#AIMemoryBullTest

Strong earnings, weak stock reaction—what happened? 📉 The latest reports from Western Digital ($WDC) and SanDisk ($SNDK) delivered outstanding historical results, yet both stocks sold off sharply. The reason? Markets are no longer rewarding strong performance alone—they expect companies to keep outperforming already sky-high expectations. 📊 Earnings Snapshot • Western Digital: Q4 revenue reached $3.75B (+44% YoY), with adjusted EPS of $3.56 (+109% YoY). • SanDisk: Q4 revenue climbed to $8.97B (+372% YoY), while data center revenue skyrocketed nearly 1,300% YoY. 📉 Why the Sell-Off? After massive rallies of roughly 200% and 469% over the past year, investors wanted even stronger forward guidance. • SanDisk's next-quarter revenue outlook came in below market expectations. • Western Digital's guidance also failed to match the bullish outlook reflected in competitor forecasts. • Both companies projected slightly softer gross margins, fueling concerns that profitability may have peaked. 🌍 Ripple Effect Across the Sector The disappointment spread quickly, weighing on storage and semiconductor names worldwide. Stocks including SK Hynix, Samsung Electronics, Kioxia, Micron, and Seagate all faced increased selling pressure. ⚖️ What's Next? The market is divided. Some investors see this pullback as a buying opportunity, pointing to tight inventories, limited production capacity, and long-term AI demand. Others believe the sector may be entering a period of slower growth after an exceptional run. The key question now is whether this correction is simply a reset in expectations—or the beginning of a broader valuation adjustment. NFA. Always manage your risk. #AIMemoryBullTest #FedHawksVsWeakJobs #SpaceXUnlockRebound

AMD Earnings Review → SanDisk $SNDK Tonight's Prediction AMD beats expectations + strong guidance → falls 9% after hours Reason: Gross margin below expectations, good news already priced in Same logic for SanDisk tonight, bearish SanDisk earnings release tonight (early morning August 6 Beijing time) Expectations: Revenue $8.3B / EPS $34.24 Price has risen 32% from the low of 1123, expectations fully priced in Price going up means short. Above 1420, short short short, live in the palace #AMD财报超预期，增长已被透支？ $AMD $SNDK $SNDK

AMD Earnings Review → SanDisk $SNDK Tonight's Prediction AMD beats expectations + strong guidance → falls 9% after hours Reason: Gross margin below expectations, good news already priced in Same logic for SanDisk tonight, bearish SanDisk earnings release tonight (early morning August 6 Beijing time) Expectations: Revenue $8.3B / EPS $34.24 Price has risen 32% from the low of 1123, expectations fully priced in Price going up means short. Above 1420, short short short, live in the palace #AMD财报超预期，增长已被透支？ $AMD $SNDK

Goldman TMT on tech weakness SanDisk & Western Digital: Both memory names traded lower after hours despite strong reported quarters, as the issue was not the current print but elevated buy-side expectations going into results. SanDisk beat on revenue, gross margin and EPS, but 3Q revenue guidance was only slightly below the Street while EPS was broadly in line, triggering some profit-taking after a big prior move; the more supportive takeaway is that SanDisk has locked in multi-year customer supply agreements covering over 50% of FY27 and 65% of FY28 planned bit production at floor pricing, supporting NAND pricing visibility and buybacks. Western Digital also beat on revenue, margin and EPS and guided 3Q revenue/margins above consensus, but the stock fell because the guide did not reset the bar enough given high HDD pricing and margin expectations, with some concern around softer exabyte growth as the company manages its 40TB ePMR ramp and HAMR qualification; overall, the read-through is a “beat but not enough” night for NAND/HDD, with fundamentals still supportive but expectations very high.

$SNDK about $1289, -4.5% after-hours. Sandisk printed a much bigger quarter than a normal storage beat, and the stock still faded. Q4 revenue was $8.97B, up 51% q/q and 372% y/y. Non-GAAP EPS was $39.25. Non-GAAP gross margin reached 84.6%, and Datacenter revenue climbed to $2.98B from $1.47B last quarter. Q1 guide is $10.3B-$10.8B with $44-$46 non-GAAP EPS. Sandisk also said it signed five more NBM agreements since April, taking the total to ten, and added $14B to its buyback authorization. The after-hours read is that the setup had already moved to a guide-and-duration test. If datacenter mix, pricing, and those NBM commitments keep lifting FY27 EPS, this fade can reverse. If the market decides Q1 only meets a stretched bar, the storage de-rating can keep running. source: company release / SEC filing

*WESTERN DIGITAL Q4 REVENUE USD 3,747 MILLION VS. IBES ESTIMATE USD 3,692 MILLION *WESTERN DIGITAL Q4 ADJUSTED EPS USD 3.56 VS. IBES ESTIMATE USD 3.3 $WDC

