#财报观察员：AI基建财报接力登场 $SNDK SanDisk grid floating profit has turned positive. Nearly 3,300 arbitrage trades were executed in 7 days, averaging 470 per day, with grid earnings of 37U, paired loss narrowed to 31U, total profit +5.92U. Previously lost over 30 points but held on hard, now starting to make money. $SPCX The unlock wave did not crash the stock price, which has risen back above the IPO price. But the next batch of 319 million shares unlocking on August 20 is already on the way, and there are 700 million shares in September. Shorts haven't fully exited; currently more than 250 million shares are still shorted. The fundamental direction hasn't changed, I continue holding the grid waiting for the August 13 investor day. SanDisk's fundamentals haven't collapsed, Citibank's 2500 target price remains, and the long-term storage logic is still intact. As long as the forced liquidation price isn't broken, I will hold steady. This week AI infrastructure earnings reports are intensive; data from optical modules, computing power cloud, and equipment manufacturers will verify whether AI capital expenditure is accelerating or slowing down. SanDisk grid closed at 1279 tonight, with an upper range of 1490 and lower range of 1219, enough room to run. Now grid earnings can cover the floating loss of the base position, making the holding mentality much steadier. Continue letting the grid run itself, waiting for what cards SanDisk can play on the 13th investor day.

#波动雷达：币种异动观察 $DOS really had a strong move today. It surged directly from 0.324 to 0.535, up 33 points, and is now pulling back near 0.52. The 24-hour trading volume is 127 million, which is quite substantial. For a new coin on its first day, this performance shows the sentiment is definitely in place. The background of this coin is actually quite interesting. DAPPOS is a Web3 AI operating system, with its core product called xBubble — you give it a simple prompt, and it automatically helps you generate documents, slides, websites, videos, and such. Essentially, it’s "doing Web3 tasks through chat," which lowers the entry barrier significantly. DOS is its ecosystem token. The funding background is not weak. In March 2024, it completed a $15.3 million Series A round led by Polychain, with a valuation of $300 million. In July 2023, there was also a $5 million seed round with investments from Sequoia China, IDG, and OKX Ventures. Altogether, it has raised over $20 million, and the shareholder list is quite strong. Why the surge today? On August 10, DOS was listed on several exchanges simultaneously — Gate launched spot trading first, Binance Alpha went live, and OKX also listed perpetual contracts. Bitget and BingX followed suit. Within one day, it was listed on five or six platforms, instantly unlocking liquidity. The premium effect typical of new coin launches naturally kicked in. OKX’s contract opened at 20:00 today, priced in USDT, with up to 20x leverage. Total supply is 1 billion tokens, with an initial circulation of 200 million. In the token allocation, the team and investors together hold 42.5%, all with a 12-month lock-up period. Short-term selling pressure is actually not large, with no major unlocks expected in the coming months. What exactly is driving this surge? The Web3 AI narrative has been quite hot recently, combined with listings on five or six exchanges in one day, liquidity premium and narrative premium stacked together, resulting in this market performance. From 0.324 to 0.535, a 65% increase before pulling back — this rise has basically priced in both the "listing expectations" and the "AI narrative" at once. What’s next depends on whether the product can truly be realized. No matter how strong the narrative is, it still needs actual users and revenue to support it. But at least today’s move was quite decisive, with good volume to match. My grid orders are still running. $SNDK on the SanDisk side averages over 500 arbitrage trades daily, and profits are slowly coming back. For a newly launched coin like DOS, volatility is indeed high, but I’m not very familiar with this project, so I’m just watching for now, not rushing to act. If there’s a pullback later, I might consider trying a small position. For now, I’ll let the grid run itself; no need to rush.

