, market sentiment toward this sector deteriorated further, and the selling momentum accelerated.
Japanese AI-related stocks were equally hit by this pullback. Since Huang's resignation, chip-related companies Kioxia Holdings and Advantest have both seen their share prices drop more than 9%, while the Topix index fell only 2.9% over the same period.
Market concerns are increasingly shifting toward corporate earnings. Many Japanese chip and AI-related companies' earnings forecasts are still based on assumptions that data center and other AI infrastructure investment will maintain strong growth. If such spending slows, these companies will be vulnerable to earnings estimate downgrades.
Yoshitaka Suda, cross-asset strategist at Nomura Singapore, said AI-related stocks are facing a "triple blow," specifically including concerns about deteriorating profitability, expectations of tighter regulation after the U.S. midterm elections, and the impact of internal turmoil at NVIDIA. He said that under these circumstances, investors may increasingly demand valuation discounts for AI stocks rather than being willing to pay premiums for them.
Hiroaki Tomori, executive fund manager at Mitsubishi UFJ Asset Management, said: "Some AI-related stocks used to enjoy an AI premium, but they have now entered an AI discount phase, where merely being associated with AI could lead to valuation discounts." He added that some large AI stocks are still trading at premium valuations and may therefore be vulnerable to further selling. In contrast, stocks that have already undergone corrections may be approaching a bottom.
Currently, analysts are still betting that AI investment will not slow down. According to compiled data, capital expenditure by six major U.S. technology companies is expected to reach $885 billion this year and grow another 64% next year to $1.45 trillion.
This week, NVIDIA gave quarterly revenue guidance above Wall Street expectations, attempting to reassure investors that market demand remains strong after Huang's departure. Capital expenditure by U.S. technology companies
However, valuation shifts in the Japanese market have already emerged. The forward 12-month expected price-to-earnings ratio of the Tokyo Stock Exchange non-ferrous metals index, which includes cable manufacturers, as well as the forward 12-month expected price-to-earnings ratio of the MSCI Japan semiconductor index, have both fallen below their end-March levels—when the AI rally had not yet accelerated.
Currently, both indices' valuation multiples are lower than those of the broader Topix index. This makes expectations for future earnings crucial.
Hiroaki Tomori said a 20% to 30% decline in data center investment cannot be ruled out.
Kioxia's situation illustrates that judging whether this sell-off has already gone too far is not easy. The chipmaker's current forward 12-month expected price-to-earnings ratio is less than 2 times, a level that fund managers generally consider to indicate cheap valuation. But this valuation is based on still-optimistic earnings forecasts. If the AI investment boom slows and corporate profit expectations are downgraded accordingly, then Kioxia's valuation may be far less cheap than it appears on the surface.
Comments