Orbit Post Sitemap
Has NVIDIA suddenly become a major shareholder of SpaceX?
The latest data shows that Nvidia holds about 122.8 million shares of SpaceX, valued at around $21 billion at the end of Q2, making it its second-largest single equity asset
But I don't think the focus is on this $21 billion. Nvidia previously invested in xAI, and as xAI merged into SpaceX, this investment ultimately became a $SPCX $NVDA of SpaceX shares
NVIDIA is moving from simply selling GPUs to moving upstream in AI industry capital
Selling GPUs → customers building data centers → NVIDIA investing→ customers continue to buy GPUs
This is actually an increasingly complete cycle of AI capital
SpaceX's Q2 revenue has reached $7.8 billion, nearly doubling year-on-year, with AI-related revenue growing even faster
So what I'm more concerned about is not how much SpaceX can rise, but whether Nvidia will continue to replicate this model. If AI capital spending keeps expanding, this approach will only get bigger
But once AI financing cools down, Nvidia will shift from being a shovel seller to one simultaneously bearing industry valuation fluctuations. This may be the real focus of this investment
Non-investment advice for DYOR
#英伟达深入AI资本链. How to balance synergy and risk Everyone with money sitting somewhere asks the same quiet questions, and none of them have clean answers.
The person with cash in a CD wonders if inflation is eating it alive. Right now it isn't top CD rates sit near 4.50%, inflation's at 3.4%. But that gap is thin, and it depends entirely on rates staying where they are.
The person eyeing gold wonders if they're already too late. Fair question gold's up over 60% this year. Rallies this steep have gone sideways for years before, more than once in gold's own history. Nobody knows which version this is yet.
The person looking at crypto hears "scam" before they hear anything else hacks, rug pulls, laundering, an asset class plenty of people still call a Ponzi scheme with extra steps. Some of that reputation is earned. Some of it ignores that $51.9B has moved into regulated Bitcoin ETFs alone. Both things are true about the same industry at once.
The person watching US stocks keeps hearing the AI story is priced in, that there's no room left to climb. Maybe. This week's AI infrastructure earnings were genuinely mixed some beat, some didn't. The market hasn't decided either.
The person circling real estate keeps waiting for it to drop further. Here's what's actually happening: new home supply sits at 9.4 months a buyer's market. Existing homes sit at 4.4 months a seller's market, because most owners are locked into mortgages under 6% and won't sell into today's rates. Two different markets, same country, same month.
Every one of these doubts is at least partly reasonable. None of them resolve by staring at the asset harder. They resolve by deciding what you can actually afford to be wrong about.🐕 $DOGE: A truly special place
The most interesting thing about DOGE may not be technology or scarcity, but that it possesses an internet cultural symbol that is hard to replicate.
Years have passed, new memes keep emerging, but DOGE remains widely known.
If the crypto market becomes increasingly dependent on brand, culture, and user perception in the future, will DOGE become one of the most enduring "internet-native assets"?
DOGE's true value may lie in its cultural presence 🐕🌐🔥苹果现在最容易被低估的一件事，可能不是它有没有做出最强的AI，而是它根本不需要做出最强的AI。
过去两年市场聊AI，最喜欢比模型能力。OpenAI、Google、Meta拼参数，$NVDA 卖算力，$AMD 想抢第二名，大家默认谁的模型更强，谁就更接近赢家。但放到 $AAPL 身上，这套逻辑其实有点不适用。苹果真正恐怖的地方从来不是某项技术第一个做出来，而是它手里已经有十几亿台设备，可以把一项原本很小众的技术直接塞进普通人的日常生活里。
这也是为什么我觉得苹果的AI故事不能只看模型排行榜。假设未来最强的大模型来自Google，最强的算力还是NVDA，苹果完全可以把这些能力整合进iPhone、Mac、Watch甚至AirPods里。普通用户不一定知道后台调用的是谁的模型，只会知道自己的手机能帮他总结邮件、处理图片、理解屏幕内容、自动完成一些原本需要手动操作的事情。苹果真正想争的不是“谁训练出了最聪明的模型”，而是“谁控制用户每天和AI发生接触的那个入口”。
这个逻辑和Google、Meta甚至OpenAI都不太一样。Google拥有搜索和Android，Meta掌握社交和内容，OpenAI在抢AI助手本身，而苹果最强的是硬件入口。你每天可以不用ChatGPT，可以不刷Instagram，甚至不用Google搜索，但一个iPhone用户很难一天不用自己的手机。只要AI最终从“偶尔打开一个App问问题”，变成系统里无处不在的功能，操作系统和硬件入口的重要性反而会越来越高。
当然，这也是苹果现在最大的压力。AI如果只是多几个功能，很难让几亿用户突然换一台新手机。真正能重新推动iPhone超级换机周期的，必须是某些旧设备根本做不了、而且用户用了以后真的回不去的能力。否则市场喊了两年“AI iPhone”，最后大家发现手里的旧手机照样能用，AI故事就很难转成真实收入。
所以接下来我看AAPL，反而不太在意苹果下一次发布会上模型跑分能不能赢Google。我更想看的是AI能不能真正改变用户换机理由。如果未来一个人买新iPhone，不再只是因为摄像头更好、芯片更快，而是因为新设备能真正替他完成更多事情，那苹果才算把AI变成了自己的生意。
NVDA赚的是所有人训练AI的钱，Google和OpenAI争的是谁的AI更聪明。
苹果赌的则是另一件事：不管最后谁的模型最好，用户每天用AI的时候，最好还是先经过我的设备。
AI时代最贵的不一定是模型，也可能是那个离用户最近的入口。
#AAPL #Apple #GOOGL #NVDA #META #AI #美股 #科技股 #欧易星球Price pressures are easing, but new alarms have already sounded.
The pillar that has long supported the U.S. economy—household consumption—is quietly softening its stance.
The recently released data is not optimistic: retail sales in July fell 0.6% month-on-month, a sharp contrast to the market's expected 0.1% increase, which is also a rather unfavorable figure in over a year. Confidence is also declining, with the University of Michigan's Consumer Confidence Index dropping to 51.0, falling short of expectations and clearly weaker than before.
The situation is changing:
What the US is currently struggling with is no longer "whether inflation will remain high," but "inflation hasn't cleared away, demand has already softened."
This is the most tricky script for the Federal Reserve.
In recent years, they've kept rates high to cool down demand and inflation. Now it seems the temperature has indeed dropped—July's CPI and PPI both show that price pressures on the production and consumer sides are easing, and weak retail sales also indicate that the effect of high interest rates has finally penetrated the household sector.
But side effects are starting to appear.
If consumption keeps falling and interest rates remain high, the economy will be squeezed even harder; But if interest rates are cut now, prices could make a comeback.
So the real key in September is not just looking at the inflation report, but how to find a balance between "stabilizing prices" and "supporting the economy."
Many people are watching the rate cut trade, but the essence of market pricing is actually this turning point:
The U.S. is moving from "demand is too hot" to "demand is turning cold."
......#消费动能转弱, September policy remains constrained by inflation
$BTC $ETH $SNDK What will 🔥 happen to $BTC $ETH next?
#消费动能转弱, September policy remains constrained by inflation
The vast majority of traders still judge the market based on a single inertia of thinking, with BTC driving the entire crypto market up and down in the same direction. But in the coming period, the divergence between the two will continue to widen, and their correlation will temporarily decrease, which is the biggest trap in the upcoming market.
Let's start with $BTC:
Bitcoin's current pricing logic is continuously "decoupling from crypto-native sectors," increasingly leaning toward commodities and alternative safe-haven assets. Macro interest rate expectations, US Treasury yields, and volatility in the US stock market are the core variables driving the market, and the popularity of the altcoin sector continues to weaken its influence.
In the short term, the market will enter a narrow consolidation bottoming mode. A large amount of uneven selling pressure accumulates above, lacking sustained incremental funds, making it difficult to break out of a continuous rally; But long-term whale chips are firmly sealed, and the potential for deep declines is also sealed.
Next, there is a high probability of repeated range shakeouts, constantly testing support and resistance to wash out contract leverage. Only two scenarios can break the volatility: first, Fed officials send clear signals of rate cuts to attract ETF funds to flow back; second, safe-haven funds flow in as buying.
Simply put: BTC has "limited decline and weak gains," with a prolonged volatility cycle that makes it difficult to break out of a one-sided trend.
Now let's look at $ETH. This is where most people tend to fall into traps:
Ethereum is now a dual-asset asset: on one hand, it is bound to macro risk sentiment following BTC, while on the other, it heavily depends on on-chain ecosystem funds, DeFi, Layer 2, and RWA market hype.
The biggest current pain point: institutional funds are in a state of "strategic recognition but tactical wait-and-see" regarding ETH. Everyone is optimistic about the long-term dividends of staking ETFs, but before policies are implemented, funds are reluctant to position in advance.
In the short term, there's a very real phenomenon: every round of rebound has strong ETH impulses, but very poor persistence. Once market sentiment weakens, Ethereum's pullback is often larger than BTC's.