A financial health check for the storage sector, focusing only on the numbers. SanDisk SNDK reported revenue of $8.96 billion this quarter, clearly exceeding the expected $8.39 billion; however, its guidance for the next quarter is set at $10.3–10.8 billion, just at the lower edge of the market expectation of $10.8 billion — causing its stock to drop over 3% in after-hours trading. On the same night, Western Digital WDC instead issued a strong guidance with a year-over-year increase of +42% to +49%. In the same sector, one is dragged down by guidance while the other is lifted by it, indicating that the divergence in this storage cycle has shifted from "market boom or bust" to "who holds stronger pricing power." Data doesn't play games with you; even in a super cycle, there are winners and losers. $BTC #ADPCoolsFedSplit #EarningsRealityCheck #SpaceXBeatEstimates

Samsung's manufacturing scale has long been one of its greatest competitive strengths. While the AI memory race is drawing intense attention, production capacity and technological execution may ultimately prove just as important as current market sentiment. Many analysts continue to focus on global supply shortages and regional demand imbalances, but history has shown that companies capable of rapidly scaling advanced manufacturing can reshape industry expectations faster than forecasts anticipate. Recent discussion has also centered on Samsung's reported zHBM technology, which is claimed to deliver significant improvements over current HBM generations—including substantially higher performance, greater memory density, improved energy efficiency, and lower thermal resistance. If these capabilities are successfully commercialized at scale, they could strengthen Samsung's position in the high-bandwidth memory market. The semiconductor industry has always been cyclical. While prices fluctuate with supply and demand, long-term leadership is often determined by innovation, manufacturing expertise, and the ability to bring next-generation technology into mass production efficiently. As always, reported technical claims should be evaluated carefully until they are independently verified and reflected in commercial products. Rehan_X Facts, Trends & Insights #AIMemoryBullTest #FedHawksVsWeakJobs #SpaceXUnlockRebound

I am Cige. The earnings season for the storage sector has sent a clear signal: performance can shatter the ceiling, but stock prices still fall. Explosive earnings are just the entry ticket; guidance is the pricing anchor. SanDisk's Q4 revenue was $8.97 billion, a year-over-year surge of 372%, far exceeding the expected $8.39 billion; adjusted EPS was $39.25, and gross margin hit a record high of 84.6%. Western Digital's revenue was $3.75 billion, up 44% year-over-year, also beating expectations. Both companies delivered impeccable results, yet SanDisk fell 7% after hours, and Western Digital dropped 11%. The core reason is only one: the guidance was not impressive enough. SanDisk's next quarter revenue outlook is $10.3 billion to $10.8 billion, with a midpoint of $10.55 billion, below FactSet's expectation of $11.148 billion. Western Digital also faced disappointment for being "not impressive enough." Citi lowered SanDisk's target price from $2500 to $2100. The market wants not just "good," but "better than expected." When expectations are already at the ceiling, any number less than "perfect" will be punished. Three forces crushing the sector are fermenting simultaneously. The sell-off of SanDisk and Western Digital quickly spread through the entire storage chain. Kioxia and SK Hynix plunged over 10%, Samsung Electronics dropped over 6%. The KOSPI index's decline widened to 5%, SK Hynix fell over 9%, Samsung Electronics dropped over 6%. Daishin Securities clearly pointed out that SanDisk's below-expectation earnings guidance weakened market confidence in the storage chip industry, and the semiconductor sector's sharp correction was the main reason for the KOSPI's decline that day. NVIDIA is evaluating reducing Rubin Ultra's HBM configuration from HBM4e 12Hi down to 8Hi or other options. #AIMemoryBullTest #FedHawksVsWeakJobs #SpaceXUnlockRebound

Here is why AI compute and consumer hardware are about to get significantly more expensive, despite what most people think. Common belief says hardware and AI token prices will decline as computing efficiency improves and memory supply increases. But that overlooks the structural reality of the supply chain. At a recent SpaceX earnings call, @elonmusk pointed out that memory supply grows by roughly 20% a year, while demand is growing by 200%. Many expected cheap memory supply from China's CXMT to ease the pressure, but Apple recently received a quote from CXMT that was actually higher than Samsung's. Having talked directly with semiconductor fab engineers, this is entirely expected. Demand is so far ahead of supply that CXMT has no incentive to price aggressively. This gap widened massively after the agent boom earlier this year. Agents consume significantly more tokens by default, causing global inference usage to skyrocket. Better performance brought in a wave of new users, as shown by Codex rapidly crossing 10M users. Meanwhile, compute efficiency gains have been very small. AI labs chose to increase model sizes to raise intelligence, consuming at least double the compute and accelerating demand further. On the supply side, hardware bottlenecks take years to fix. SK Hynix's Yongin cluster only starts its first fab in 2027, with full operations planned for 2033. Even with massive construction starting today, it takes about 5 years for actual supply to land on the market. This is an unsolvable short-term bottleneck. AI inference costs will inevitably keep rising, and consumer hardware will follow. This is why I keep telling everyone to buy the personal hardware now.