摸鱼小能手

$ANKER(00668.HK) dropped over six percentage points, even the star stock in the overseas consumer electronics sector couldn't avoid the pullback. As long as the shadow of tariffs persists, the valuation of the export chain won't go up. 😔

$LINGYI ITECH(01688.HK) and $HUAQIN(03296.HK) both saw slight declines, sentiment in the OEM chain is lukewarm; $SUNNY OPTICAL(02382.HK), however, edged up slightly. The story of optical upgrades counts as a rare growth point within the smartphone supply chain.

$Cerebras(CBRS.US) rose over eight percentage points, with the expansion signal released at the Paris AI Summit being directly bought by funds; $Nebius(NBIS.US) rose slightly, $Super Micro Computer(SMCI.US) moved sideways, $NetApp(NTAP.US) fell slightly🤔

All on the AI computing power chain, the new story of wafer-scale chips is being chased, while the old story of servers and storage isn't getting any additional funds.

The valuation divergence within the computing power chain has already widened significantly: $Cerebras(CBRS.US)' valuation has been rising continuously since its IPO, while $Super Micro Computer(SMCI.US) is still digesting the trust discount. Chasing the new story now is equivalent to betting that "the spillover of inference computing power demand to non-NVIDIA camps" will materialize in the next two quarters—do you think the speed of this spillover can support the current pricing?

$Vnet(VNET.US) and $GDS(GDS.US) both rose around eight percentage points. Chinese concept data centers are strengthening along with the demand for AI cloud and expectations of increased purchases of H200.

$Digitalocean(DOCN.US) rose over two points, as cloud service demand from small and medium-sized customers is also recovering. IDC directly corresponds to the implementation of AI computing power—rack occupancy rate and power supply determine revenue. These two Chinese concept companies hold data center resources in first-tier cities, offering greater flexibility than their overseas peers. The capital expenditure for this round of AI cloud is still in an acceleration phase, and the certainty on the data center side is relatively high among this batch.

$JOYSON ELEC(00699.HK) and $SANHUA(02050.HK) have both dropped 4-5% in this wave, with both auto parts stocks weakening together. The new energy vehicle (NEV) supply chain has recently lacked news of new volume expansion or orders. Parts stocks are pulling back along with the cautious sentiment in the overall vehicle market. The thermal management systems of $SANHUA(02050.HK) and the automotive electronics of $JOYSON ELEC(00699.HK) are both segments that follow downstream demand. There's no specific negative news; it's more like sector-wide waiting. Upward momentum will only come after overall vehicle sales data shows signs of recovery.

:dart: $Chagee(CHA.US), $MAO GEPING(01318.HK), $CHERY AUTO(09973.HK)

$Chagee(CHA.US) fell slightly, $MAO GEPING(01318.HK) dropped over four percentage points, and $CHERY AUTO(09973.HK) dipped a bit. China's new consumption sector fell in both Hong Kong and US markets today — the tea beverage, cosmetics, and vehicle manufacturing sectors all pulled back simultaneously.

Is this the Hong Kong stock new consumption sector entering a "post-IPO adjustment period," or are domestic consumption indicators marginally weakening? Catalysts for all three companies remain — $Chagee(CHA.US) delayed its US store opening in June, $MAO GEPING(01318.HK) is waiting for the pace of Japanese stock offline channels, and $CHERY AUTO(09973.HK) is yet to release its June sales data. The synchronized pullback is more likely due to rotation in Hong Kong's newly listed and post-IPO stocks, not a fundamental turning point.

The AI computing infrastructure theme was collectively sold off today.

$Vertiv(VRT.US) and $Hewlett Packard Enterprise(HPE.US) both fell over six percentage points. $Applied Digital(APLD.US) fell over four percentage points. $IREN(IREN.US) and $CIG(06166.HK) fell one to two percentage points, respectively.

Two signals of a loosening sector narrative: first, "unsustainable AI Capex" is starting to be written into research reports by mainstream sell-side analysts; second, news of OpenAI delaying its IPO has spread skepticism about the pace of AI capital expenditure from the primary market to the secondary market. High-beta names like $Vertiv(VRT.US) will take the hit first.

In terms of pricing, this round of correction has pulled $Vertiv(VRT.US) and $Hewlett Packard Enterprise(HPE.US) back from their previous highs to near their levels at the beginning of the year. Whether computing infrastructure has entered a valuation repair zone is one of the core questions for the next 2-3 months.

$WEICHAI POWER(02338.HK) rose slightly, $DONGYUE GROUP(00189.HK) dipped slightly, $TIANLI HOLDINGS(00117.HK) fell seven percentage points, and $CNBM(03323.HK) dropped nine percentage points, which was quite noticeable—the chemical and cyclical sectors are moving in a very divergent manner. $CNBM(03323.HK) fell deeply; the fundamentals of cement were already weak. When risk appetite rebounded, capital flowed from these old-economy cyclical stocks towards AI and semiconductors, and combined with its own profit pressure, the decline was amplified. In contrast, stocks like $WEICHAI POWER(02338.HK), which carry expectations related to heavy-duty trucks and data center generators, held up. Within the same cyclical sector, those touched by the AI narrative and purely traditional ones have moved in two different directions.

$Hewlett Packard Enterprise(HPE.US) and $Vertiv(VRT.US) both dropped over three percentage points, and $CIG(06166.HK) fell even deeper, by over seven percentage points😪

The entire chain of computing hardware—AI servers, power supply and cooling, optical modules—from US stocks to Hong Kong stocks is all in a state of profit-taking. No one doubts the capital expenditure story, but the debate over whether valuations are too high is getting louder. In times when the VIX rises, high-beta hardware is always among the first to be sold.

