NVIDIA and Intel are both raising money for AI, but their fundraising methods are completely different. NVIDIA doesn't put up its own money nor dilute equity; it purely leverages its platform capabilities and industry position to mobilize Wall Street's funds. Intel, on the other hand, directly issues new shares, reportedly increasing from 15 billion to 20 billion, with subscription demand exceeding 100 billion. These two paths laid out reflect the completely different positions of the two companies in the AI industry chain. NVIDIA can create a financing platform because GPU demand far exceeds supply. Customers want to buy computing power but lack funds to build data centers; Wall Street wants to invest in AI but lacks project entry points. NVIDIA sits right in the middle, connecting these two ends, raising customers' capital expenditures without spending a dime itself, while continuing to firmly control GPU pricing and allocation rights. Intel doesn't have this condition. It can only sell stock to raise money for capital expenditures and AI chip R&D. Subscription demand exceeding 100 billion shows the market is willing to provide funds, but the cost is equity dilution. Financing capability is becoming a key variable in the AI race. Whoever can get more money at a lower cost can run longer. NVIDIA's model is more sophisticated and non-dilutive but depends on continued external capital cooperation; Intel's issuance can quickly replenish funds but at the cost of equity. Neither model is absolutely good or bad, but in this money-burning race of AI infrastructure, financing ability itself is a moat. Winners on this track must not only be good at making chips but also at raising money. #AI基建融资升温，英伟达英特尔路径分化 $SNDK $NVDA $INTC

I'm honestly speechless—are the altcoins collectively going on a kill-off? Yesterday, I was watching $BEAT T fall from 3.4 to 2.6, but today it dropped straight to 1.07—40 points, and the day was gone $BICO also crashed, dropping from 0.07 to 0.038, halved $LAB Even more frustrating, my Martingal multi-single, which was running fine, today it was a tie for me This wave of altcoins is really fierce The CPI hasn't been released yet, and the market hasn't crashed; the altcoins collapsed first, and it was a completely unstoppable crash This indicates that funds are accelerating the withdrawal of small coins, and they have lost patience even for CPI Right now, the market is short-handed with two words: cash is shrinking, liquidity is shrinking, and capital would rather wait for CPI to land before moving than spend the night in small coins. The vulnerability of altcoins is fully exposed; when they fall, there's no bottom, not even a decent rebound. My own LAB grid was originally planned for long-distance running. But this afternoon, a waterfall was directly restored to a forced draw. This round can only be said to be a bit overconfident. When I used to trade Lab Strategy long orders, I really overestimated myself, which led to this huge drawdown. Honestly, I felt very upset because I lost a few hundred yuan. I still need to be cautious. From now on, I'll just play a small game. $SPCX is also the source of my confidence. I earned quite a bit from long orders, and now I pay it all back 🥹. I must be extra cautious in the future. #波动雷达: Monitor currency fluctuations

The data hasn't been released yet, but the market is already starting to stir. CME says the probability of no rate hike is 55.6%, while Polymarket and Kalshi push it above 63%. For the same event, there's nearly a 10 percentage point difference. Both markets are betting, but the direction and intensity of their bets differ. This divergence itself is a signal; the market lacks consensus, and Wednesday's CPI will force one side to concede. Expectations after the non-farm payrolls have already been priced in; BTC has been flat around 65000 for this reason—the good news is already baked in. Now the CPI expectation is for continued cooling, with overall at 3.4% and core at 2.5%. If it meets expectations, the market reaction won't be significant because the rise has already happened. But if CPI rebounds beyond expectations, the reversal in expectations will catch the market off guard. I have three possibilities in mind: · CPI continues to decline: no rate hike expectation confirmed, BTC may rally to break through 66000, gold slightly pulls back. · CPI meets expectations but no surprise: market reaction is muted, BTC continues to hover around 65000, waiting for the next catalyst. This is the most likely scenario. · CPI rebounds beyond expectations: rate cut expectations reverse, BTC retests 64000 or even lower, gold and energy benefit, US stocks come under pressure. Low probability events often cause the biggest volatility. The difference in expectations is the anchor for pricing. $SNDK $XAU $BTC #本周三CPI公布，9月加息定价会改写吗？

