#闪迪投资者日后股价大涨，长期目标待验证 How to put it $AXTI is a company investing in semiconductors, and these days chips and storage have basically all been rebounding after a deep drop, especially $SNDK SanDisk. After yesterday's investor day, SanDisk set its long-term goals, leading a major counterattack in storage, chips, and related stocks. After stopping the SanDisk grid yesterday, at the opening of the US stock market today, SanDisk dipped down sharply. I think this is a good opportunity to enter the semiconductor industry, so I took the chance to open a $AXTI grid.

#闪迪投资者日后股价大涨，长期目标待验证 After closing the $SNDK SanDisk position and pocketing the profits, I immediately opened a grid on $AXTI. With 70U principal, 10x leverage, range 62-93, current price around 77.45. After just a few minutes, it has already arbitraged 27 times. AXTI is a semiconductor equipment company, making wafer manufacturing equipment and measurement systems in the upstream segment. When storage manufacturers like SanDisk make money, equipment suppliers will inevitably be driven along. The position size is even smaller than Xiaomi’s, just testing the waters first. The direction hasn’t changed — still continuing to believe in semiconductors. SanDisk’s performance and investor day have turned the AI storage demand story into a bill. As the upstream equipment side, rotation is just a matter of time. With a smaller position, drawdowns are acceptable. The grid parameters are set wide enough, 62-93, to capture short-term volatility. Yesterday’s SanDisk trade was a swing operation, capturing the rise from 1391 to 1580. Today’s two grids are a different approach: small position, wide range, letting the grid run automatically, no directional judgment, just capturing volatility. Opening and closing positions must be decisive. The SanDisk profits are already secured; for the remaining trades, just let the grid run itself, no rush to add positions or stop losses. Wait for the signal of the next major semiconductor rally before acting.

#交易之声：你的经验值得被听到 $XIAOMI Xiaomi Grid is already open, with 160U principal, 10x leverage, range 2.5-3.8, current price 3.275, basically in the middle position. The earnings report will be revealed on August 18. Market Expectations Morgan Stanley maintains an overweight rating, expecting Q2 revenue to exceed 100 billion yuan, with recurring net profit around 6 billion. Xiaomi's smartphone shipments in Q2 reached 31.2 million units, 15% higher than expected, ASP hit a record high, and gross margin is expected to remain above 8%. Daiwa expects 106 billion yuan, and the moo platform shows a consensus market revenue expectation of 108.8 billion. Automobiles Are the Biggest Variable July SU7 deliveries reached 21,044 units, maintaining above 20,000 units for four consecutive months. Xiaomi Auto's total deliveries in July exceeded 30,000 units, with Q2 total deliveries over 90,000 units, and cumulative deliveries in the first half of the year about 180,000 units. However, there is disagreement in the market about the full-year target of 550,000 units. Morgan Stanley is relatively optimistic, while JPMorgan believes the target is "difficult to achieve," citing weak momentum in electric vehicles before new model launches. Daiwa lowered the target price to 32 yuan. My Judgment Morgan Stanley says smartphone shipments exceeded expectations by 15%, ASP hit a record high; Daiwa worries about subsidy reductions and macro weakness; JPMorgan thinks the electric vehicle target is hard to reach; Huaxing raised the target price to HKD 47. Who is right or wrong will be revealed next Tuesday.

$SNDK SanDisk's large position was just closed, and I opened a small position on $EDEN to test the waters. With 20U principal, 5x short, grid range 0.065-0.097, ran arbitrage 19 times in 17 minutes, earning 0.05U. EDEN surged strongly today, rising from 0.045 to 0.086, up more than 60%. After such a sharp rise, my judgment is that a pullback is highly likely. The grid short position is exactly stuck at this point—if the price falls back, the grid continuously arbitrages during the decline; if it continues to rise, the position is small, so losses won't be significant. The liquidation price is 0.21, more than double the current price, so basically no risk of liquidation, purely for practice. The advantage of playing altcoin grids with a small position is a stable mindset. When holding a large SanDisk position through a pullback, I can stay calm because I know what I'm doing; when testing the waters with a small position, I can be relaxed because I can afford the loss. These two states do not conflict; the key is having a clear position for each trade. This position is small and should be closed tonight; any profit of a few U is a gain. Nothing much to say, just testing the feel. #交易之声：你的经验值得被听到

