Sci-Tech and semiconductors are weakening across the board this time: the chip ETF — until now the only one of the four still up for the year — has flipped to −2.0% year-to-date and fallen another −30.9% over the past month, so all four lines are now down for the year. The real laggard is the STAR 50 Index itself, down −13.6% year-to-date, −47.4% off its high, with this week's decline reaching −9.7%; its year-to-date loss is closing in on the AI ETF (−13.7%), with the two now nearly even — but the biggest decliner among the four is still the optical module ETF (−17.1%). Constituent breadth (the share of stocks above the 200-day moving average) has held at 3/50 stocks (6%) for two straight days without falling further alongside price, but the looser 50-day line slipped this time to 8%, one fewer than before. Import substitution: confirming the other way; semiconductor sales: not triggered. Next to watch: whether price drags 200-day breadth down with it.
Watch two things: whether domestic equipment/materials' share on leading fabs' production lines has turned down, and whether top-tier fabs' orders have systematically shifted back to overseas supply chains. Single-month swings, stock price moves, and short-term news about one company don't count.
Watch the SIA global semiconductor monthly sales, year-over-year — only a decline sustained for 2–3 consecutive months counts as a peak. A single month's pullback, month-over-month noise, or a standalone cycle in one category (e.g., memory) don't count.
This time the four lines aren't about who gained more — it's about where the money is moving. The STAR Chip ETF just flipped: its year-to-date reading went from a gain to a loss, now at −2.0%, down another −30.9% over the past month — all four lines are now in the red for the year, with none left unscathed. The real leader of the decline is still the STAR 50 itself: it closed at 1,161, a fresh 52-week low, down −13.6% year-to-date and off −47.4% from its 52-week high — a bigger move today than any of the other three lines. The Optical Module ETF still has the steepest year-to-date drop of the four, now at −17.1%; the STAR 50's own year-to-date loss has nearly caught up with the AI ETF (−13.7%), with the gap narrowing from about 1.8 percentage points to about 0.1 — essentially even. All four lines remain well below their year-to-date highs: what's changed today is the STAR 50's role, from a sideline benchmark to the single most volatile line of the day — that doesn't mean the thesis has changed.
The index is propped up by heavyweight stocks; breadth tracks how most stocks are actually doing. When breadth (the share of stocks above their 200-day moving average) drops below 30% while the index barely falls, it means only a handful of heavyweights are rising; a real recovery needs breadth and the index to turn up together. The number to watch most on this page is the share of constituents above their 200-day moving average, which has held at 3/50 (6%) for two days straight — further price declines haven't dragged it down, but it hasn't recovered either; instead, the looser 50-day line slipped this time to 8%, one fewer company holding above it. The STAR 50's weekly decline widened to −9.7%, and price hit a fresh 52-week low today — a divergence from the 200-day line's steady reading, as price fell while breadth didn't follow this time. The good news is that falling prices and deteriorating fundamentals are two different things: domestic equipment makers' order share hasn't been affected by this bout of volatility. What to watch next: whether the 200-day line eventually gets dragged down by price, or price stabilizes first. Current readings:
Not an opinion — closing prices land and breadth and IPO readings recompute themselves.
This page is just a snapshot — the dashboard itself is live. The full version includes the complete definitions of both theses — what to watch, what doesn't count, the sources to check, and the update frequency — along with price tables for the STAR/chip/AI/optical module ETFs, constituent breadth, and STAR Market IPO details.
Current: domestic substitution confirming the other way. Q3 earnings season landed in bulk: AMEC and Hwatsing revenue grew 20-30%, while ACM Shanghai's order backlog rose against the trend, up 88%; but most of the surge in AMEC's and Piotech's net profit came from paper gains on equity stakes they hold in other companies, not from selling more or pricier equipment — the one whose profit actually doubled on core business alone was Jiangfeng Electronic; only Naura's growth slowed in Q3, with no sign yet of orders shifting overseas. Semiconductor sales not triggered (YoY growth still rising). The real variable is price: the STAR 50's weekly decline widened to −9.7%, hitting a fresh 52-week low intraday, with its year-to-date −13.6% and gap from the high −47.4% both widening in step. Breadth among constituents has split: the share of stocks above the 200-day average has stalled at 6% (3/50) — not falling further with price, but not recovering either; the 50-day measure has slipped this time, down to 8%. Equipment makers' revenue and orders are still rising while their stock prices fall — the business isn't getting worse, the price the market is now willing to pay is just lower. Next, watch whether price can recover in step with orders.
An inflection is a broken thesis, not a moving price. Price noise is noise; only broken logic is a signal.
If neither thesis has been triggered, but constituent breadth stays below three in ten for a whole quarter and the STAR 50 keeps weakening — that means the market is seeing something the thesis hasn't yet captured, and it's time to add a signal, not keep waiting.