The earnings season for AI infrastructure has arrived, and tonight is a critical juncture. Palantir set the example two weeks ago, with revenue growth of 93% and a 12% stock increase. SpaceX reported revenue of 7.8 billion, but expenses reached 18.37 billion, a surge of 550%, with cash burn far outpacing revenue growth. Now, everyone is watching to see if AI can turn a profit. This week's earnings reports are tightly scheduled. Lumentum and CoreWeave will report after the market closes tonight, Coherent tomorrow after the close, Applied Materials the day after, and Cisco is also on the watchlist. All five segments are covered: AI cloud is the computing brain, optical communications are the data veins, equipment is the chip factory, and networks are the connecting skeleton. The pressure varies across segments. Optical communications and AI cloud have the highest order visibility, directly tied to the expansion of computing clusters, with guidance likely to be strong. The equipment sector faces more pressure, with capital expenditure slowing and marginal growth declining. Enterprise network AI has the weakest pull and depends on whether traditional demand can recover. The rules for this earnings season have been set: the current quarter's numbers are just the entry ticket; guidance sets the valuation. The market doesn't care how much you earned in the past three months; it only cares about how much you can earn going forward. Strong guidance gets the buy-in; weak guidance renders good results useless. Optical communications and AI cloud are worth watching closely, as their order visibility is high and they are most likely to exceed expectations. The equipment sector should be observed cautiously, as the capital expenditure risks have not yet been fully cleared. $SNDK $SPCX $SKHYNIX #财报观察员：AI基建财报接力登场

The Hormuz agreement is stuck again, and oil prices have directly soared. WTI crude oil $CL has risen to $80.98, and Brent crude oil $BZ has also risen in sync. Iran has laid out its conditions: lifting the US maritime blockade, US military withdrawal from around Iran, compensation for conflict damages, and unfreezing Iranian assets. Each of these conditions is tougher than the last, making the likelihood of reaching an agreement in the short term quite low. Oman is mediating safe passage negotiations, but full reopening is still some distance away. Geopolitical risks have been heating up all day. The Houthi forces bombed Saudi Aramco's refinery again, and the overall tension in the Middle East shows no signs of easing, keeping oil prices uncertain in their pricing. So why are BTC and ETH falling? The market is recalculating; sustained oil price increases mean rising inflation expectations, which in turn mean the Federal Reserve will find it harder to cut interest rates, and might even raise them again. BTC has dropped to around 64,300, ETH to 1,878, giving back previous gains. Mainstream coins are in a rather awkward position right now. Geopolitical risks push up oil prices → oil prices push up inflation expectations → inflation expectations delay rate cuts → risk assets come under pressure. This transmission chain is currently active, and BTC and ETH are being pressured by macroeconomic logic in the short term. Don't rush to bottom-fish now; geopolitical risks and oil price trends are still unclear, and CPI data is about to be released, so uncertainty is too high. Wait for these two variables to settle before making moves; hold your hand, and act only when the direction is clear. $BTC $ETH #霍尔木兹协议未落地，油价风险再升温？

#闪迪8月13日投资者日临近，财报分歧待解 SanDisk $SNDK tonight surged straight up to 1277, nearly a 100-point increase. The volume-backed rebound at this level indicates that around 1200, there are buyers willing to take positions, and market sentiment is gradually recovering from pessimism. The core catalyst is the Investor Day on August 13. The biggest issue with the earnings report is the weak guidance; the market has been worried whether this is due to conservative management or if AI storage demand is truly cooling down. The execution details of the 14 billion buyback plan have also not been clarified. Investor Day is a perfect window; this rally tonight is likely funds positioning early, betting on Investor Day delivering better-than-expected content. The earnings report itself is not bad, with revenue of 8.97 billion, gross margin of 84.6%, and EPS of 39.25; the numbers alone show no flaws. After nearly a week of digesting the weak guidance, the negative impact is gradually clearing. The market is starting to reassess this earnings report, which also explains why there are buyers around 1200. How far this rebound can go depends on two signals: whether 1200 can hold and whether Investor Day delivers better-than-expected content. Before these two signals come out, treat this as a rebound and wait for confirmation before adjusting expectations. $SKHYNIX $SPCX