There are two scenario simulations for subsequent trends:
Scenario 1 (Positive): Fidelity's staked ETH ETF has made positive progress, attracting institutional long-term capital inflows, stabilizing and rebounding the ETH/BTC exchange rate, marking a phase where the market is stronger than Bitcoin;
Scenario 2 (Continued Weakness): Policy remains unresponsive, the market is dominated by existing funds competing, funds keep clustering around BTC, ETH continues to passively follow, and rebound highs keep moving downward
(Personal opinion analysis only, no investment advice)
Everyone moves forward steadily. Wishing you great wealth and better and better timesSNDK's strength is driven by a short squeeze, but the true meaning behind breaking through 1580 has yet to be verified. Is this rally driven by new demand, or is it a temporary rebound caused by liquidation? While SNDK's recent upward trend may seem strong based solely on price fluctuations, the data reveals a different nature. Over the 24 hours, positions liquidated were overwhelmingly more short than long. This means that the main driving force behind this rally is forced liquidation of short positions—a typical short squeeze structure—rather than increased buying demand. Positions that had been holding short positions below 1400 rushed to stop losses and liquidations, pushing the price higher. Looking at the market structure, when short liquidations occur in a chain, prices surge in a short time, but at the same time, they tend to exhaust future buying demand in advance. In other words, the current rise is close to price distortion caused by leverage imbalance. If you actually see funding costs entering overheated territory or a sharp increase in basis, it is more likely to be a profit-taking zone for longs already entered rather than chasing buying.🚨 ANOTHER TWIST IN CRYPTO REGULATION
The agency canceled a scheduled crypto rules meeting, while its tokenization innovation exemption remains delayed.
The Senate’s next CLARITY Act test has also been pushed to September 15.
For ecosystems like $SUI , clearer rules for tokenized assets could become increasingly important.
Three moving pieces. One evolving regulatory story. ⚖️🔵Every major BTC rally begins at a turning point in macro liquidity.
March 2020 — The pandemic collapsed, and the Federal Reserve implemented unlimited QE. BTC rose from 3,800 to 69,000.
Early 2023 — The pace of rate hikes slowed, and the market began to "shift" pricing. BTC rose from 16,000 to 70,000+.
What about this time?
At the July 29 FOMC, the Federal Reserve kept rates unchanged for the fifth consecutive time, at 3.50%-3.75%.
The key point is—rate hike expectations are collapsing.
At the beginning of August, the market priced in a 55% chance of a rate hike in September.
After the CPI was released, it dropped to 44.1%.
By August 15, CME data showed that the probability of keeping rates unchanged in September had risen to 67.5%, while the probability of a rate hike dropped to just 32.5%.
From 55% to 32.5%—this is not the end, but a signal that the Fed's narrative is starting to loosen.
Short-term traders see "BTC not rising."
Long-term holders see that "the spark has been lit."
The probability of a rate hike dropped from 55% to 32.5%. This is not the end, but a sign that the Fed's narrative is beginning to collapse.
Consumer data shifted from "strong" to "unexpected decline"—this is not volatility, but a trend.
The trend has already taken shape, just waiting for confirmation from the Federal Reserve.
And once confirmed—BTC's explosion always begins when most people are still hesitating.$BTC 美股那边7月零售销售环比跌了0.6%，大家都不买单了，消费端这不就开始撑不住了吗？而且信心指数也一路滑坡。虽然通胀预期还有点反复，但9月加息估计是彻底没戏了，甚至市场都在提前博弈降息。资金要是从美债流出来，黄金和BTC绝对是第一受益者。
但有意思的是，宏观经济虽然喊着要衰退，AI这块却跟独立行情似的。OpenAI年化搞到400亿，Anthropic二季度直接翻倍，估值冲着两万亿去。这说明市场上根本不缺钱，大家只是不敢乱投，全都抱团堆到AI大模型这种有硬需求的龙头上了。
底层硬件也跟着疯狂卷，海力士半年砸了18万亿韩元扩产HBM。我现在唯一担心的就是，如果宏观消费真被高利率拖垮了，光靠这几家AI巨头买算力，能不能消化掉存储巨头们疯狂吐出来的产能？
简言之，短线看宏观数据互掐，震荡少不了；但中长线逻辑太清晰了，降息预期+AI算力真需求，BTC和AI相关板块肯定还是主线。
大家现在是清仓防衰退，还是逢低继续抄底？$BTC $SNDK $OKB
#加密估值转向收入，BTC如何定价？ The CPI data is out, with a year-on-year increase of 3.4%, and the PPI has softened. Expectations for rate cuts are being hyped up. 📊 Logically, this should be a bull market charge, but what about our big brother Bitcoin and second brother Ethereum? One was happily around $63,500, with a daily range below 500 points; the other was hovering around $1,890, repeatedly touching the 1,900 level but still couldn't break through. 😅 It's like dealing a good deck to a player, and then they glance at it, swipe the tiles, and say, "I'm done playing"—isn't that infuriating? The reason is actually not mysterious: the market has always only speculated on expectations, not on the news itself. This positive data had already been consumed by the gains from the previous few days. When the real data came in, those smart money that had been lying in wait not only didn't increase their holdings but actually left with profits. This is what the old chives often say: "When good news is released, bad news comes out." In plain language: by the time you hear good news, the price of good news has already been paid. 💰 Not to mention that there are about $140 million worth of options to expire and be delivered tonight. Both bulls and bears are walking like they are carrying a thunderstorm, with neither daring to take the first big step. The scene looked like a group of people on a mountaintop before a thunderstorm, watching the distant lightning while weighing whether to hurry down the mountain or bet that the rain wouldn't fall on them. ☔️ So, don't think that good data means prices should rise. The market never plays by the rules. When everyone is crowded at the same door, waiting to get rich, that door is likely not the road to wealth, but rather the gateway for institutions to sell goods. The real signal never comesNvidia's deep involvement in the AI capital chain is no longer just a chip company.
It sells GPUs while pulling Wall Street to finance clients, packaging computing power leasing, data centers, and chip residual value as investable assets. This move is smart: if customers can't afford it, help them find money; cloud providers spend too much capital, so they bring in pensions, insurance funds, and private loans together.
But this is where risk starts to become more complex.
Previously, NVIDIA only needed to prove chips sold well; now it also has to prove that chips bought by margin financing can truly generate cash flow in the future. GPU depreciation is fast, model competition is fast, and the payback cycle for data centers is very long. If AI revenue falls short of expectations, who will take on these debts and residual value risks?
I don't think this is a signal of a bubble burst, but it indicates that AI has entered the financialization stage.
When technology narratives are at their peak, capital often invents new levers. The real question is: does collaboration amplify efficiency, or does risk hide within the asset package?
#英伟达深入AI资本链. How to balance synergy and risk SanDisk rose 37% in two weeks from 1226 to 1687—Old Mo tells you whether to keep flying or take a breather next week at this level
Brothers, SanDisk hit a high of 1687 today and closed at 1652, up 1.82%. Starting from the August 5th earnings report low of 1226, it has rebounded over 460 points over two weeks, with a cumulative gain of 37%. This bullish candlestick before Friday's close fixed the week's gain at about 30%.
Let's start with the technical side, with a few key signals.
On the 4-hour chart, the upper Bollinger Band is at 1764, the middle band at 1502, and the lower band at 1239. The price of 1652 is trading between the middle and upper bands, forming a strong zone, but about 112 pips away from the upper band. The SAR steering signal at 1597 is being pushed below — trend confirmation is bullish. SuperTrend 1522 forms support below.
MACD fast line at 97.47, slow line at 86.61, energy bar at 21.73. Compared to yesterday's chart (fast line 12.19, slow line 15.34, energy bar -6.29), bears have converged, bulls are regaining their head. The fast line surged directly from 12 to 97, indicating strong momentum in this rally. The energy bars have turned positive from negative to positive, indicating that the short-term correction may be over.
Key levels: First resistance above is 1687-1700; a breakout is at 1750-1764; The first support below is at 1630-1650; a break below targets 1590-1600, and further down is 1560-1570.
What happened this week? Three things.
First, Investor Day provided guidance beyond expectations. An 80% gross margin target, 100% excess cash return to shareholders, and an HBF roadmap—the market voted in favor with real money.
Second, the storage industry's prosperity remains unchanged. 93.9 billion yuan long-term contracts lock in future revenue, HBF as an additional option not included in the financial model, and the new QLC platform just launched. Industry analysts believe that even if the new Dalian factory is successfully put into operation, it will most likely only ease supply pressure moderately and will not directly change the overall global NAND market supply shortage.
Third, the overall market environment is cooperating. CPI and PPI have cooled simultaneously, with the probability of a rate hike in September falling below 40%. The S&P 500 hit new highs, risk appetite is rebounding, and funds are flowing back into growth stocks and tech hardware sectors.
What do you think about next week? Lao Mo shares a few judgments.
With the weekend closed, liquidity will shrink, and the direction of Monday's opening will be crucial.
If Monday can break through 1687-1700 with increased volume, the upside potential opens up, with 1750-1764 or even higher. The upper Bollinger band at 1764 is the next real test—at this level, the market will reassess whether short-term valuations are reasonable.
If the price shrinks and consolidates sideways in the 1687-1700 range, or even pulls back, it indicates that short-term profit-taking needs to be digested. A pullback to 1630-1650 is a healthy correction; as long as it doesn't break 1590-1600, the bullish structure is sound.
If it unexpectedly breaks below 1590-1600, it means the strongest phase of this rebound has ended and a longer period of consolidation is needed.
The operation suggestion is divided into two scenarios.
Brothers with positions: If your cost is below 1400, the profit is already substantial. I suggest cutting positions by one-third to half around 1687-1700, and moving the stop-loss up to below 1590 with the remaining positions, aiming for a chance to break through 1750-1764. Why do this? It rose 37% in two weeks, and there are already quite a few short-term profit-taking positions. Reduce positions to lock in profits, set protective stop-losses on the remaining positions, and let profits run. This is a method that balances risk and return—if it rises, you still have positions; if it falls, you won't give back all your gains.
Brothers wanting to enter with short positions: Chasing at 1652 now, stop loss is hard to hold. Wait for two signals: either wait for a pullback to stabilize at 1630-1650 before buying, set stop-loss below 1590, target 1687-1700, and look for a breakout at 1750-1764. Or wait for volume to break above 1687-1700 before chasing to the right, stop loss below 1650, target 1750-1764. Directly buy heavily near 1650, not cost-effective—upside potential is about 30-50 points, downward stop loss may require 60-80 points, odds not very good.
