CITIC Securities says the recent tech equity correction reflects forward pricing
risks in AI-related names, not solely high US long-term yields. Three key
narrative variables determine valuation: whether commercialization speed and
market scope meet expectations; whether compute advantage yields share and
pricing power; and whether current compute gaps will widen future AI model
divergence. The current consensus concern centers on commercialization speed and
scope; the largest latent risk is whether "anti-distillation" will re-widen
model gaps. On macro, the US Treasury’s announced long-note buybacks have
limited impact; near-term dollar weakness and weaker rate-hike expectations
support narrowing of K-shaped divergences in global markets. However, drivers of
sustained US long-end yield rises remain intact and further volatility is
possible. Given external shocks and the A-share market’s short-term funding
structure, market complexity is increasing; investors should manage expectations
and avoid broad, conviction-heavy narratives in this volatile phase.