#黄金升破4300美元，资金在押降息还是避险？ $XAU Gold has been quite strong in this move. Before the non-farm payrolls, it traded between 3900 and 4100 for a full two months, unable to break up or down. Once the non-farm data came out showing a decrease of 23,000 jobs and cooling rate hike expectations, gold surged through 4100 with a strong bullish candle, now pushing above 4400, reaching a high of 4421 today, up nearly 8% in this rally. Honestly, the core logic for gold holding at this level is the easing of rate hike expectations. Both ADP and non-farm data were weak, and the market started repricing the likelihood of a September rate hike, with CME's no-rate-hike probability rising to 55%. The dollar weakened, U.S. Treasury yields turned down, and naturally, gold went up. But I’m a bit hesitant about the 4400 level. It jumped straight from 4100 to 4400 with almost no decent pullback, and the RSI is already above 73, entering the short-term overbought zone. Chasing in at this level has a poor risk-reward ratio. I think the short-term peak is likely near 4400, and a pullback is expected next, targeting around 4170. If gold can retrace to that level, digest the overbought signals, and confirm the validity of the 4100 breakout, I will enter a long position near 4170. After all, the long-term logic for gold hasn’t changed: geopolitical risks remain, and the Fed will eventually pivot, it’s just a matter of timing. During this period, $SNDK SanDisk’s grid has been running continuously with sufficient volatility, averaging over 500 arbitrage trades daily. If gold really gives a pullback opportunity to 4170, I will shift part of my position there. It’s not that I’m bearish on gold, I just want to find a better entry point. No rush to chase, waiting for it to come down on its own.

#本周三CPI公布，9月加息定价会改写吗？ The CPI will be released on Wednesday night (August 12). The market expects the overall CPI annual rate to drop from 3.5% to 3.4%, and the core CPI annual rate to decrease from 2.6% to 2.5%. The probability of a rate hike in September has already fallen significantly. After the non-farm payrolls data came out, CME FedWatch shows the probability of maintaining the current rate in September has risen to 55.6%, while the probability of a rate hike dropped to 44.4%. Polymarket leans more towards no rate hike, with a 63% chance of holding steady, and Kalshi shows 65%. A week ago, CME's rate hike expectation was still 67%, but it dropped immediately after the non-farm data was released. Interestingly, although market expectations are changing, institutional views are still divided. Bank of America insists on a September rate hike, arguing that the Fed will focus more on inflation data rather than employment data. Citibank says if CPI softens for a second consecutive month, a September rate hike is basically off the table. The Cleveland Fed's Nowcasting model shows July core CPI is expected to rise 0.21% month-over-month and 2.52% year-over-year, rebounding compared to June's flat month-over-month data. The subtlety here is that 0.21% is not high, but it is a rebound from June's 0%. Hawks will say "inflation is accelerating again," while doves will say "it's still below the average level of the past 12 months." So if Wednesday's data is below expectations, the probability of a September rate hike may continue to decline, and risk assets might get a short-term breather. If the data exceeds expectations, rate hike expectations could bounce back, and the market may need to reprice again. My $SNDK SanDisk grid is still running; it dropped to 1197 pre-market but bounced back. The CPI variable is a source of volatility for me. Good data or bad data is fine, as long as the price oscillates within the range, the grid can keep capturing profits. Tomorrow night, we'll see how the market reacts to the data. $BTC

#存储股抛压缓和，AI内存牛市还稳吗？ $SNDK SanDisk is stronger in this wave than I expected; it bounced at 1191 without reaching 1160. Now it’s back to 1255, the grid is running again, and the floating loss has shrunk from -30U to -8U. Most likely, it will fluctuate between 1200-1300 next, then wait for the Investor Day on August 13. This range is perfect for the grid: sellers at the upper edge, buyers at the lower edge, and free movement in between. My grid has recovered, averaging over 500 arbitrage trades daily. As long as the price stays within the range, profits will gradually return. Investor Day is the first real catalyst after this SanDisk pullback. The HBF roadmap and BiCS10 commercialization schedule are the core variables that will determine if SanDisk can turn this oversold rebound into a trend reversal. If the market sees a clear technical path and commercialization timeline, SanDisk’s valuation logic might be repriced. At this stage, watch more and act less. Let the grid run itself and wait to see what cards SanDisk plays on August 13. Next week might be when SanDisk starts telling its story again.