Lao Mo concludes: SanDisk rose from 1226 to 1687, a 37% increase in two weeks. The positive news from Investor Day is still being digested, and the medium- to long-term logic hasn't changed. But in the short term, 1687-1700 is a key watershed—once it's gone, a new world awaits; If you can't get past it, pull back and gather strength to push again. We'll see the outcome on Monday.
Did you get a share of SanDisk this time? What do you think about next week? Let's talk in the comments.
If you think Lao Mo is so clear-cut, give a like and follow. When I reach the key opening position on Monday, I'll call you right away. $BTC $ETH $SNDK #交易之声: Your experience deserves to be heard Latest update on US-Iran situation: (According to current mainstream open-source media information) 1. Trump confirmed that the USS Lincoln has withdrawn and replaced the new carrier, dispelling previous concerns about US military escalation in the Middle East. If the new carrier is deployed jointly with the USS Lincoln, it signals military upgrades, and this replacement can be seen as preparation for prolonged operations. 2. On August 14, the Canadian government imposed a new round of sanctions on Iran, totaling five people. This signal signals that the diplomatic rivalry between the US and Iran has expanded to include collective pressure from Western allies on Iran, representing increased diplomatic pressure. #消费动能转弱, September policy remains constrained by inflation. 3. Iranian Foreign Minister responded to Trump's remark that "the Strait of Hormuz is U.S. territory," stating that Qatar and Pakistan still maintain communication with Iran but have not decided to resume negotiations with the U.S. This message is a positive signal, with the key being that the U.S. and Iran have directly stated they will close negotiations with the U.S. It is good that the ≠ has not been decided to shut down completely and that communication with the mediating country is still being maintained. 4. According to the latest Kpler statistics, only two ships confirmed transit in Hormuz on Friday, indicating a near-stall phase. This news will actually affect the international crude oil supply balance and future expectations, which is positive for crude oil. This week, the initial daily number of ships increased from 5 to 9, and now to 2, which is indeed not very optimistic. 5. Israeli airstrikes on southern Lebanon have caused multiple deaths and serious injuries. On Friday, the Houthi forces fired six ballistic missiles into the Red Sea region, resulting in multiple deaths and injuries. Both pieces of news are negative, meaning that although there has been no clash between the US and Iran, the proxies have already acted on their behalf#OpenAI与Anthropic估值竞赛升温
Let's talk about something the community has been watching closely these past couple of days—the valuation race between the two AI giants, OpenAI and Anthropic. Honestly, it has a bigger impact on us Crypto players than you might think.
On OpenAI's side, they just completed a $7 billion employee stock buyback. The valuation is anchored at $852 billion, consistent with the $122 billion valuation from the March funding round. The company secretly filed for an IPO in June, but this buyback move actually suggests the listing won't happen anytime soon.
Why? Employee stock buybacks usually provide liquidity to employees and early investors while stabilizing the equity structure—in other words, "guys, don't rush, we'll cash out some money for you now, but hold off on the IPO."
Anthropic, on the other hand, is much more aggressive. After completing its Series H funding in May, its valuation hit $965 billion, directly surpassing OpenAI. Recently, the secondary market valuation soared to $1.5 trillion, a 25% increase in one month. The most astonishing part is that several investors expect the IPO valuation in October to reach $2 trillion or even higher, surpassing SpaceX to become the largest IPO in history.
The revenue data is also impressive—Q2 quarterly revenue was $11.5 billion, a 14-fold year-over-year increase, and it achieved adjusted operating profit for the first time. Annualized revenue surged from $14 billion at the start of the year to $47 billion in May, with investors predicting it could reach $100-120 billion by year-end.
What does this have to do with us?
Did you read Arthur Hayes' "Situationship" article a few days ago? He said the AI infrastructure investment boom is very much like the real estate bubble, and if it bursts, it could trigger a credit crisis similar to 2008. But the aftermath might force central banks worldwide to "super-print" money to rescue the market, which could actually become the catalyst for the next Bitcoin bull run.
In the short term, it's a different story—the AI infrastructure financing is currently draining liquidity from the Crypto market. Market maker GSR also mentioned that equity issuances by large tech companies building AI infrastructure are tightening liquidity across all asset classes, including crypto.
So this moment is quite delicate—OpenAI's $852 billion valuation is hanging there, with the $7 billion buyback hinting at a delayed IPO; Anthropic's valuation is at least $965 billion, with the secondary market already at $1.5 trillion, and possibly a $2 trillion IPO in October. Whoever goes public first, and at whatever valuation, the liquidity siphoning effect on the market is undeniable.
What do you think about this AI valuation bubble? Do you believe Hayes' logic of "AI bubble burst → super money printing → BTC bull market"? Or will the short-term liquidity drain hammer the market first?Bears face a test amid market resonance: $SNDK Analysis of the momentum behind the short squeeze In the current crypto and financial market landscape, traders who choose to short $SNDK against the trend are facing significant risks. This strong short squeeze was not caused by a single factor, but rather the result of multiple favorable factors including market chip structure, fundamental news, chain reactions, and the overall economic environment. First, the chip side shows an extremely asymmetric overcrowded state, laying a heavy short-fueled foundation for the subsequent surge. During the previous price correction, many investors blindly predicted that the rally had peaked, rushing to build short positions. This one-sided tilt is vividly displayed on major trading platforms like OKX—the total number of short accounts once reached 1.8 times that of long ones. When the market direction suddenly reverses, the total amount of short liquidations triggered within 24 hours surges to nearly $40 million. These forced liquidations directly convert into the strongest driving force behind price surges. Second, the releasing of major positive news from industry and company fundamentals has become the direct trigger for this short squeeze. During this period, SanDisk officially announced a long-term operating outlook that far exceeded market expectations and successfully signed a long-term supply agreement worth up to $93.9 billion; Combined with the ongoing spread of shortage expectations across the entire storage chip industry, strong fundamental support instantly disrupted the bears' defenses, forcing the first batch of short positions to trigger stop-losses and exit. Furthermore, the market evolves into a standard chainCapital expenditure surged 70% in the first half of the year! Is SK Hynix's gamble truly legendary, or is it just setting a trap?
SK Hynix's capital expenditure in the first half of the year surged by over 70% year-on-year, pouring all its resources and ammunition into HBM, advanced packaging, and next-generation NAND production lines.
Anyone who's worked in the semiconductor industry for a few years would feel a chill down their spine when they saw these terrifying capital expenditure figures. Because the memory chip industry has played out the same script over the past decades: crazy capacity ➡️ expansion during booms, concentrated capacity release ➡️, price crashes ➡️, and industry-wide losses cutting capital expenditures.
Many people are asking: Is SK Hynix's aggressive move this time a strategic move to seize the initiative, or is it overextending at the top of the cycle?
To be fair, Hynix is now "on the verge of being roasted and has no choice but to gamble." Without expansion, Samsung and Micron, both eyeing the market, could snatch HBM market share in no time; But frenzied capacity expansion is essentially betting on huge future depreciation costs to bet that downstream AI computing demand will never slow down.
But reality is often not perfect. Everyone must recognize a trick in the current hardware supply chain — double booking.
When AI computing power cards are at their most scarce, major downstream manufacturers often inflate order demand several times to lock in capacity to ensure they can secure supply. Once the training computing power of large models reaches a bottleneck, or when major cloud giants start to seriously assess ROI, the first to be cut from orders and trigger a price crash will definitely not be Nvidia's GPUs, but storage chips as bulk components.
Once spot and contract prices plunge, expensive new production lines like Hynix instantly turn from money printing machines into profit-eating depreciation black holes.
That's why, in the AI hardware supply chain, I never dare to hold a heavy position in the storage sector. Ultimately, storage is still a cyclical commodity logic. Those who truly cross cycles and have absolute bargaining power are always at the top of the chain, controlling chip core architecture and advanced foundry pricing.
If AI investment cools down slightly in the next one to two years, do you think the first to crash will be memory chip prices or TSMC's foundry gross margins?