SPCX hit 134.9 today, just a breath away from the IPO pricing of 135. Before the lock-up expiration, the market was shouting "trillions in sell pressure will crash the stock." But on the day of the lock-up expiration, it rose 6%, continued to rally the next day, gaining 23% in two days, adding over 327 billion in market value. Short positions stand at $24.6 billion, even higher than Tesla. When the stock price rebounds, it directly forces them to cover their positions, and the covering itself pushes the price even higher. Why didn’t the lock-up expiration cause a crash? Because after the earnings report, SPCX had already dropped to a historic low of 104.85, and those who wanted to exit have done so. The lock-up expiration actually signaled that the bad news was fully priced in. Currently, over 250 million shares are still shorted, accounting for 16% of tradable shares. Shorts remain, but the capital structure is increasingly unfavorable to them. The next variable is the second batch of lock-up expiration on August 20, just 10 days away. If 135 holds, there is short-term upside; if it fails to break through, a pullback may occur. Regarding capital expenditure, revenue was 7.8 billion, AI losses narrowed, but capital expenditure was 18.37 billion, a 550% year-over-year increase. Citi set a target price of 220, but to sustain that valuation, the burn rate must slow down. SanDisk $SNDK is still hovering around 1200, and $BICO dropped 49% in one day. SPCX is rising because it has a "new story": lock-up expiration bad news fully priced in plus short covering. Wait for the CPI data and the August 20 lock-up expiration to unfold; then the story will have its conclusion. $SPCX #财报观察员：空头回补成焦点，SpaceX后续怎么看？

After the non-farm payroll data came out last Friday, I had only one feeling: money is looking for a place to stay, but it hasn't found its way to BTC yet. Gold was the first to react to this non-farm data. It gained more than 7% in a week, steadily standing above 4300. The logic itself is not complicated: poor employment data, falling probability of rate hikes, the dollar softens accordingly, and gold gets pushed up. But the problem is, with the same data, gold has already fully priced in the "rate cut expectations" benefit, while BTC and ETH are still dragging their feet. ETFs have actually been providing support all along, with a net inflow of $865 million last week, nearly $700 million of which was from BlackRock itself. Ethereum has seen net inflows for five consecutive weeks without interruption. Money is indeed flowing in, but prices are holding steady. This indicates that some are taking the risk to buy, while others are selling at high levels. Chips are exchanging hands on both sides, and no one has won in the short term. Coinbase premium has been negative continuously, meaning institutions in the US are selling while Asian funds are buying, two forces opposing each other. Gold's current position is like setting an example for the market: as long as rate cut expectations exist, assets with logical coherence can rise. BTC not keeping up is not because the macro logic is wrong, but because it carries heavier burdens itself. It needs the CPI data to truly settle before confirming the next direction. If CPI continues to cool down on Wednesday, the probability of BTC catching up is not low; gold has already paved the way. If CPI rebounds, rate cut expectations retreat, gold pulls back, and BTC may have to keep grinding. At this position, before the direction becomes clear, it's not worth heavy betting; now is just waiting for the data. $XAU $BTC $ETH #现货ETF资金回流，BTC与ETH能否接力？

