Asian AI Stocks Face Turbulent Recalibration as Optimism Fades

HSBC has drawn parallels between the recent performance of Asian AI stocks and the post-peak adjustment observed in China’s internet sector during 2020-2021. The optimism surrounding the AI sector is waning, as substantial investments in AI have yet to yield clear financial returns. This shift in sentiment is leading to downward revisions in target prices and earnings forecasts for these stocks.

Since July 2024, AI-related stocks in Asia have experienced a significant decline, with HSBC’s custom Asian AI Index dropping by 15% from its peak on July 11. This represents the most severe sell-off since the launch of ChatGPT in 2022. As AI stock prices recede rapidly, HSBC analysts believe that the sector may have reached its peak and entered a new phase of adjustment.

The HSBC report highlights three primary challenges facing the Asian AI market: weakening demand, regulatory pressures, and downward earnings revisions. The latter has only recently begun to impact the sector. Similar to the 2020-2021 trend in Chinese internet stocks, the downgrades in earnings expectations and corresponding adjustments to target prices have already signaled a shift.

Despite 89% of AI stocks still holding a “buy” rating, analysts are adopting a more cautious approach towards their earnings outlook. Earlier this year saw an upswing in target price assessments, but since July, these have been trimmed by 2%. Given some targets are based on overall industry benchmarks, further significant reductions are anticipated in the weeks ahead.

There's particular uncertainty regarding when AI's substantial investments will pay off. Although tech giants are heavily investing in AI infrastructure, doubts linger over how these investments will be monetized. For instance, NVIDIA’s CEO projected a $1 trillion global AI infrastructure investment over the next four to five years, but current demand hasn't aligned with such projections.

Strategic investments from major tech companies, with capital expenditures up by 50% year-over-year in Q2 2024 for significant players, far outpace the existing demand. Questions arise over the adequacy of market demand to justify these investments.

Further complicating matters are potential regulatory risks. The significant energy consumption of AI data centers poses challenges, with regions like Europe introducing stricter data privacy and AI regulations. The U.S., too, has initiated antitrust investigations, such as against NVIDIA, and the massive electricity demands highlighted by ChatGPT’s usage are seen as future regulatory hurdles.