---
The above content represents personal views only and does not constitute any investment advice. DYOR，NFA。
#海力士扩产提速, whether capital expenditures can deliver returns ETH现在最难受的是被机构重新算账了
$ETH 最尴尬的地方，不是它没有叙事，而是它的叙事太容易被拿去计算。
BTC可以说自己是数字黄金，可以不谈现金流；DOGE可以说自己是meme，可以不谈估值；但ETH不行。ETH有质押收益，有链上费用，有L2结算，有稳定币和DeFi沉淀，所以市场会很自然地把它放进一张表里，和美债、货币基金、科技股、平台类资产一起比较。
这听起来像进步，实际上也是压力。成熟资产才会被算账，但被算账就不能只靠愿景涨。机构看ETH，不会只听“以太坊生态最大”这句话，它会问质押收益扣掉费用以后还有多少，波动率值不值得承受，监管会不会影响托管，L2把费用分走以后ETH本身还能捕获多少价值。
所以ETH现在最需要的可能不是一个新口号，而是更便宜的钱。只要无风险利率还在高位，ETH那点链上收益就没有想象中性感。机构可以喜欢ETH，但喜欢不等于马上配置，尤其当国债和现金类产品还能提供不错回报的时候。
这就是ETH和BTC最根本的差异。BTC的核心叙事是稀缺，ETH的核心叙事是使用。稀缺可以对抗财政扩张，使用却要面对收益率比较。前者更像保险，后者更像生产资料。市场紧张时，保险更容易被买；市场愿意承担风险时，生产资料才会被重新定价。
但我不认为ETH现在的处境是坏事。被拿去算账，说明它已经进入机构资产池的语言体系。问题只是，当前这个宏观环境对它不够友好。利率不下来，风险偏好不起来，链上活动又没有进入极端繁荣阶段，ETH就容易卡在一个“有基本面但没弹性”的状态。
真正值得看的是稳定币和RWA。它们不像meme那样热闹，却是ETH最硬的需求来源。只要高价值资产继续愿意在以太坊体系里结算、抵押、托管，ETH的地基就还在。短期价格不兴奋，不代表长期结构消失。
ETH最怕的不是BTC强，而是自己生态变大却捕获不到价值。如果L2增长、稳定币增长、机构应用增长，最后都不能转化为ETH的需求，那才是大问题。现在市场只是在等证据，不是在宣判死刑。
这轮ETH要重新赢回资金，靠的不会是一句“世界计算机”，而是把收益、结算、安全和资产沉淀重新连成闭环。等这条线重新顺起来，市场自然会给它换一套估值表。
对ETH来说，最好的修复路径不是一天暴涨，而是几项数据慢慢对齐：稳定币继续留在以太坊体系，L2活动没有削弱安全需求，质押收益在降息周期里重新有吸引力，机构产品没有被监管卡住。只要这些条件逐步出现，ETH的估值会比情绪更先修复；如果这些条件迟迟不来，反弹就容易变成短线资金自救。
ETH最怕的不是没人讨论，而是讨论太多却没有价格锚。等收益率、费用和结算需求重新对齐，它才会从被动挨打变成主动定价。在那之前，任何反弹都要先看资金是不是愿意把短线仓位变成配置仓位。现在的ETH不是缺故事，而是缺一个能让机构停止犹豫的确认信号和更便宜的宏观资金环境支撑。
---What will happen to BTC and ETH next?
The vast majority of traders still judge the market based on a single inertia of thinking, with BTC driving the entire crypto market up and down in the same direction. But in the coming period, the divergence between the two will continue to widen, and their correlation will temporarily decrease, which is the biggest trap in the upcoming market.
Let's start with $BTC
Bitcoin's current pricing logic is continuously "decoupling from crypto-native sectors," increasingly leaning toward commodities and alternative safe-haven assets. Macro interest rate expectations, US Treasury yields, and volatility in the US stock market are the core variables driving the market, and the popularity of the altcoin sector continues to weaken its influence.
In the short term, the market will enter a narrow consolidation bottoming mode. A large amount of uneven selling pressure accumulates above, lacking sustained incremental funds, making it difficult to break out of a continuous rally; But long-term whale chips are firmly sealed, and the potential for deep declines is also sealed.
Next, there is a high probability of repeated range shakeouts, constantly testing support and resistance to wash out contract leverage. Only two scenarios can break the volatility: first, Fed officials send clear signals of rate cuts to attract ETF funds to flow back; second, safe-haven funds flow in as buying.
Simply put: BTC has "limited decline and weak gains," with a prolonged volatility cycle that makes it difficult to break out of a one-sided trend.
Looking at $ETH, this is where most people tend to fall in:
Ethereum is now a dual-asset asset: on one hand, it is bound to macro risk sentiment following BTC, while on the other, it heavily depends on on-chain ecosystem funds, DeFi, Layer 2, and RWA market hype.
The biggest current pain point: institutional funds are in a state of "strategic recognition but tactical wait-and-see" regarding ETH. Everyone is optimistic about the long-term dividends of staking ETFs, but before policies are implemented, funds are reluctant to position in advance.
In the short term, there's a very real phenomenon: every round of rebound has strong ETH impulses, but very poor persistence. Once market sentiment weakens, Ethereum's pullback is often larger than BTC's.
There are two scenarios for the subsequent trajectory:
Scenario 1 (Positive): Fidelity's staked ETH ETF has made positive progress, attracting institutional long-term capital inflows, stabilizing and rebounding the ETH/BTC exchange rate, marking a phase where the market is stronger than Bitcoin;
Scenario 2 (Continued Weakness): Policy remains unresponsive, the market is dominated by existing funds competing, funds keep clustering around BTC, ETH continues to passively follow, and rebound highs keep moving downward# Capital Inflows and Selling Pressure Cosom: Ethereum Stalls Near the $1800 Level Despite recent positive news in the market, Ether ($ETH) price remains rather sluggish, consolidating around $1880 without showing strong follow-up momentum. From the perspective of capital flows, institutional interest in the overall cryptocurrency market is actually quite high. Last week, net purchases of Bitcoin and Ethereum ETFs reached $1.1 billion, and the related net inflow trend has continued for five consecutive days, with an increase of $6.7 million recorded on Wednesday alone. Meanwhile, traditional financial institutions are actively making moves—Fidelity has officially submitted an application, planning to stake all its $ETH holdings; UBS has further expanded its service scale, fully opening up trading channels for crypto assets for retail investors. However, these highly positive industry developments have not directly translated into strong price increases, driven by intense profit-taking pressure. Data shows that a whale-level investor who built a position at a low of $1,637 in June this year has recently started frequent sell-offs, transferring over 10,000 $ETH to the FalconX platform, successfully realizing a profit of about $2.47 million. This kind of early long-term holder reduction on rallies directly led to significant resistance in the key price range of $1,861 to $1,899, with weekly price fluctuations even being compressed to within 2%. All things considered$BTC $ETH #消费动能转弱，9月政策仍受通胀制约 What happened to Ethereum? It has been tentatively testing support breaks but hasn't maintained accelerated declines, forming a consolidation pattern. Is this a bear trap or a bull trap?
📊 Overall Capital Flow Overview
As of August 15, the total net asset value of U.S. Ethereum spot ETFs was approximately $10.521 billion, with the ETF net asset ratio (market cap as a percentage of ETH total market cap) reaching 4.64%. The historical cumulative net inflow has reached $11.453 billion.
However, since the beginning of the year, Ethereum ETFs have still experienced a net outflow of about $873 million, consistent with the broader environment where Bitcoin ETFs have seen a net outflow of about $4.44 billion since the start of the year.
📈 Recent Capital Flow Trends
August Capital Flow Timeline
Date Capital Flow Key Events
August 3 Net outflow of $11.9 million BlackRock ETHA redemption of $9 million, Grayscale ETHE outflow of $7.8 million
August 5 Net inflow of $60.85 million BlackRock ETHA single-day inflow of $50.34 million, hitting a recent high
August 12 Net inflow of $7.4 million Entirely absorbed by BlackRock ETHA, zero flow in other competitors
August 13 Net inflow of $6.7 million Fifth consecutive week of net inflow, longest positive inflow streak
August 14 Zero inflow Ten ETFs had no net inflow or outflow all day
In early August, a single-week net inflow of $244.9 million was recorded, the best weekly performance since mid-April, but capital flow then fluctuated, with zero inflow on August 14, indicating institutional funds entered a wait-and-see mode after a brief concentrated entry.
🏦 Fund Holdings and Competitive Landscape
BlackRock — The Absolute Leader
- ETHA (Spot ETF): As of August 5, cumulative net inflow of $11.53 billion; as of August 4, assets under management about $831 million, fee rate 0.25%
- ETHB (Staking ETF): As of August 5, cumulative net inflow of $555 million; on August 3, a counter-trend net inflow of $5.8 million, the only product with positive inflow that day
- Data as of August 4 shows BlackRock ETHA holdings increased 12.9% over the past seven days, far exceeding other competitors
- According to CoinShares research head James Butterfill on August 4, 2026, BlackRock has surpassed Grayscale to become the largest digital asset fund manager by total assets under management
Grayscale — Dragged Down by High Fees
- ETHE: As of August 4, assets under management about $5.068 billion, but with a high fee of 2.5%, far above the industry average of 0.25%, causing continuous capital outflows
- Morgan Stanley Q2 13F filings show an increase in Grayscale ETHE holdings by about 26% to 5.1 million shares, but overall Grayscale remains in net outflow at the market level
- ETH Mini ETF: As of August 6, 80.8% of ETH has been staked, with an annual net staking yield of 2.61%, having earned $27.3 million in staking rewards
Fidelity
- FETH: As of August 14, cumulative net inflow of about $2.13 billion; actively applying to the SEC to stake all ETH holdings and distribute earnings quarterly, which if approved will greatly enhance product competitiveness
Other Players
- Bitwise ETHW, VanEck ETHV, etc., are smaller in scale, with capital flows basically zero or slight outflows
🔄 Core Trend Summary
Trend 1: Highly Concentrated Capital, "Winner Takes All"
Institutional funds show extreme concentration. On August 12, the entire market's $7.4 million net inflow was absorbed by BlackRock, with Fidelity, Bitwise, and Grayscale showing zero net flow that day. This reflects institutional investors' cautious sentiment, only willing to bet on the most liquid and strongest brand leading products.
Trend 2: Staking Yields Become the New Competitive Focus
- BlackRock launched the independent staking product ETHB in February
- Grayscale amended its trust agreement in August to make staking the default option
- Fidelity is applying to stake all FETH holdings
Staking yields (around 2.6% annualized) are becoming the core differentiator for ETF products; in the future, "whether staking yields are offered" may determine a product's survival.
Trend 3: Traditional Financial Institutions Accelerate Entry
- Morgan Stanley significantly increased crypto ETF holdings in Q2, with BlackRock ETHA holdings up 202% to 4.6 million shares, while systematically reducing Coinbase and other crypto stocks
- Wells Fargo increased ETHA holdings by 63.5% and Bitwise ETHW by 37% in Q1 2026
This marks a structural shift from "crypto stock speculation" to "standardized ETF allocation" by traditional institutions.
