We should worry — but measure first. China’s listed companies closed the half-year disclosure window in late August with a headline that looks like a clear tide: 5,550 companies posted combined revenue of 37.74 trillion yuan, up 7.61 percent, and attributable net profit of 3.58 trillion yuan, up 19.4 percent — the first time the profit total has crossed 3.5 trillion, and the fastest growth since 2022. Nearly three quarters of the companies, 74.54 percent of them, were profitable.
Those are the numbers a press release would lead with. The steady way to read them is to ask what is doing the work, because a headline profit number can hide a narrow current as easily as it can reveal a broad one. When I measure the composition carefully, the story is not the tide level. It is where the current is running.
The headline, measured
Let me be precise about what the aggregate says. Revenue up 7.61 percent against net profit up 19.4 percent means margins expanded, not just volumes — profit grew more than two and a half times faster than revenue. That is the first meaningful signal in the file: the market is not just selling more; it is keeping more per unit sold. That combination is usually the signature of pricing power or cost discipline, and both matter.
The 74.54 percent profitability share is the second useful measurement. It says the profit is not coming from a handful of outliers dragging a flat average — three in four companies actually made money. So the breadth is real at the level of the profit-or-loss line. What the headline does not tell you is how unevenly the growth inside that profitable half is distributed, and that is where the composition becomes the story.
The narrow, fast currents
Pull the growth out by segment and the picture sharpens immediately. The STAR Market — the board built for hard-tech and biotech listings — reported revenue of 1.01 trillion yuan, up 38.59 percent, and net profit of 1,448.87 billion yuan, up 437.59 percent. A 437 percent profit growth rate in one reporting period is not a trend; it is a wave, and it is moving through the technology and innovation companies at speed.
Inside that wave, the fastest current is semiconductors. The electronics industry as a whole posted net profit up 195.11 percent; the semiconductor segment specifically was up 607.2 percent. Let me sit with that number for a moment, because it deserves more than a glance. A 607 percent jump in half-year profit is the kind of figure that either marks a genuine structural change or a low base from the prior year, and in this case the two are connected: the segment is coming off a painful down-cycle, and the recovery is real but it is also steep partly because the starting point was deep. Both things are true, and a careful reading holds both.
I started writing this piece from the aggregate angle — total revenue, total profit, the familiar headline order — and dropped it. The aggregate is fine, but it flattens the very thing that matters, which is the unevenness. The companies doing the heavy lifting are not spread across the market; they are concentrated in a thin band of hard-tech and semiconductor names, and the concentration is the story.
The export current
The second fast current is offshore. A-share companies reported total overseas revenue of 6.06 trillion yuan, up 22.68 percent, and that overseas share has risen to 16.05 percent of total revenue, up two percentage points from the year before. Measured against the overall revenue growth of 7.61 percent, the export line is running nearly three times faster than the domestic one.
That gap is the part I find most quietly significant. It says the marginal growth in the market is increasingly being earned outside China’s domestic demand — in export orders, overseas manufacturing, and foreign markets — and the companies positioned there are the ones compounding. The overseas share crossing sixteen percent is not a headline number, but it is a structural marker: the profit the market is reporting is progressively more global in origin.
Investing and paying at the same time
Then there is the double movement that is unusual enough to flag. A-share research and development spending crossed 800 billion yuan in the first half — the market is reinvesting heavily in innovation. At the same time, 867 companies announced interim dividends, planning to pay out 716.67 billion yuan to shareholders. Investing and distributing in the same reporting period is not what a company does when it is uncertain; it is what a company does when it has cash flow it trusts.
No, that is not quite right, and I want to correct myself before the point travels. The simultaneous rise in R&D and dividends does not mean every company is healthy; some are distributing to keep investors while margins are squeezed. But in aggregate, the two numbers together — 800 billion in R&D, 716.67 billion in planned payouts — describe a market that is both funding its own future and rewarding holders in the present. That combination, at market scale, is the kind of thing that tends to repeat in the years after it first appears.
What the composition warns about
A steady journal notes the warnings as well as the currents. The first is dependence: when growth is concentrated in semiconductors and exports, the market’s headline rate is hostage to two factors — the global semiconductor cycle and the health of overseas demand. If the chip cycle rolls over or trade conditions sour, the 19.4 percent aggregate is the first number to fall, because the concentration cuts both ways.