#黄金维持高位，韩国央行重返市场 The Bank of Korea, which hadn't bought gold in 13 years, suddenly made a move. In Q2, the Bank of Korea purchased about $250 million worth of SPDR Gold ETF. The amount isn't actually that large; placed within $427.36 billion of foreign reserves, it barely causes a ripple. But what I'm really watching isn't this $250 million. It's why the Bank of Korea, which hasn't touched gold in 13 years, suddenly started buying now? The latest LBMA survey is also quite interesting: 16 analysts predict a median year-end gold price of $4500, an average annual price of $4604, with the highest forecast reaching $5100. Of course, predictions are just for reference. If analysts could accurately predict gold prices, they probably wouldn't be making money by writing reports. But the trend is worth paying attention to. Central banks buying gold, geopolitical risks, and expectations of rate cuts—these logics have not disappeared yet. What's even more noteworthy is $BTC. This year, gold $XAU has risen about 9%, while BTC has actually dropped about 11%. Facing the same inflation, geopolitical risks, and interest rate expectations, gold is treated as a safe-haven asset, but BTC still carries the label of "high volatility risk asset + tech stock." So I actually think the most interesting thing now isn't how much gold has risen, but that the market is re-pricing BTC. The logic for gold is still intact. The Bank of Korea buying gold again after 13 years just adds another brick to this logic. I won't chase gold. Waiting for a pullback. Only a truly comfortable position is worth getting in.

#EarningsObserver: AI Infrastructure Earnings Reports Take the Stage $SNDK SanDisk finally closed this trade. Entered at 1391, peaked at 1580, profit 52U, 66%, enough. Investor Day content was solid—BiCS10 bit density increased by 60%, gross margin target of 80% for fiscal years 2028-2030, 8 customers signed NBM long-term contracts locking in $93.9 billion. Citi calls 2500, Bernstein calls 3000, likely to be revised upward after this wave. Next, preparing to look for opportunities in gold and crude oil. $XAU Gold: Slightly Overheated in the Short Term This afternoon gold and silver suddenly plunged, spot gold dropped over $35 intraday to $4372, earlier approaching $4450. Market participants said Japan's Prime Minister supports recent central bank rate hikes, yen surged, dollar index returned to 100, gold got hit accordingly. Waiting for a pullback. The previously mentioned target near 4170 remains; will act when it reaches there. $CL Crude Oil: Geopolitics Supportive, but Inventory Is a Problem No breakthrough in US-Iran talks, Trump says US "100% controls" the Strait, Iran says the Strait remains closed. OPEC lowered demand forecasts for the fourth consecutive time, US crude inventories unexpectedly rose by 17.4 million barrels. Grid trading loves this kind of market, crude oil volatility range is large, news triggers a single big move. Waiting for SanDisk's funds to be ready, then see if crude offers a suitable entry point. No rush, the market has no shortage of opportunities.

#交易之声：你的经验值得被听到 I stopped the $SNDK SanDisk grid. Entered at 1391, peaked near 1580, profit 52U, 66% return, enough. What was said at Investor Day? The technical roadmap was clarified. BiCS10 QLC bit density improved by 60% over the previous generation, HBF was defined as an "important solution for the AI inference era," and the CEO clarified that shipment volume will be flexibly adjusted based on profitability. The financial model provided hard metrics. For fiscal years 2028-2030, non-GAAP gross margin is about 80%, operating margin close to 75%, with a commitment to 100% excess cash return to shareholders. Long-term agreements locked in performance. Eight customers signed NBM agreements covering about 50% of bit shipments for fiscal year 2027, with a total contract value of $93.9 billion. Reason for the surge After the meeting started, the stock price took off directly, rising 12%, intraday high 1579. The market was waiting for the commercialization timeline, technical details, and profit model; SanDisk delivered all, with numbers stronger than expected. Citi called 2500, Bernstein called 3000, and target prices may be adjusted upward after this. My grid order Stopped. The upper range of 1490 was broken, invalidating the grid logic. Profit 52U, 66% return. Next, waiting for a pullback, no rush. The market never lacks opportunities.

