July 13 recap — short-term trading back in oversold rebound, but take profits when you see gains

These past few days, the child was born, and I didn’t post or comment. Today, my wife and child were discharged from the hospital, so I’ll briefly talk to everyone about the latest market situation.

[Taoguba]

1、Indexes, There’s not much to say. When, at a certain stage, an index clearly can’t move up, and in the short term it has multiple big drops, then theoretically the index has no real expectations. Overall, it’s highly likely to be a downtrend channel. No need to explain—if the index itself is fine, why would it repeatedly see so many big drops in such a short time? The converse is also true. If the overall trend at this stage is down or expectations are bottoming, then the index has no meaning for most retail investors, because most individual stocks are priced for downside. “Repairs,” “rallies,” and intraday recoveries are mostly selling points. You can’t possibly expect the index to reverse during this stage. Stepping back, even if the index rises for 3 straight days, most stocks only get a 1 to 1.5 day oversold rebound, and then they fall again. Even if the index rises for 3 straight days, your holdings might still be down for 3 straight days.

2、Sentiment & themes, Here “sentiment & themes” mainly refers to non-AI-tech areas such as pharma, robots, and commercial aerospace—these are quant-driven sentiment themes, so I don’t recommend everyone paying too much attention. Back when money from prop-trading firms (游资) guided the sentiment themes, there were still some things you could do—continuation plays, board-hunting, strong-goes-strong, weak-to-strong, “watch A then trade B,” and so on. After all, prop traders are still human; if they guide the sentiment, you follow along and you can still drink some soup.

But today, the prop traders who used to dominate are basically nowhere to be seen. For quant-based sentiment themes, it’s basically a one-day trip. They’re machines; they won’t proactively build some “consecutive-limit-up dragon” or “space dragon.” If you follow and chase price, you’ll basically end up deeply trapped. Quant sentiment has always been about harvesting market liquidity—especially when the market is chaotic and listless. Quant-driven sentiment themes most often do a quick rise for a day and then immediately get slammed (suppressed/“murdered”) on the next day.

For example, the pharma and robot themes from earlier: if you hesitated and didn’t exit, you got trapped. Then last Friday, commercial aerospace saw a large volume of limit-ups; today it was immediately slammed. Most of the stocks that hit limit-up last Friday didn’t even have any premium today—most had no premium at all.

Here’s a joke: commercial aerospace “fired a rocket” at the cost of a rocket just to get you guys to come in—and it was recovered successfully.

So, for any quant-driven sentiment theme lift that isn’t led by prop traders—personally, I generally don’t look at it, and I don’t recommend following it. Especially when the market is weak, quant-driven sentiment themes are usually selling points on the day they surge; hitting limit-up also means selling. Otherwise, if you hesitate the next day, it will smash you into deep water.

3、Technology, This round of tech has fallen terribly. Over this period, high-level tech has basically averaged a 30%补跌 (catch-up drop)—it’s truly brutal. The trigger was the news that meta出租 (rental) computing power was exposed. Even if a bunch of later interpretations and clarifications came out, tech didn’t recover. Then the catalyst should have been Apple saying it plans to raise prices: the reason given is that storage chips are rising in price too aggressively, and they can’t digest the costs. This is indeed a major negative. Don’t look at how the hardware side is starting a price-rise wave—it looks great—but for terminal manufacturers like Apple that need to purchase huge amounts of components, they either (1) keep the original price to preserve sales volume, but then their return on cost and profit could be squeezed, or (2) raise prices, but then sales will naturally drop. Big companies can still hold on; small ones might go straight bankrupt.

This leads to people thinking: these seemingly endless hardware price hikes make the top hardware makers earn a lot, but if terminals don’t buy it, can such price hikes keep going? For example, if Apple’s terminal products raise prices, theoretically sales should fall. Also, this weekend’s赛力周末 (赛力斯) performance is expected to亏损（pre-loss） due to the storage price hikes. In fact, for any terminal product, it ultimately has to follow the price hikes—or the terminals can’t digest the costs.

So don’t just look at storage financial reports being dazzling—for overseas players like Samsung, SK hynix, Micron, SanDisk, etc., and for domestic ones like 江波龙, 兆易创新, 德明利, etc. plus some upstream and midstream suppliers that can profit along with it. But the downstream, every terminal consumption end is losing money. In that scenario, the path of “only hardware makers profit, while terminal makers lose money” can’t be sustained. Unless a large number of consumers are willing to pay the bill, there can’t be a closed loop. It may even impact the next global AI tech expansion cycle. Therefore, this time the big drop in tech hardware—besides the need for补跌 on the ones that were already high—has the biggest negative being that terminal makers issue price-hike news.

Now let’s talk about performance. Many people have been thinking that in Q3 tech should run another performance-driven wave of “earnings lines.” I also previously thought Q3 tech performance could still run another wave. But the key node for this performance wave was when 江波龙 released a midterm results forecast with a同比 600-700倍 increase, and the next day the price action was far below expectations. Such a scale of revenue and half-year同比 increase is, by far, one of the most explosive earnings in A-share history. It could even be unprecedented. Since there wasn’t a limit-up the next day, and then it fell all the way afterward. Around the same time, 杭电股份, 兆易创新, 永鼎股份’s midterm performance forecasts with同比 also jumped sharply—and then all fell sharply, even seeing A-shares slaughter (A-kill). Even 香农芯创, whose midterm forecast同比 increased by over 20 times today, hit a limit-down.

