Foreign investors dumped nearly $2.5 billion in Japanese equities during the week ending May 29, according to Japan Exchange Group data reported by Bloomberg. The sell-off followed the Nikkei 225 breaching 65,000 points for the first time—a rally fueled almost entirely by AI-linked stocks like SoftBank Group, Kioxia Holdings, and Murata Manufacturing.

For data and ML engineers watching infrastructure valuations, the pattern is familiar. The Nikkei’s rapid ascent has triggered profit-taking, with Pelham Smithers, managing director at Pelham Smithers Associates, estimating that 70% of Japan’s 2026 market growth will come from AI ecosystem companies. “There’s a growing sense that AI is turning into a bubble,” he said, noting global investors are pivoting to less AI-saturated markets like Europe.

The International Monetary Fund warns the current AI boom risks a collapse comparable to the 2000-2001 dot-com crash, driven by overly optimistic expectations. From an engineering perspective, the question isn’t whether AI will transform workloads—it already has—but whether current valuations reflect sustainable infrastructure demand or speculative overhang. The sell-off suggests institutional investors are hedging against a correction in compute and data pipeline valuations, particularly in markets where AI exposure is concentrated in a few high-beta names. For practitioners, this signals a need to decouple engineering roadmaps from market hype: build for latency and cost efficiency, not stock price momentum.

Source: Lenta.RU