Several stocks in the Hong Kong stock market's computing power chain surged together: $ILUVATAR COREX(09903.HK) rose by over ten percent, with the dual positive catalysts of inclusion in the Stock Connect and the tech index boosting sentiment for domestic GPUs; $ASMPT(00522.HK) rose by nearly ten points, as its advanced packaging equipment secured repeat orders from major manufacturers; $SICC(02631.HK) also rose by over eight points, with its silicon carbide capacity utilization running at full tilt.

$BIREN TECH(06082.HK) was the only one in this group to decline, falling by over four points. Also a domestic GPU, with no news, its high beta led it to take the brunt of the pullback first 😪

Fastly basically didn't move, but Shopify fell three or four points, The Trade Desk and Cadence Design Systems both fell two or three points—the software batch's internal movements were a bit scattered today. E-commerce, advertising, and EDA each do their own business, but today they were all dragged down together, with only the small-cap edge CDN stock bucking the trend and holding steady. In this broad decline, is the sporadic resistance of Fastly due to independent catalysts, or is it just random volatility of a small-cap stock? It looks more like the latter—today, what's weighing on this batch is valuation compression under interest rate expectations, with high-valuation growth stocks like Shopify and The Trade Desk bearing the brunt. Fastly has a small market cap and isn't closely tied to the major themes, so it wasn't dragged in instead.

Cambridge Technology and Sanhua Intelligent Controls both fell by four or five percentage points. The two segments of small-cap Hong Kong stocks, optical modules and data center thermal management, pulled back together 😪. Both are segments with relatively high elasticity on the AI computing chain. When the broader market fluctuates and risk appetite shrinks, high beta small-cap stocks always take the first hit. The more they rose earlier, the harder they fall back.

Huaqin Technology rose 4% in this wave, as the AI terminal business on the ODM manufacturing side is being repriced;

Foxconn Interconnect Technology fell 3%, with no new updates on the connector business in the short term;

Sunny Optical fell slightly, with the optical sector waiting for mobile phone shipment data;

Changchun Changguang Chenxin remained largely flat😷

BigBear.ai crashed 31% today, and the earnings report immediately exposed its true form😭

This is the dilemma of AI application companies—easy to spin a narrative, hard to monetize. The valuation that was propped up by concepts in previous quarters was laid bare the moment it met the real quarterly numbers. Losses were deeper than expected, and revenue growth was disappointing. We were still discussing the support level yesterday, but there's no need to look for it today😰

$Venture Global(VG.US) is a rising star in U.S. LNG exports. European procurement was re-locked during the Hormuz crisis, and the long-term LNG contract value logic remains valid after the ceasefire—a two-week ceasefire does not alter the medium-term trend of energy diversification.

$EOS Energy Enterprises(EOSE.US) targets long-duration energy storage with its iron-air battery technology, supported by policy amid U.S. grid investment acceleration, but commercialization is still in early stages.$Lightbridge(LTBR.US) focuses on nuclear fuel technology, driven by demand for small modular reactor support.$Navitas Semiconductor(NVTS.US) works on GaN/SiC power chips, with downstream applications in EV charging and AI data centers. It overlaps with Wolfspeed in the market but Navitas is relatively more focused on GaN.

Stayed up late using the newly released Longbridge skill to screen several major tech stocks in Hong Kong and the US. Data like moving averages and volume were pulled directly by the skill, saving a lot of hassle compared to manual calculation. But after reviewing all the data, I'm feeling a bit disheartened 😮💨. The market is so weak; barely any stock meets my screening criteria except for $Arm(ARM.US). Most others are underperforming. It's really tough!

According to the real-time data checked with Longbridge skill, ARM's current price is 149.11. Although the moving averages show a bullish alignment, with positive slopes for both MA5 and MA20, and volume exploded 4.46 times in the last 30 days—even surging over 4 times the average volume on March 25th, pushing the price from 148 to 157—the overall market is too weak. It's hard to say if it can hold steady after pulling back from the highs. Even with its architecture relied upon by Apple, Qualcomm, and NVIDIA, and a deep moat, it can't withstand the poor macro environment. I dare not be blindly optimistic.

After screening with Longbridge skill, the only stocks worth tracking are NVIDIA and TSMC, but even those are just for watching, not for action. NVIDIA has been range-bound between 165-195 for 6 weeks and is still below MA20. Even if it breaks out on high volume, a weak market could make it a flash in the pan. TSMC is in a narrow range, waiting for tariff relief, but who knows how long that will take? At this stage, it's just a waiting game.

The remaining Meta, Microsoft, Tencent, and Alibaba are directly flagged by Longbridge skill as not meeting the criteria—either their moving averages are declining or they haven't seen volume spikes. They're untouchable for now. Especially Hong Kong stocks, the overall sentiment is freezing cold, with Tencent and Alibaba showing no signs of recovery. It feels like a waste of energy even to watch them. Is there any buddy out there like me, relying on Longbridge skill to monitor the market, now too afraid to make a move? How much longer do we have to endure this market!!!

On March 31st, the quarter ended, and the crypto market's sentiment index plummeted to 11, indicating "extreme fear." Although $Grayscale Bitcoin Mini Trust ETF(BTC.US) is still holding around $67,000, $Strategy(MSTR.US) (MicroStrategy) crashed 3.6% at yesterday's close, dropping to $121, and $Coinbase(COIN.US) (Coinbase) followed with a slow decline. BlackRock's Larry Fink has been very active recently, and institutions seem to be using the panic to shake out weak hands. I think this is actually a good opportunity to accumulate MSTR in batches. After all, as long as BTC's long-term thesis remains intact, MSTR, as a "Bitcoin shadow stock," is essentially a leveraged call option. Do you think it can rebound by tonight's close?