GuoTou Securities analyst Lin Rongxiong argues that A-share AI tech stocks have not yet formed their first M-top peak, with the second top — typically 10% to 15% lower — being the true sell signal.
GuoTou Securities analyst Lin Rongxiong argues that A-share AI tech stocks have not yet formed their first M-top peak, with the second top — typically 10% to 15% lower — being the true sell signal.

China's AI tech rally has not yet reached its first M-top peak, according to GuoTou Securities, as key signals including trend-line breaks and valuation overextension remain absent despite rising rotation pressure.
"All trends are ultimately driven by crowding," Lin Rongxiong, a strategy analyst at GuoTou Securities, said in a June 4 report. "The first top is rarely the right exit — the second top is where the real call is made."
Historical M-top patterns show the first peak typically forms when single-quarter earnings growth is at or near its high, while the second peak aligns with the trailing-12-month earnings high, usually arriving one to two quarters — roughly six months — later. The second top historically sits 10% to 15% below the first, except during the 2015 leverage-driven bull market.
The framework matters for the roughly $1.2 trillion A-share AI sector because premature selling risks leaving 10% to 15% of upside on the table. Lin identifies five pre-peak signals: peak single-quarter earnings, extreme trading congestion, core leaders undergoing a final valuation re-rating, a "large-to-small-to-large" rotation pattern, and capital crowding into the theme.
Five Signals That Precede the First Top
Lin's framework identifies five conditions that historically precede a first M-top peak. The first is single-quarter earnings growth reaching a cyclical high — as seen with Kweichow Moutai and Contemporary Amperex Technology Co. Ltd. before their respective peaks. The second is extreme trading congestion: valuation percentiles, institutional ownership, and turnover ratios all rising simultaneously. The third is core leaders undergoing a final valuation re-rating phase, where large-cap names make their last leg up before the broader market turns. The fourth is a "large-to-small-to-large" rotation pattern, where blue chips lead first, then mid- and small-cap names catch up, then large caps lead a final push. The fifth is capital crowding, where money flows so heavily into the theme that other indices decline.
What's Missing in Today's AI Rally
Applying the framework to current conditions, Lin argues that several key signals are not yet flashing red. The representative AI index has not broken below its 60-day moving average — a threshold that historically distinguishes healthy rotation from trend reversal. No macro gray rhino event has emerged to disrupt the industry narrative. Core AI leaders have not yet undergone a final valuation re-rating; second- and third-tier names have outperformed the large caps, the opposite of what typically happens near a peak. Earnings overhang for most AI-related companies remains below three years, far from the three-to-five-year levels seen at the 2021 New Energy Portfolio and 2015 tech bubble peaks.
The one signal that has already lit up is capital crowding. Money is concentrating into AI names at the expense of other sectors, a pattern that historically precedes the first top. But Lin says one signal alone is insufficient to call the peak.
The Real Constraints — CapEx and Competition
Two structural factors will determine the sell point, according to Lin. The first is the capital expenditure cycle: AI-related CapEx remains difficult to disprove over the next six months, but sustainability beyond 2027 is a key watchpoint. The second is competitive dynamics: if the industry's competitive structure deteriorates, even strong demand growth would become a sell signal.
"Valuation-based selling is a left-side call — you may miss the final re-rating leg," Lin said. "Industry-based selling requires watching for macro gray rhinos and competitive breakdowns. Neither condition is present today."
For investors holding A-share AI exposure, Lin's framework suggests the current rotation pressure is a trading phenomenon, not a structural reversal. The key thresholds to monitor are a 60-day moving average break, a macro shock, or core leaders beginning a final valuation push — any of which would shift the call from "first top not yet arrived" to "first top forming now."
This article is for informational purposes only and does not constitute investment advice.