#闪迪8月13日投资者日临近，财报分歧待解 $SNDK SanDisk has already dropped to 1197 pre-market, with a low touching 1191. 1160 is the support level tonight; if it can hold until the August 13 investor day, it could surge upward. From the candlestick chart, SanDisk has fallen from 1391, but every time it reaches around 1190-1200, there is support. Last time, the low was 1186 and it bounced back. If it can hold above 1160 this time, it would indeed signal short-term stabilization. Technical indicators also support this view. MACD and moving averages are starting to converge, indicating weakening downward momentum and the market is waiting for new catalysts. RSI is in the forties, neither oversold nor overbought. Coupled with potential positive news from next week's investor day, bears may lack the motivation to continue pushing the price down at this level. However, the key is still the plan on August 13. The HBF roadmap and BiCS10 commercialization schedule are the core variables determining whether SanDisk can shift from an "oversold rebound" to a "trend reversal." If a clear technical roadmap and commercialization path are provided that day, the market may reprice. If it remains vague, SanDisk might have to consolidate around 1200 for a while longer. My grid orders are still running; the price has now returned above 1219, so the strategy should have resumed. The liquidation price is 930, which is still over 260 points away, so no need to panic for now. If 1160 can hold, the grid can continue to arbitrage.

#闪迪8月13日投资者日临近，财报分歧待解 $SNDK SanDisk will hold its Investor Day on August 13. What will be discussed at this Investor Day? According to the official schedule, it starts at 9 AM Eastern Time on Thursday, August 13. The core topics include: HBF product roadmap, commercialization timeline, BiCS10 3D NAND technology, SSD capacity expansion plans, and long-term supply contracts. What is HBF? HBF is a new storage technology jointly launched by SanDisk and SK Hynix, positioned between HBM and traditional SSDs. HBM is too expensive and has small capacity, SSDs are too slow with insufficient bandwidth, and HBF fits right in the middle — using NAND flash stacking, 8 or 16 layers stacked, with a maximum of 512GB per module and bandwidth from 0.4 to 3.0TB/s. It uses a UCIe interface, allowing direct connection to GPU and CPU. Google and Tenstorrent have already joined the ecosystem. The goal is to solve the "storage wall" problem in AI inference — as models grow larger, no matter how powerful the GPU is, it’s useless if data can’t be fed in. What is BiCS10? BiCS10 is the tenth-generation 3D NAND jointly developed by SanDisk and Kioxia, with 332 layers stacked, QLC version, bit density increased by about 59% compared to the previous generation. Samples will be delivered in the second half of the year. It uses CBA wafer bonding technology, optimizing logic circuits and storage arrays separately. It features high density and low power consumption, designed specifically for AI data lakes and RAG knowledge bases. Why is it important? SanDisk’s current stock price is 1214, down 48% from the June high of 2354, with a price-to-earnings ratio of about 6 times. The earnings report showed 8.97 billion, exceeding expectations, but guidance was lowered by 250 million, causing the market to turn negative immediately. Citi’s target price is 2500, and the average analyst target price is 2220. If the Investor Day provides a clear enough HBF roadmap and commercialization timeline, the market may reprice the stock. After all, the dual technologies of HBF and BiCS10 are not minor upgrades. Recent insider selling by executives is indeed unsettling, but given the high price level, selling is understandable. The key is whether HBF can transform from a "technology story" into a "revenue story." On August 13, if the roadmap is clear enough, the market may assign a premium again. If it remains vague, SanDisk may have to linger around 1200 for a while longer.