Trend 4: Grayscale's High Fee Model Faces Challenges
Grayscale ETHE fees at 2.5% vs. industry average 0.25%, a 10x difference. Although Grayscale still leads in assets under management, continuous capital outflows and BlackRock's rapid catch-up in holdings are eroding Grayscale's leading position.
💡 Impact Assessment on ETH Price
ETF capital flows impact ETH price as "strong support but insufficient push":
- Support: Five consecutive weeks of net inflows + continuous whale accumulation (whales accumulated 112,000 ETH within three weeks) provide bottom support
- Pressure: $873 million net outflow since the start of the year + FG Nexus selling 50,000 ETH + historically weak August seasonality (median return -1.87% over past 11 years) limit upside
In the short term, ETF capital flows are more likely to provide support in the $1,860–$1,900 range, but to push ETH past the key resistance at $1,930 (100-day EMA), larger and more sustained net inflows are needed.When will $CORE public chain explode as fast as possible?
1. Scenario A: Triggered fastest (low probability, 12-18 months, around mid-2027)
At least two heavy catalysts must be hit simultaneously:
(1) The U.S. SEC has approved BTC yield-type LST ETFs based on Core underlying layers, allowing compliant funds from European and American institutions to enter the market;
(2) Custody institutions like BitGo/HexTrust, through Core's lstBTC, saw institutional staking scale increase (billions of dollars), generating real on-chain business income and initiating continuous token buybacks;
(3) Combined with Bitcoin being in the main rally phase of a new bull market, overall risk appetite across the market remains high.
2. Scenario B: Neutral scenario (highly probable, 2028-2029, mid-to-late stage of the next Bitcoin bull market)
US ETF approval delays without super compliance benefits;
The BTCFi sector is booming overall, with a large amount of existing Bitcoin assets starting to be staked for interest; Core, as one of BTCFi's infrastructure, follows market cycles to realize valuations;
However, funds will be diverted by projects in the same sector like Stacks and Babylon, reducing flexibility.
3. Scenario C: No Outbreak (High-Risk Reality Path)
Summary
- Theoretical fastest: around mid-2027, but this is a low-probability event, requiring a double catalyst for a US ETF + institutional staking scale explosion;
- Neutral time window: mid to late Bitcoin bull market in 2028-2029;US retail data for July fell far short of expectations, sending a macro catalyst signal to the crypto market.
July retail sales fell 0.6% month-on-month, compared to expectations of +0.1%, marking the first decline in nine months and the largest month-on-month drop in 14 months; Excluding automobiles, the data also weakened. Online retail and automotive sectors cooled significantly, with only clothing and catering remaining resilient. Overall, consumption is cooling but not yet a hard landing.
The market has directly repriced rate hike expectations: the probability of a rate hike in September has dropped to 30.6%, compared to 50% a month ago. Consumer spending hitting the brakes will limit the Fed's room for further tightening, theoretically opening a rebound window for risk assets.
But it's important to distinguish between expectations and reality: ≠ data turns dovish, the coin price directly surges. Currently, $BTC is still stuck at the key support range of 63,000, with volume shrinking in the market. Macro positive news only provides conditions; what really depends on whether ETF funds can cooperate with the flow back.
In a favorable environment, $ETH are more sensitive; Conversely, if the market fails to hold key support, highly elastic stocks can be even more damaging when pullbacks occur. $CORE When will public blockchains explode?
When will $CORE public chain explode as fast as possible?
1. Scenario A: Triggered fastest (low probability, 12-18 months, around mid-2027)
At least two heavy catalysts must be hit simultaneously:
(1) The U.S. SEC has approved BTC yield-type LST ETFs based on Core underlying layers, allowing compliant funds from European and American institutions to enter the market;
(2) Custody institutions like BitGo/HexTrust, through Core's lstBTC, saw institutional staking scale increase (billions of dollars), generating real on-chain business income and initiating continuous token buybacks;
(3) Combined with Bitcoin being in the main rally phase of a new bull market, overall risk appetite across the market remains high.
2. Scenario B: Neutral scenario (highly probable, 2028-2029, mid-to-late stage of the next Bitcoin bull market)
US ETF approval delays without super compliance benefits;
The BTCFi sector is booming overall, with a large amount of existing Bitcoin assets starting to be staked for interest; Core, as one of BTCFi's infrastructure, follows market cycles to realize valuations;
However, funds will be diverted by projects in the same sector like Stacks and Babylon, reducing flexibility.
3. Scenario C: No Outbreak (High-Risk Reality Path)
Summary
- Theoretical fastest: around mid-2027, but this is a low-probability event, requiring a double catalyst for a US ETF + institutional staking scale explosion;
- Neutral time window: mid to late Bitcoin bull market in 2028-2029;WEAK CONSUMPTION, FED STILL CAUTIOUS
The U.S. economy is sending mixed signals: consumption is cooling, but inflation keeps the Fed cautious. This leaves short-term liquidity expectations too weak to trigger a fresh risk-on wave. $BTC still has an advantage through spot ETF flows and its market-leading position, while $ETH needs stronger liquidity and real demand to regain relative strength. With the Fed’s path still #WeakConsumptionFedSplit #OpenAIAnthropicRace #SKHynixCapexSurge WEAK CONSUMPTION, FED STILL CAUTIOUS
The U.S. economy is sending mixed signals: consumption is cooling, but inflation keeps the Fed cautious. This leaves short-term liquidity expectations too weak to trigger a fresh risk-on wave. $BTC still has an advantage through spot ETF flows and its market-leading position, while $ETH needs stronger liquidity and real demand to regain relative strength. With the Fed’s path still uncertain #WeakConsumptionFedSplit #OpenAIAnthropicRace #SKHynixCapexSurge $SNDK
Don't short SanDisk, don't short SanDisk, don't short SanDisk.
There may be a short-term pullback, but in the long term, SanDisk is expected to stay around $2,000.
Domestic channel rumors (consumer drives): Agents report that wafer quotas for regular USB drives and entry-level TF cards continue to shrink, with production capacity prioritized for enterprise eSSDs; Retail regular models have reduced promotional prices and increased shortage rates, so traders are reluctant to stock up on consumption flash storage, focusing on clearing old inventory.
SanDisk (SNDK) surged strongly this week, closing at $1,641.11 on August 14, up 7.39%, with a five-day cumulative increase of about 35%. JPMorgan raised its target price to 'overweight,' with a target price of $2,250, driven by the investor day guidance on August 13. The company expects mid-to-high double-digit revenue growth for fiscal years 2028–2030, with a long-term non-GAAP gross margin target of 80%; Flash memory demand for AI data centers has become the main theme, with a market size of 1.2ZB projected by 2030, focusing on BiCS9 QLC and HBF solutions. It has signed multi-year long-term agreements with leading cloud providers to lock in most of its capacity, with enterprise-level SSD quotas and strong spot supply, making a short-term drop unlikely.
#海力士扩产提速, whether capital expenditures can deliver returns #英伟达深入AI资本链, how to balance synergy and risk, #闪迪投资者日后股价大涨 long-term goals remain to be verified SNDK at $1,650, did you miss out?
Let's look at the surface first: from hell to heaven in just two weeks.
After the August 5 earnings report, it once dropped to 1,350, hitting a low of 998 at the end of July — nearly halving from the ATH of 2,354. But starting August 10, it surged violently, skyrocketing 17.6% on Investor Day, August 13, and rising another 7.39% on August 14 to close at 1,641. In two weeks, it bounced back from 1,000 to 1,650, a rebound of over 60%.
Daily candles show consecutive gains with volume, weekly chart shows a strong reversal, RSI around 44 not overbought yet, this is not a rebound, it's a trend reversal.
First thing: Investor Day blew up, SNDK is no longer a "cyclical stock."
On August 13, SanDisk held its 2026 Investor Day in New York, unveiling a long-term financial model that stunned the audience:
FY2028-2030:
Revenue growth in the mid-to-high double digits
Non-GAAP gross margin about 80%
Non-GAAP operating margin about 75%
Adjusted free cash flow margin about 50%
100% of excess cash after investments returned to shareholders
Second thing: JPMorgan directly says: there's still 47% upside.
On August 14, JPMorgan upgraded SNDK from "Neutral" to "Overweight" with a target price of $2,250.
Analyst Harlan Sur said: SNDK is uniquely positioned in many ways to capture the structural inflection point in NAND demand driven by AI inference.
He also emphasized the value of the NBM long-term agreements — 8 NBM agreements signed, total contract value about $94 billion, average term over 4 years, structurally resetting SNDK's margins and significantly reducing cyclicality.
Third thing: 8 NBM agreements locking in $93.9 billion guaranteed minimum revenue.
This is SNDK's strongest fundamental. The company has signed NBM long-term supply agreements with 8 data center customers, including 3 major US hyperscale cloud providers.
Hard data:
Guaranteed minimum total revenue $93.9 billion
Remaining performance obligations $91.1 billion
Financial guarantee mechanism $16.5 billion
Covers over 50% capacity in fiscal 2027, about 2/3 capacity in fiscal 2028
Trading strategy
Short-term traders:
Light long positions near 1650, wait for a pullback to 1600-1620, stop loss at 1540-1550, target 1720-1750 → 1800-1850
If volume breaks through 1680-1700 on Monday and holds, add to longs, target 1800
Swing traders:
Partial profit-taking near 1680-1700, keep base positions for higher targets. Consider reducing positions if it breaks below 1550 with volume
Long-term believers:
Ignore short-term noise, invest based on fundamentals. Betting on AI storage super cycle + valuation logic reconstruction, target $2,250-$3,000+#SK Hynix Expansion Accelerates, Can Capital Expenditure Deliver Returns?