#Gold remains high, institutions still bullish at year-end LBMA year-end forecast is 4500, with a high of 5100 and a low of 3879, a difference of over a thousand dollars. The large divergence among analysts indicates no unified market answer, but the mainstream expectation leans upward. The 90-day correlation coefficient between $XAU Gold and $BTC changed from -0.9 to +0.7, shifting from inverse to positive correlation. Funds are starting to price both in the same basket—geopolitical risk, inflation expectations, real interest rates—all affecting both sides. However, after the CPI release, they diverged: gold broke through 4400, while BTC fell from 64400 to 63800. The market positions gold as a defensive safe-haven asset and BTC as a highly elastic risk asset; the window for them to rise and fall together has not truly opened yet. The core factors supporting gold remain unchanged: Middle East geopolitical risk, US inflation, Federal Reserve policy path, and central bank gold purchases. Among 16 respondents, 5 listed Iran as the primary concern. CICC continues to recommend overweighting gold, believing the two narratives that previously suppressed gold—global liquidity tightening and de-dollarization—are being disproven. I personally wait for gold to pull back to around 4170 to enter long positions; I hesitate to act below 4380. LBMA's big picture is institutional bullishness, but the exact timing and price to enter still need to be determined by the market itself. No rush. If 4380 holds, it may go straight up; if it breaks, 4170 is a better position. I will continue to update the next long-term plan and short-term ideas on the community platform.

#Chip stocks lead the rise, South Korean stocks rebound over 22% in ten days That whale with 18 consecutive wins in storage also thinks it's expensive. $SNDK took profit at 1390, earning 400,000, then bought back in batches between 1201-1401, with an average price of 1309. Now SanDisk is at 1347, he has an unrealized loss of 270,000 and is still waiting. Reports say he hasn't added more positions at this level—indicating even he thinks the current price isn't cheap enough. From June 25 until now, he has made 16 rounds of trades on $MU and SanDisk, profiting in 14 rounds, with a total gain of 5.63 million. If he doesn't move, I don't dare to be too aggressive either. My own grid order: opened at 1391, now at 1347, total profit +31U, grid profit 43U, unpaired loss -11U. The price has pulled back, but the grid is still running. Arbitraged 3758 times, profits are still ongoing. Investor Day starts tonight, but I don't plan to chase. The whale is watching at this level, indicating he thinks the cost-performance above 1350 is not high. I think similarly, the grid keeps running. Keep it up $OKB

#马斯克称AI将占SpaceX价值99% $SPCX Musk dropped a bold statement at the all-hands meeting: AI revenue is expected to surpass the company's other businesses combined by September, and in five years, AI will account for 99% of SpaceX's value. He did the math — by the end of next year, building 10 gigawatts of computing power, at $30-50 per watt, corresponds to annual revenue of $300 billion to $500 billion. SpaceX's AI business revenue in Q2 has already grown 213% quarter-over-quarter and 247% year-over-year, reaching $2.6 billion. Of the $7.8 billion total revenue in Q2, AI already accounts for nearly one-third. The last successful short squeeze happened because the short positions were too crowded — 34% of the float was shorted. Now the short positions have dropped to only 11%, so even if the stock price continues to rise, the number of shorts forced to cover has significantly decreased. The intensity of the previous "short squeeze stampede" is hard to replicate. A bigger issue is the continuous release of shares. On August 20, 319 million shares (about 7%) will be unlocked, about 700 million shares in September, and nearly 700 million shares in October. By the end of the year, the float will surge from 639 million shares to 5.33 billion shares. Last week's unlock didn't crash the market because the short positions were too crowded. This week, the shorts have mostly exited; can we really expect the same script to play out again? The real factor determining the trend after August 20 is not the shorts, but the insiders and early investors who can finally sell their shares.