#财报观察员：空头回补成焦点，SpaceX后续怎么看？ $SPCX still has several rounds of unlocking ahead, and selling pressure will indeed come wave after wave. 📅 Unlocking timeline After the first batch of 911.5 million shares unlocks on August 6, the following are expected: · August 20: about 319 million shares · September: approximately 700 million shares · October: close to 700 million shares The entire unlocking process uses a nine-stage batch mechanism and will continue until 2027. Elon Musk and some specific shareholders have an extended lock-up period that ends next June. Shorts are not out yet Currently, more than 250 million SPCX shares are still shorted, and short positions remain. If insiders start mass selling after unlocking, shorts will have more ammunition; if selling pressure is not as big as expected, shorts may continue to be forced out. The battle between both sides is not over. My view The market withstood the August 6 unlocking wave; the stock price didn’t collapse but actually rose. However, that doesn’t mean the following rounds will hold up—each batch of unlocking is a new supply shock. The 700 million shares in September and 700 million shares in October combined exceed the first batch of 911.5 million shares. Chips are continuously being released, which is the real test. SPCX’s current level can be seen as cheap or expensive depending on your calculation cycle. There are still several rounds of unlocking ahead, so I won’t heavily bet at this position. I’ll watch first and wait for the chip structure to stabilize. No rush to jump in, no rush to run.

#存储股抛压缓和，AI内存牛市还稳吗？ $SNDK SanDisk grid ran for 5 days and 10 hours, arbitraging 2,929 times, averaging 540 times per day. Grid profit is 33.45U, floating loss narrowed from over 30% to 25%, total profit -20.32U. A simple calculation: average daily grid profit is 6.17U, based on 80U principal, daily yield is about 7.7%. Previously I said about 5%, which was conservative. The grid is making profits, but the base position is losing. However, as long as SanDisk continues to fluctuate, the grid can keep arbitraging. Floating loss narrowing from over 30% to 25% indicates the base position is slowly recovering. If it rebounds above 1300, total profit will most likely turn positive. This is the advantage of grid orders — no need to predict direction, only need volatility. SanDisk has been jumping up and down recently, which is exactly the rhythm the grid loves. I don’t need to guess if it will rise to 1400 or worry if it will fall to 1100. As long as the price stays within the range, the grid can automatically arbitrage. It paused below 1219, and resumed when it rose back. This logic is quite simple: no direction judgment, capture all volatility. Not greedy, not panicking, just let the grid run slowly.

#Spot ETF capital inflow, can BTC and ETH take over? Saylor is at it again. Last night he posted a picture with just two words: "Doing ₿usiness." Those familiar with him should know this basically means: getting ready to buy. He’s very experienced with this kind of pre-announcement pattern. On August 2nd, he posted "Bitcoin Drive engaged," and the market was guessing the Strategy would restart buying coins. At that time, the company actually didn’t buy from July 27 to August 2, but sold 1,638 BTC instead. However, that sale was to pay preferred stock dividends and repurchase discounted shares, not an active liquidation. This time, the "Doing Business" message came at an interesting moment—posted on August 9. Based on his usual rhythm, submitting documents to the SEC on Monday to disclose new purchases is routine. So the earliest news could come tomorrow (Monday). The current conditions are different from before. Strategy holds about $4 billion in cash. The funds from the previous coin sales haven’t been fully spent yet. Currently holding 842,138 BTC with an average cost of $75,419, floating loss is about $10.8 billion. The price is now around $65,000, nearly $10,000 below his cost. The last time he bought heavily was in June, averaging $63,024 for 1,587 BTC. The current $65,000 price is close to that last purchase price. If he really restarts buying this week, it means he still thinks this range is worth adding to his position. Market reaction is quite direct. $BTC broke above $65,000 yesterday and stayed around that level after the news. Saylor’s hint definitely fueled market sentiment, with whale activity and institutional optimism heating up simultaneously. But one thing to keep in mind: what he posted is just a hint, not an announcement. If he really bought, the SEC filing tomorrow will confirm it; if not, it’s just another "trailer." But judging by his past style, posting this kind of picture is rarely an empty shot.