The news about SK Hynix's expansion is indeed quite strong.
A massive expansion order worth 54 trillion KRW, tied with $NVDA Nvidia, is solidifying its position as the HBM leader. Since August, $SKHY has also risen by 15 points, and market sentiment has surged. But the question arises: with accelerated expansion and such a huge investment, will it really translate into profits?
First, let's talk about the news itself. The core of SK Hynix's expansion is HBM. With the explosive demand for AI servers, $NVDA's GPUs are in short supply, and HBM, as the supporting memory, is almost a bottleneck. SK Hynix indeed has a strong voice in this field; otherwise, it wouldn't have secured such a large order. So from an industry trend perspective, the direction is sound.
What does this mean for us?
First, it's a short-term positive for sentiment. Expansion means strong order certainty and guaranteed revenue for the coming years. $SKHY and related memory stocks have short-term support.
Second, the capital expenditure is huge, which will extend the return cycle. 54 trillion KRW is no small amount—building factories, buying equipment, expanding production lines—the money is spent upfront, and profit realization will take one to two years. If the market starts to factor this in, valuations will fluctuate. Concerns about overcapacity are not unfounded. Historically, the memory industry often faces price wars following expansion waves.
Third, this also reflects on US-listed memory stocks like $SNDK and $MU. SK Hynix's expansion indicates high industry prosperity, but conversely, increased future supply will pressure NAND and HBM prices. So the US memory sector will see internal differentiation—those with orders will rise, those without will follow trends, with different rhythms.
Here’s my personal view.
I don’t hold a direct position in $SKHY. I had previously followed $MU in the US market but never entered. After this SK Hynix expansion news, I’m even less eager to jump in.
The reason is simple: expansion is a long-term story, but the short-term stock price has already priced in part of the expectations. $SKHY rose 15 points in August, and the market is already trading on the logic of “HBM leader with no order worries.” But whether capital expenditure can deliver returns depends crucially on whether downstream demand can continuously absorb the new capacity. If AI server shipments fall short of expectations or $NVDA’s pace slows, HBM supply pressure will show up earlier.
My current approach is to keep observing and not chase the highs. I’m focusing on two signals: one is $NVDA’s subsequent orders and earnings guidance; the other is SK Hynix’s own capital expenditure rhythm and capacity utilization. If later we see stable HBM prices and smooth ramp-up of new capacity, I’ll consider finding an entry point. At this stage, the story is very attractive, but the price is not cheap.
Accelerated expansion is good, but good things also need good prices. When the market shifts from "speculating on expectations" to "looking at realization," that will be the true test of this round of capital expenditure quality.
$NVDA $ETH $BTC #SK Hynix Expansion Accelerates, Can Capital Expenditure Deliver Returns? The truth behind Bitcoin's $63K correction: orderly selling pressure without liquidation collapse (capitulation).
Although Bitcoin has experienced a significant pullback, dropping from its high to the $63,000 level, unlike previous cycles, the futures market has not seen a surge in large-scale long liquidations.
Long Liquidations: The scale of forced liquidation of leveraged buying positions due to insufficient margin, which is an important indicator for determining whether panic selling has formed a cycle bottom.
Forced Sell vs. Voluntary Sell: Unlike previous years (2020 and 2021), this decline was not triggered by forced liquidation, but rather by intentional reductions mainly in spot and low-leverage holdings
Reduced leverage structure: The market is not being shaken out by severe liquidation crashes, but is orderly and gradually reducing positions and moving downward
Caution in bottoming judgment: Historically, the "mass liquidation (surrender)" phenomenon seen at cycle lows is missing, so slow inventory allocation may last for months
A sharp drop without liquidation does not mean the market is healthy; rather, selling pressure has not been exhausted all at once but is advancing slowly.
Close attention is needed to see if the last large-scale liquidation surge occurs.#消费动能转弱, September policy remains constrained by inflation
The market easily translates "bad data" as "the Fed is about to loosen." But this time, the script is not so smooth.
Retail sales in July fell 0.6% month-on-month, and the control group, which better reflects GDP consumption of goods, also dropped by 0.4%.
In September, the probability of holding rates steady once rose to about 69%, yet BTC still fell below $63,000.
The problem lies in another set of data: Michigan consumer confidence fell from 55.2 to 51.0, while one-year inflation expectations rose from 4.2% to 4.3%.
This is not a comfortable "inflation down, demand stable," but rather consumers starting to hold back, while price anxiety persists.
The Fed can raise rates one less time, but there is no reason to declare victory early; For BTC, pausing rate hikes does not mean new liquidity is entering immediately.
I won't chase long positions just because of a weak retail data. Still holding spot stocks, short-term funds waiting first: 2-year US Treasury yields and US dollar continue to fall, $BTC stabilizes above 63,000–64,000 USD and increases volume.
If at least two of the three signals appear, then pick up in batches; If it's just officials giving a dovish stance and prices not keeping up, keep watching and waiting.
Bad news turns into positive news and needs market confirmation. Otherwise, it might really just be bad news.【周期推演：为什么当前的纠结震荡，大概率在等待最后一跌与“真平底”？】
回顾我之前那篇关于“比特币熊市平底论”的分析：自由市场的出清从不靠V反，而是由“深跌刺穿 + 2至3个月极度死寂的平底磨盘”共同完成。
对照当下的盘面，行情在关键均线与通道间反复拉锯、纠结震荡。这种上不去又下不透的形态，绝非大周期见底的特征：
筹码未彻底投降： 真正的平底建立在杠杆出清后的“波动率真空”与情绪极度绝望之上，而当前市场仍充斥着博弈与抄底预期。
物理出清仍缺一环： 历史周期表明，没有经历那记抽干流动性的“最后一刺”，上方的套牢盘与获利盘就无法在低位完成沉淀。
因此，当前的纠结更像是中继修正。后市大概率还需要一次加速下探去刺破最后的防线，随后在更低的区间走出长达数月的死寂平底。
耐住性子，把更多的子弹留给恐慌宣泄后的平底吸筹期。$DOGE The core current conflict lies in the struggle between pure sentiment speculation and the unproven implementation of X Money payments; relying solely on social media slogans can no longer lock in long-term risk capital.
The position structure is significantly less sensitive to social media news, and short-term gambling funds mainly rely on Bitcoin-driven overall market risk appetite rather than independent premiums. New public chain ecosystems like Solana and Base are diverting marginal liquidity from the meme market, and the siphoning effect of old assets on capital continues to weaken.
The priority of driving variables has shifted: the high Beta rebound driven by macro risk appetite determines short-term elasticity, while the acceleration of actual payment and transfer scenarios within X platform sets the medium- to long-term valuation center. Relying solely on trading volume amplification can only sustain short-term volatility; real on-chain and in-platform usage is the decisive factor in changing valuation models.
The upside scenario is built on the rebound of risk appetite combined with specific integration events. If X Money suddenly publicly announces the deep integration of $DOGE into the micro-tip or settlement system, the market will quickly price its certainty as the platform's default payment gateway, triggering overlapping short positions and funds chasing highs. This scenario requires observing continuous improvements in on-chain transfer frequency and settlement efficiency. If follow-up funds only remain at the growth of derivatives trading volume without support from on-chain active addresses, the scenario will immediately fail.
The downside scenario is based on the assumption of narrative gaps and ongoing liquidity squeeze. If X Money ultimately becomes a standard compliant payment tool without $DOGE core functional positioning, its Musk premium will be squeezed out more rapidly. As other emerging public chain meme assets erode liquidity, traders will tend to reduce their positions as old bull market assets, further converging valuations toward traditional memes with no real returns.
The critical point for judging failure depends on how the market allocates capital. If, without actual payment implementation, $DOGE breakthrough is achieved relying on extreme market preference, it indicates that sentiment funds still dominate; Conversely, if on-chain data remains weak after payment functionality is implemented, it indicates that the market's distortion of the payment narrative has been completed early.
Over the next 7 days, focus on monitoring the regulatory licensing progress for X Pay and changes in on-chain holding addresses, while monitoring the overall allocation of risk assets and the alignment between derivatives holdings.
#海力士扩产提速. Can capital expenditures deliver returns? #AMD完成历史最大美元债发行: Financing of $4.75 billionSanDisk $SNDK: 1641 SanDisk, rising or falling?
First, the conclusion: the medium-term outlook is bullish, but in the short term, wait for a pullback.
Market view: Friday closed at 1641, +13.68% for the day, up 35% for the week, severe short-term overheating. 43% short of the 52-week high of 2354—this is both pressure and room for gain. 1500-1550 is the confirmation zone for pullback after a sharp rise; if it doesn't break through, the upward structure will still exist.
From a fundamental perspective, this rally is not sentiment speculation, but earnings revaluation:
Fiscal year 2026 revenue is $20.25 billion, up +175% year-on-year; Net profit is $11.43 billion, turning losses into profits and taking off immediately. The biggest blowout is gross margin: 26.2% a year ago, but dropped to 84.6% in Q4! There are very few companies worldwide that can turn half of revenue into cash flow.
Why I think it will still rise:
1. The supply-demand tightness for HBM and NAND will persist until 2027, with new capacity not until 2028 to be launched, indicating a high ceiling for prosperity;
2. Performance realization is one of the highest links in the AI industry chain, not just storytelling;
3. Hillhouse increased holdings in Q2, Temasek directly entered the investment and storage sector, with top-tier funds lining up.
The risks are clear: after a short-term 13% surge, profit-taking is expected, which is normal. Historically, SanDisk dropped from its peak in June to cut a third of its market value, indicating its extreme volatility and that chasing highs can easily get you wiped out!!