Apple is testing Changxin Memory's DRAM chips, planning to use them across multiple product lines of iPhone and MacBook. If this news had come out two years ago, probably no one would have paid much attention. But at this point in time, the signal it sends is much deeper than it appears on the surface; end manufacturers are actively seeking alternative suppliers, and the landscape of memory chips may be loosening. The position of memory stocks is indeed very delicate now; earnings reports have all exceeded expectations, yet stock prices are falling. SanDisk $SNDK revenue surged 372%, gross margin 84.6%, and approved a 14 billion buyback, yet still fell 7% after hours. Western Digital's profits doubled but dropped 11% after hours. SK Hynix $SKHYNIX had a maximum drawdown of 54%, Samsung Electronics 42%. The better the performance, the harder the fall; the market simply doesn't buy it. However, after the Korean stock market opened today, SK Hynix rose over 4%, Samsung Electronics over 3%, and KOSPI once rose over 2%. The previous leveraged selling pressure is easing, and after the new regulations on leveraged ETFs took effect, the scale of related products is shrinking. With fewer sell orders, stock prices naturally bounce. Note this is a short-term sentiment repair, not a fundamental reversal. Many people treated Apple's move to Changxin as ordinary news and moved on, but its significance is actually considerable. Apple's supply chain certification threshold is extremely high. Once Changxin truly enters the chain, it means Chinese memory chips have entered the supply system of a global top-tier end manufacturer for the first time. For Samsung and SK Hynix, the short-term threat is not significant, as Changxin's capacity and yield still lag behind the Korean giants. But the "broken window effect" is more meaningful than the actual order volume—it breaks the certainty of the original pattern. The price increase cycle of memory chips over the past year was built on two foundations: "highly concentrated supply + AI demand explosion." If end manufacturers start actively seeking alternative supplies, even if only at the "testing" stage, it sends a signal to the market that the supply-demand pattern is not immutable. This marginal change in expectations may have a more lasting impact than a single earnings report. Bank of America is also signaling that Samsung may launch a special dividend exceeding 30 trillion KRW plus a 40 trillion KRW buyback, and SK Hynix may use 50% of free cash flow to reward shareholders. Korean conglomerates have historically been stingy with shareholder returns. If even they are pushing return plans, it shows management realizes "just telling stories can't sustain valuations anymore." Shareholder returns are "something companies have to do after stock prices fall to a certain level," not "a signal that fundamentals have improved." The current state of memory stocks is that short-term rebounds exist, but medium- to long-term pressure remains. Leverage selling pressure has eased, so a stock price bounce is normal. Apple choosing Changxin is a variable in the long-term landscape; it won't change anything immediately, but the direction is clear: the supply side is loosening. Expansion is also proceeding simultaneously; the three major memory manufacturers are expected to invest $146 billion in equipment next year, 3.4 times that of 2024. AI demand is indeed still there, but the "scarcity premium" is gradually being diluted. The memory sector won't die; AI demand remains, but the days of "price increases while lying down" may be gone. Today's rebound is more a natural recovery after sell-off exhaustion, not a strong entry of new buyers. Wait for real signals to emerge; what we need to do now is wait. Patient waiting will bring good results. $SAMSUNG #存储股抛压缓和，AI内存牛市还稳吗？

Wednesday's CPI, honestly, this data is quite critical. The current expectation is that the overall CPI year-on-year will drop from 3.5% to 3.4%, and the core CPI from 2.6% to 2.5%. The power of capital is indeed strong, easily pushing it down. But the question is, how much of a drop is enough? This is the real dilemma for the market. Whether to raise interest rates in September or not, the non-farm payrolls have already shifted expectations toward a rate cut, but CPI is the true judge. If CPI continues to fall, the probability of a rate hike will decrease further, and risk assets might rally again. If CPI rebounds, rate cut expectations will retreat, and the market will have to reprice. $BTC is still hovering around 65000 without moving. ETFs have been seeing net inflows; BlackRock put in nearly 700 million this week, money keeps coming in. But the price just can't push up; 65000 feels like a wall. If CPI cools down, there might be a catch-up rally, possibly breaking through 66000 or even 67000. If CPI rebounds, it might retest 64000 or even lower. $BICO has really been giving me a hard time lately. The two short positions I opened a few days ago both got squeezed out. I opened one at 0.0314, it got pushed up. I wasn't convinced and added another, but it still got pushed up. However, last night I finally got a correct bet, and this wave basically recovered the losses, but honestly, closing the position was a bit early, I didn't maximize the gains. $BEAT fell from 3.4 to 2.6, down nearly 30%. For coins that rise fast and fall fast, once volume shrinks, it's hard to chase. I suggest just watching from the sidelines. $SNDK is still stuck at 1200, not moving up. The storage sector's expectation correction is ongoing. Is storage really hopeless? All the funds have fled now. Can no one save SanDisk? Before the CPI data comes out, avoid major position adjustments. Instead of betting on direction, wait for the direction; decide after the data lands. Now it's just a waiting game. Wednesday is the moment to decide the outcome. #本周三CPI公布，9月加息定价会改写吗？