The second is the base effect hiding inside the growth rates. A 607.2 percent semiconductor number and a 437.59 percent STAR number both carry the imprint of a low prior-year base. The growth is real, but the slope is steeper than the underlying trend, and anyone who treats the slope as the new normal is measuring the wave, not the tide.
The third is breadth at the margin. Seventy-four percent profitability is reassuring, but the profit distribution inside that number is not visible in the aggregate. The export-dependent and chip-linked names are pulling the average up; the companies serving purely domestic, non-tech demand are reporting the quiet numbers. That is not a prediction of weakness; it is a reminder that the average is a summary, and a summary hides its own shape.
The third current: innovation beyond chips
There is a third current in the half-year file, and it is easier to miss than the semiconductor line because it does not have a headline percentage attached in the facts I am working from. The description of the season points to innovation-driven industries as a group — hard-tech and biotech along with computing — and the aggregate R&D number above 800 billion yuan is the marker for all of them. When a market spends that much on research in six months, the profit cycle it reports this year is, in part, an investment in the profit cycle of three years out. The companies spending on research today are borrowing from the current quarter to buy the next phase of the current.
I have seen this pattern in the marine world in miniature: a research program that pays for itself quietly for a decade, then produces a result that changes the picture for everyone downstream. The market’s 800-billion-yuan research line is that kind of program at scale. It is the least glamorous number in the report — no one will build a headline around it — and it may be the most consequential one, because it describes the market’s own opinion of its future.
Let me be careful not to overread it. R&D spending can be vanity, and some of that 800 billion will go to projects that never produce a return; every innovation wave carries its share of dead ends. But the direction of the spending — up, across a market that is also paying out record dividends — is the direction a confident market chooses. The two numbers together say more than either alone.
One more measurement belongs in the file, and it is the one investors feel before analysts see it: the breadth of the season. When a reporting period is genuinely broad, the number of companies beating their own prior-year earnings is itself a statistic. When it is narrow, the index rises while the median company barely moves. I do not have that median figure verified in front of me, so I will not quote one — but the composition of this report, with semiconductors at 607.2 percent and the aggregate at 19.4 percent, already tells you which shape this season has. The arithmetic is the evidence.
How I read the half-year ledger
So here is the honest read, in the order the evidence supports. The aggregate is genuinely strong: first-half profit above 3.5 trillion, up 19.4 percent, fastest since 2022, with healthy breadth in the profit-loss line. The engine of that strength is narrow: hard-tech and semiconductor names are growing at multiples of the market, and overseas revenue is compounding at nearly three times the domestic rate. And the market is doing something unusual — reinvesting more than 800 billion in R&D while paying out 716.67 billion in interim dividends — which is the signature of cash-generative, forward-looking firms.
I have been wrong about earnings season readouts before, and I want to say that plainly. Aggregate numbers are the easiest to misread because they hide composition, and composition is exactly where this report differs from a routine up-cycle. The signal that matters is not the 19.4 percent; it is the shape of the growth underneath it.
I will also admit what I do not know: I have not individually verified each of the 5,550 filings, and no reader should mistake an aggregate for an audit. What an aggregate can do is name the currents, and the currents are what I watch. On the evening the last report landed, I printed the composition table and pinned it above my desk — the 607.2 percent line in one column, the 19.4 percent total in another — because the gap between those two lines is where the next two quarters will be decided.
The current to keep watching
The number to watch going forward is the overseas revenue share, now at 16.05 percent. If it keeps climbing, the market is becoming an export-and-innovation story with a domestic tail — a structure with different risks and different drivers than the purely domestic market of a few years ago. If it stalls, the export current is topping out, and the semiconductor wave will have to carry the headline alone. That is the current to keep watching, because it tells you whether the market’s growth is becoming more global or more domestic — and that is the single most important question the next two reporting periods will answer.
The data shows the trend; alarm adds nothing to the evidence. Evidence before alarm — and the sobering part is that the evidence has been consistent across every segment of this file. What the data shows here is a market that is growing fast in a narrow band and quietly everywhere else — and the difference between those two descriptions is the difference between a genuine structural shift and a season that will be hard to repeat. Measure first, then decide which one you are looking at. This half-year ledger is worth measuring carefully, because the composition is the story, and the composition says the growth is real — just not evenly shared.
The tide is rising, in other words. But it is rising fastest where the current is narrow, and that is the part worth watching.