#闪迪投资者日后股价大涨, long-term goals remain to be verified American consumers have begun to hit the brakes, and the Fed's toughest moment may have just 🔥 begun. Inflation has finally cooled, but an even bigger problem is emerging—the once strongest pillar of the U.S. economy, consumption, is visibly cooling off. The latest data shows that U.S. retail sales in July fell 0.6% month-on-month, not only far below the market expectation of 0.1% growth, but also the most significant single-month drop in over a year. Meanwhile, the University of Michigan's Consumer Sentiment Index fell to 51.0, also below expectations. It seems inflation has been curbed, but the price is that the speed of the consumption engine is slowing. This means the U.S. economy is no longer simply facing "excessive inflation," but a more complex new problem: inflation has not completely disappeared, but consumption momentum has already started to turn downward. This is precisely the Fed's toughest dilemma to resolve. Over the past two years, the Fed has used persistently high interest rates to suppress demand and cool the economy to combat inflation. Now, this strategy has indeed worked—both CPI and PPI in July showed price pressures easing, and inflation on both the production and consumption sides cooled simultaneously. But on the flip side, the high interest rate environment is being transmitted to ordinary consumers through credit, mortgages, car loans, and other channels, with the decline in retail data being the most direct evidence. This raises the question: if consumption continues to weaken while the Fed maintains high interest rates, the momentum for economic recovery will be further damaged; If the shift to easing is too early, the risk of an inflation rebound could return. So the key to the September policy meeting is not just about watching CPsFundamental Research Report $TON / The Open Network (Public Chain/L1) $1.33 (24h +0.88%)
To put it plainly: The Open Network ($TON) has an overall score of 64/100, with a rating focused on narrative over implementation. Looking at the three layers, the company team has cash reserves, the protocol network already shows signs of paid usage, and token capture has been implemented.
Let's look at the project first: The Open Network (token $TON), public chain/L1 track. Focuses on the Telegram ecosystem, payments/wallets. Benchmarked against SOL and NOT. Traditional enterprise collaboration relies on cloud servers and contract reconciliation; during high concurrency, gas surges, TPS is limited, and cross-chain bridge security incidents are frequent. Public blockchains use a unified state machine for trustless settlement, reducing reconciliation costs. Average order value is $50-500/month, requiring USDC or fiat currency settlement. Narrative-driven track, bear market usage cut by 60-80%. Positioned as an end-to-end vertical platform. Product launch: protocol layer officially operational, on-chain dashboard shows protocol fees accumulating, showing signs of paid usage. Latest version v2026.07, valid submissions 2,407 times in the past 90 days.
On the user side, address MAU is not disclosed, DAU is not disclosed, 24h transaction volume is $15.04M, TVL not found. Wallet addresses do not equal monthly active users of natural persons; concentrated holdings of large addresses tend to overestimate real user numbers. On the revenue side, user fees are undisclosed. Supply-side revenue is about 80-90% of user fees (to LPs and nodes), protocol treasury revenue is $787.3K, token holders buy back and burn annualized without a burn mechanism. 24h transaction volume is business revenue, not revenue. Company profits do not equal protocol profits, protocol profits do not equal token holders profit. Code side: 2,407 valid submissions in 90 days, 72 active contributors, latest version v2026.07. GitHub is A-level evidence that can be directly verified. Investment background: For company equity financing, look at PitchBook/Crunchbase (A-level); for token private and public funding, refer to whitepapers, release curves, and on-chain unlock contracts (A-level); market makers and ecosystem funding are B-level but do not represent long-term holdings by tech VCs; for technical integration, see API/SDK access evidence (B-level); strategic partnerships and logo walls are D-level. Using NVIDIA GPUs does not mean NVIDIA investment, and going public on exchanges does not mean strategic investment.
On the token side, total supply is 5,231,656,820.025723, circulating is 2,758,490,316.428023 (52.7%), FDV $6.98B, next unlock undisclosed (share of circulating undisclosed), burn buyback annualized rate no clear buyback burn. Must you buy coins when using the product? Yes, strong value capture (Gas/Collateral/Service Access). Looking together with peers (unified caliber, no cross-sector random comparison): In terms of circulating market capitalization, The Open Network $3.68B, SOL undisclosed, NOT undisclosed. FDV: The Open Network $6.98B, SOL undisclosed, NOT undisclosed. Annualized revenue: The Open Network $787.3K, SOL undisclosed, NOT undisclosed. Monthly active addresses or users: The Open Network undisclosed, SOL undisclosed, NOT undisclosed. Figures are based on public data snapshots; some omissions will be supplemented by official self-reports or industry standards. Valuation: circulating market cap $3.68B, FDV $6.98B, P/S 4671.9x, FDV divided by revenue 8860.5x. Pessimistic outlook: $3.68B at 50-70% of the original price, neutral range oscillation; optimistic outlook: revenue doubled, burn landed, enterprise clients inflow, FDV P/S, aligned with the top companies. In summary: Solid fundamentals (score 64/100). Token value capture has been implemented (buyback/burn/gas). Circulating market cap is relatively high relative to fundamentals, overdrawing expectations, and FDV is moderate. Three major risks: short-term large-scale unlocking and sell-off, long-term protocol revenue reverting to zero, token demand relying solely on incentives (once incentives break and usage collapses). Key points to look at next: protocol fee weekliness, burn amount, active address retention, TVL/loan balance, GitHub version releases. Information sources are public, logic is self-developed, and does not constitute buy or sell advice. Data deviations over 30% require revaluation.
That's all for this research report. If you found it useful, please give it a follow.
#基本面研报 #加密 #研究 #OKXOrbitWEAK CONSUMPTION, FED STILL CAUTIOUS
The U.S. economy is sending mixed signals: consumption is cooling, but inflation keeps the Fed cautious. This leaves short-term liquidity expectations too weak to trigger a fresh risk-on wave. $BTC still has an advantage through spot ETF flows and its market-leading position, while $ETH needs stronger liquidity and real demand to regain relative strength. With the Fed’s path still uncertain, risk management remains more important than chasing FOMO. 我真佩服，做空$SNDK 的这波人，真是拿命在赌。
这轮逼空能这么猛，明显不是单一因素，是好几个条件撞到一起了。
1、空头自己把自己逼上了绝路。前期回调的时候，一堆人觉得见顶了，纷纷跑去开空。OKX盘面上空头账户数一度是多头的1.8倍，24小时空单爆仓接近4000万美元。这么大的空头仓位堆在那，本身就是最大的上涨燃料。价格稍微一拉，空单就开始排队强平。
2、基本面利好一个接一个。闪迪先抛超预期长期业绩规划，接着939亿美元长期供货协议落地，再叠加行业缺货预期发酵。这些消息如果放在平时，可能也就涨一涨，但偏偏赶上空头最拥挤的时候，直接点爆了止损潮。
3、连环强平，价格自我加速。价格小涨一点，就有空单触及强平线，强平产生的被动买盘又把价格推高，然后引爆更多空单。循环往复，越涨越凶，根本停不下来。这种行情里，空头不是被市场打败的，是被自己的仓位踩死的。
4、宏观环境也在帮忙。美国通胀数据降温，降息预期升温，成长赛道整体估值修复，市场风险偏好回暖。大环境不拖后腿，主动大额抛盘也少，拉升阻力自然就小。
5、资金抱团，只炒龙头。存储赛道热度高，闪迪作为板块龙头，多头资金源源不断进场。所有人都知道它短期超买了，但在逼空行情里，超买根本不是理由，资金只看谁最强。
多重条件共振，这轮逼空的爆发力确实罕见。逆势做空$SNDK，承担的风险比平时高太多。这种行情下，空头要么有极强的纪律，要么就别碰。
以上只是行情复盘，不构成投资建议。‼️$BTC $ETH
#消费动能转弱，9月政策仍受通胀制约 What Musk gave DOGE was not a perpetual motion machine, but a countdown
$DOGE The most fascinating thing about it before was that it didn't need to be explained. Once Musk posted, the market naturally filled in all the narratives: payments, Mars, Tesla, X, meme culture, retail investor rebellion—every word could become a reason to buy.
Now, this magic isn't as effective.
The reason isn't that Musk isn't important, but that the market has heard the same story too many times. The first time DOGE was mentioned, everyone found it novel; The tenth time DOGE was mentioned, and people thought it still had potential; Up to now, without the real progress of X Money, no payment usage, no merchant or content ecosystem data, just shouting won't keep funds going long-term.
DOGE's biggest contradiction is interesting: the more it tries to prove its usefulness, the less it resembles the original meme; the more it maintains its meme attributes, the harder it is to attract more serious funding. Pure sentiment assets can surge through hype, but once the payment narrative begins, the market starts asking about transfer frequency, user base, settlement efficiency, fees, and compliance boundaries. The moment a joke becomes a product, valuation rules change.
Musk can still bring traffic to DOGE. As long as X Money continues to advance, DOGE still has the potential to be imagined as part of the X payment system. But this path is not as simple as in 2021. Back then, the market was buying "what Musk might do," but now the market is watching "what Musk has actually accomplished." The former depends on sentiment, the latter on data.
For traders, DOGE should be viewed in two ways going forward. One is a meme rebound driven by the broader market, with strong BTC and strong risk appetite, leading DOGE to rise with high beta; The other is a reboot of its own narrative, such as X Pay showing verifiable scenarios. The former comes quickly and moves quickly; The latter is difficult, but if it happens, the price potential is more stable.
I don't think DOGE has completely lost its chance, since it's still one of the most recognized meme coins globally. The problem is, awareness isn't always free. The longer time goes on, the more the real product needs to keep the Musk premium going, or it will gradually become a commemorative mark of the old bull market.