So from a performance perspective, since 江波龙 released that performance expectation and the price action was below expectations, the whole “tech performance expectation” narrative basically died immediately. So at this point, don’t keep considering a performance line trade. Fundamentally, once earnings are released, it’s basically priced in. Earnings are just reflecting last year’s and current-period performance, but most tech stocks’ prices have already priced in expectations for the next three years. At this time, in combination with some negative news—such as terminal price hikes—the whole tech hardware side comes under pressure. Because if terminal price hikes don’t get customers to pay, that path can’t continue. Then they’ll be forced to cut prices, or at least the pace of price hikes will be greatly slowed. And this tech cycle has been driven by price-hike expectations. Conversely, the price-hike-related stocks recently are the most under pressure.

Lastly, I’ll talk about the overall tech trend. It’s clearly a “tide leaving” (退潮). In every cycle, there must be at least one补跌 “tide-leaving.” After the tide leaves, whether it can stabilize and then start another leg is another story. But our response approach should be to treat it like a tide-leaving environment. I’ve told everyone before: in a big-level tide-leaving, don’t subjectively fantasize that after falling a bit, it’ll reverse—that’s living in a dream. A big-level tide-leaving at least goes through “退一反一” and “退二反二.” And if the overall environment still isn’t good, there can also be “退三反三.” Also, every time you retreat on a big scale, it’s very likely to set new lows. That’s why I tell everyone: repairs are for selling—don’t hesitate—because most stocks will fall again after they “repair.”

This wave of tech is split into overseas computing power and domestic computing power, then upstream raw materials. The upstream raw materials are mainly oriented toward domestic computing power.

Overseas computing power is mainly optical communications. This wave of optical comms was the first to see the tide leaving. Most stocks went through A-kill. Even the big names—like 大哥大中际旭创—are nearly breaking down to new lows. And groups like 长飞, 中天, 亨通光纤 (the three giants) all got A-killed. After this, what’s left is only oversold rebound; I don’t look for another leg up for now.

Domestic computing power is mainly semiconductor chips—i.e., domestic substitution: advanced node production and mature node production expanding capacity domestically. From the recent price action, the relatively resilient leaders are 上海复旦? (not correct) 种芯国际, 华虹公司, 沐曦股份. The key areas are 晶圆制造 and GPU—mainly pure domestic substitution expansion directions. There’s also the “advanced packaging” angle fermented by韬定律, which is more short-term sentiment. But from a medium-term perspective, the tech main surge in April, May, and June was driven by price-hike expectations. At present, terminals temporarily can’t complete a price-hike closed loop, so price-hike stocks are under pressure. Add the certainty that global AI expansion is coming—once expansion lands, it will also push existing price-hike stocks toward price cuts or at least slower price increases. So if you still like tech, it’s suggested to focus on the domestic substitution capacity-expansion “must-have” segments.

And in the capacity-expansion must-have segments, the main areas are semiconductor equipment, wafer foundry (晶圆代工), core components, upstream supporting materials, packaging, compute infrastructure, etc. There are also sub-segments—everyone should map them to their own understanding. For example: In semiconductor front-end equipment: etching, thin-film deposition, ion implantation, polishing, DUV lithography machines, cleanrooms, etc. For core components: heaters, chucks, vacuum components, precision metal structural parts, RF power supplies, etc. For upstream materials: photoresists, special gases, developer liquids, polishing liquids, copper-clad laminates (覆铜板), indium phosphide (磷化铟), etc. For compute infrastructure: AIDC data center rooms, liquid-cooling supporting systems, etc. For packaging/testing: large packaging/testing houses, testing equipment, probe stations, etc. For wafer manufacturing: domestically, mainly two companies—one is中芯 (SMIC), one is华虹 (Hua Hong).

There may be some sub-segments I haven’t thought of; later, if I have time, I’ll organize a separate list for everyone.

In short, summary:

1、Tech is a big-level tide-leaving. It needs to go through “退一反一,” “退二反二,” and even “退三反三.” In such a tide-leaving, even if it ends and stabilizes, a large batch of stocks will directly lag behind or even disappear. Therefore, for most tech stocks right now, repairs are selling points. Don’t fantasize subjectively that your holdings can make new highs again—there’s no meaning.

2、Tech from April, May, and June went through an earnings-driven route, essentially also a price-hike-driven route. With price-hike headwinds stacking up right now—one is that terminal giants are announcing price hikes one after another; unless consumers actually pay, the hardware price hikes can’t complete a closed loop. Another is that once global AI expansion lands, it will push those price-hike stocks into price cuts. So the point where the price-hike trade is being speculated has effectively ended. Repairs are selling points, and only the stocks involved in expansion have expectations.

3、Since the price-hike side is under pressure, if tech has “more to come,” in August it should focus on advanced and mature-node mass purchases of equipment dominated by semiconductor equipment, with domestic completion of domestic substitution. It’s recommended to pay attention to this direction.

4、For quant-era sentiment themes, I don’t recommend you follow. Most of them are gone the next day. Especially when the overall environment is bad, on the day of the surge, most are selling points rather than buying points—let alone if you chase in the next day; one careless move and you’ll be trapped deep.

5、As for the index: all that exists is an oversold rebound. Don’t think about anything else. In this kind of decline, it can’t directly reverse. Tomorrow there may be a one-day “short-term gambling” opportunity. For the market right now, you can only buy on green days—never buy on red—and definitely don’t chase. Otherwise you’ll die horribly.

Thanks to friends for the encouragement and support, @思绪走了光8

Thanks for the donations and support from friends, @莉莉莉莉李 @忘记了呼吸 @爱笑的华仔 @背后灵 @惊蛰万物生 @思绪走了光8