Chinese tech stocks lose nearly half trillion
A significant reversal has taken place in China’s technology sector, with the STAR50 index, which is heavily weighted towards AI and semiconductor stocks, down about 40 per cent from its July high in less than four months and an estimated US$480 billion ($682 billion) in market value erased. The wider CSI300 blue-chip index has also fallen about 20 per cent from its June peak.
According to Filip Tortevski, senior analyst at Wealth Within, investors have reassessed the prices they were willing to pay for AI-linked shares as concerns about future demand, tax bills and possible overseas restrictions on Chinese equipment added pressure. Tortevski noted that expectations had run ahead of company earnings, leading to a correction in the market. “I think that expectations ran well ahead of earnings, leaving prices vulnerable to a change in sentiment. Tax bills and fears of overseas restrictions may have added selling pressure, but I’d put stretched valuations at the centre of it,” Tortevski said.
Tortevski suggested that the correction may reflect speculative excess unwinding while the industry itself continued to grow. However, he warned that technological promise did not guarantee returns for shareholders. “AI could transform the economy and investors could still lose money buying the wrong companies at the wrong prices,” he said.
Valuations remain high despite the fall
The STAR50 trades at about 100 times trailing earnings, compared with roughly 30 times for the Nasdaq 100. Dr Liew Chee Yoong, associate professor of finance at UCSI University Malaysia, observed that the two scenarios can coexist, with speculative expectations potentially deflating while valuations remain high. “The two descriptions can coexist. My interpretation is that speculative expectations may be deflating while valuations remain demanding,” Dr Liew said.
Dr Liew cautioned that comparisons between indices must be considered in context, taking into account differences in businesses, growth rates, and earnings maturity. “At 100 times trailing earnings, investors are still paying heavily for future profit growth. A large fall does not automatically make stocks cheap,” he said. Additionally, Dr Liew noted that the reported decline in share prices and valuations needs to be reconciled before the decline can be attributed solely to a collapse in valuations.
Related Post: Anthropic eyes Australian market for investors
A fall in the price-to-earnings multiple from nearly 250 to 100 would amount to a 60 per cent contraction, while the reported share-price decline was 40 per cent; if the periods and methods matched, that difference would imply aggregate earnings had risen by about 50 per cent. Dr Liew said that determining the relative impact of various factors on the decline would require access to investor trading data, fund flows, margin balances, and order-book evidence. The share-price drop alone does not provide insight into who has been selling, as various factors such as retail investors unwinding speculative positions, fund redemptions, and leveraged investors reducing exposure could all have contributed.
Fund redemptions can force sales
Dr Liew noted that “fund redemptions can force managers to sell overlapping holdings, depressing prices for other funds and potentially triggering further withdrawals.” He emphasized that the critical issue is the balance between the amount investors need to sell and what buyers can absorb. “The critical issue is how much investors need to sell relative to what buyers can absorb,” he said.
Tortevski recommended that position sizes reflect the losses an investor can tolerate, noting that a 10 per cent allocation that halves would reduce a portfolio by roughly 5 per cent before other movements. Investors should also add their exposure across direct shares, ETFs, and managed funds to identify repeated bets.
US tech results will be key
Tortevski said that the decline in China’s technology sector does not automatically predict the same outcome for US technology shares, because major US companies have established earnings engines. However, forecasts that the five major US hyperscalers would spend more on infrastructure this year than they generate in operating cash flow make this week’s Microsoft, Alphabet, and Amazon results particularly important. Dr Liew noted that the shared risk across markets was that AI infrastructure spending could grow faster than the profits it ultimately generated, while stressing that Chinese and Western companies had different exposures. “The useful comparison is between valuations and credible future cash flows, rather than between national index performances alone,” he said.