DOGE is not waiting for a tweet now, but for a scenario. Without a scenario, every time Musk appears, it's just consuming leftover credit; Only with a scenario does DOGE qualify to turn memes into payment gateways.
Here, it's also important to distinguish between "transaction volume" and "usage." Even if DOGE's trading volume surges one day, it might just be short-term funds betting on Musk's news; What truly changes valuations is the continuous on-chain or in-platform data for payments, tips, transfers, and consumption. The former can heat up prices for a day, while the latter may redefine DOGE in the market.
If X Money ultimately remains just an ordinary payment wallet and DOGE has no core position, its Musk premium will continue to be squeezed; If DOGE becomes the default option for certain types of micropayments or community interactions, it will regain an unrepeatable story. DOGE's future is not in the crypto world's mouths, but in users' hands.
There's a more real pressure: the meme market itself has become crowded. Last round, DOGE was almost the default entry point for retail memes, but now each chain has its own meme ecosystem, new coins on SOL are listed faster, and Base and other chains also divert attention. DOGE's advantage is being the oldest brand and most widely recognized, but its disadvantage is a lack of novelty. Musk can help it regain old attention, but whether it can be retained after recall depends on new scenarios.
So DOGE does have a moat; its moat is called global cognition; But global cognition cannot always be used as cash flow. The market is willing to pay a premium for cognition, but it will also discount when cognition cannot be converted into usage. DOGE is now seeking new outlets in this discount process.$UNI After breaking below the $3.9 neckline support, the decline accelerated, with long leveraged liquidations resonating with spot selling pressure in the short term.
On the spot side, Cumberland transferred 3.72 million tokens into exchanges, releasing large amounts of liquidity within hours and directly breaking through buying depth.
On the derivatives side, nearly $3 million were closed in a single day, and the price decline triggered passive margin settlement, amplifying the support pressure in the spot market.
Previously, market expectations for fee switch dividends were realized early, and positive news turned into liquidity outflows. Concentrated selling of chips intensified the downward breakout of the technical pattern.
If subsequent on-chain dividend details are implemented beyond expectations and large deposits on exchanges slow down, spot buying coverage may trigger a pulse rebound after overselling.
If macro risk aversion persists and market makers continue to move large inventories to trading venues, a lack of buyer liquidity will lead prices to further downward in search of lower support.
As long as large one-way deposits by large players to exchanges continue, signals of stabilization on the right side are hard to establish.
In the next 24 hours, the most critical variable is whether market maker wallet addresses undergo a new round of large-scale spot transfers to exchanges.
#霍尔木兹通航谈判未果, pressure from the US and Iran escalated, #高盛收购Neos crypto ETFs shifted to a #特朗普因TruthSocial付费数据流遭起诉 of yield competitionAt the beginning of June, I saw SanDisk $SNDK's monthly RSI enter an extreme range and started building short positions, which I gradually closed out by the end of July. At this moment, SNDK has rebounded more than 60% from its low point, with extreme volatility. To anticipate its next movement, I think gold is a good reference asset.
Similar path: a massive and fierce rally — accelerated like a crescent moon blade — a guillotine-style deleveraging stampede — a violent rebound
The upcoming trading opportunity generally favors the short side. Starting from the current price, gradually enter short positions; if the price breaks new highs and the daily chart holds above, stop loss is a feasible approach.Nvidia fell 0.07% today. I saw an analysis on Xueqiu saying "the AI capital spending cycle has peaked."
I checked the data: global AI chip purchases only rose 8% quarter-on-quarter in Q2, compared to 22% in Q1. Growth is slowing but not peaking yet. More importantly, NVIDIA's H200/H300 capacity is already scheduled through 2027, with full orders.
But the market isn't concerned about capacity, but "how much longer can prices rise?" I can't answer this question, but we can look at one indicator: the number of AI data centers under construction worldwide. Public data shows over 1,200 of them, two-thirds of which were newly built in the past two years, and these data centers will continue to purchase GPUs over the next three years. So AI infrastructure demand can at least last until 2029, and it's only 2026.
However, NVIDIA's risk is clear: if the monetization ability of AI applications cannot support current valuations, it will burst the bubble. Currently, the average monetization cycle for AI applications is 3-5 years, much longer than the iteration cycle of AI hardware (1-2 years). This time gap is the risk.
Brothers, the relationship between AI and crypto is very delicate now: AI infrastructure is absorbing capital, crypto is depreciating; But in the long run, AI and crypto will merge—DeFi needs AI for risk management, AI needs crypto for value exchange. It's a "seesaw" now; the future is a "dual-engine" drive.
#英伟达 #AI #加密 #英伟达深入AI资本链. How to balance collaboration and risk #海力士扩产提速, can capital expenditures pay off? Everyone, SK Hynix's recent round of expansion is indeed quite aggressive.
Cash expenditure on purchasing tangible assets in the first half exceeded 18 trillion KRW, up more than 70 percentage points year-on-year, with most of the money poured into HBM, advanced packaging, and NAND capacity. Profits and cash flow from AI memory business are being rapidly converted into the next round of capacity expansion. Logically, it's logical—if you don't expand now, wait for others to take your share, then it'll be too late to catch up.
But the market is no longer focused on whether to expand, but whether the money invested can be recouped. Whether high-intensity investment can meet the demand for AI servers and maintain technological advantages depends on whether orders can keep up, capacity utilization can be stabilized, and storage prices can continue to be sustained. If any link goes wrong, large-scale capital expenditure becomes a burden on profits and cash flow.
It's too early to judge right or wrong; the steps toward expansion have already been taken, and the real test will be in 2027 or 2028. When new capacity is released in phases, whether AI demand will still exist and storage prices can remain high will be seen whether the bill is worth the calculation.
The long-term direction is smooth, but the short-term pace is up to you. What do you all think about SK Hynix's recent expansion—is it positioning itself for the future or overdrawing expectations? Share in the comments. Wishing everyone a great weekend $BTC $SNDK $SKHYNIX SpaceX completed its largest IPO in history, OpenAI and Anthropic are about to go public one after another, and the Trump administration plans to invest in AI companies. Looking at these three factors together, the signal is clear: the U.S. is replacing the petrodollar with AI dollars. This is the largest swap for the dollar since it left gold in 1971, and also the third. The previous two anchor changes each reshaped the flow of global wealth. This time, it might be the most concentrated and exclusive one. 1. The US Dollar's Three "Anchor Changes": The Logic Has Never Changed The essence of money is not paper, but credit. And the premise of trust is that everyone needs the same thing. In 1944, the dollar was pegged to gold. Why gold? Not because of scarcity, but platinum is much scarcer than it is. Nobel laureate Robert Mondale once made a classic statement: gold became currency not for its industrial use, but because its verification cost was almost zero: anyone could independently verify authenticity through weighing and density tests, without any authoritative endorsement. In the era when information transmission relied on telegrams, this was an irreplaceable advantage, essentially a "trustless" settlement mechanism reminiscent of the pre-internet era. In 1973, the anchor switched to oil. Harvard economic historian Neil Ferguson details this turning point in "The Rise of Money": the core of Kissinger's agreement with Saudi Arabia was not oil itself, but locking "the fuel for global industry" and "the dollar" in the same trading channel. If you open a factory, you need oil; to get oil, you must first have dollars—this is the chainPure handwritten copying, not AI
July retail sales fell 0.6% month-on-month, very weak. But $IWM closed at $305.09, rising and hovering near the intraday high; The market fell during the same period. Bad news arrived, but the small-cap market did not kneel.
This is where the trouble lies. Consumer confidence is weakening, one-year inflation expectations are rising, and long-term bond yields are still rising. Small caps have not benefited from a broad rate decline. This current strength is mostly due to funds betting early on policy space, not that macro conditions have already loosened.
Small positions tested the bullish position, entering in batches from $304.3 to $305, stop-loss at 302.6, target 309.5; maximum loss per trade 0.5%, no leverage needed. Falling below 302.6 indicates that profit concerns caused by weakening consumption outweighed policy expectations. Data as of 18:24 Beijing time.BTC miners suddenly behaving like AI stocks indicate the market has shifted its valuation lens
Miner stocks have recently shown more resilience than many pure crypto stocks, which is even more worth pondering than BTC prices themselves. Because the market hasn't suddenly started to love mining more, but rather has started to re-examine miners through the lens of AI infrastructure.
In the past, miners' valuations were simple: when BTC prices rose, miners were happy; when BTC prices fell, miners suffered. Hash rate, electricity costs, machine efficiency, and halving cycles were the main variables in this business. But after AI emerged, miners' resources suddenly became more than just Bitcoin. Power, data centers, cooling, land, grid connection capabilities, and operations teams—these would all be data center assets in the AI era.
This is where the market shifts its perspective. The same mining farm, once seen as a "place to mine BTC," can now be seen as "infrastructure that can transition to AI/HPC hosting." When BTC prices fall, pure miner logic comes under pressure; But if a company can secure AI computing power leases, data center upgrades, or long-term power contracts, valuations no longer follow BTC entirely.
This line also has a negative meaning for [$BTC ](https://www.okx.com/zh-hans/trade-spot/btc-usdt) itself. If more miners turn to AI, it means the Bitcoin network's supply chain will become more complex. The good thing is that miners' income sources are dispersed, and their ability to resist cyclicality improves; The downside is that some capital expenditures may no longer prioritize BTC mining, and mining company investors will care more about AI contracts than Bitcoin beliefs.
So buying miner stocks and buying BTC are no longer the same thing. Buying BTC is buying non-sovereign scarce assets; buying miners is buying energy, hashrate, operations, and cyclical flexibility. In a bull market, both rise together; in bear or volatile markets, the gap widens. BTC may move sideways, miners rise because of AI data center contracts; BTC may rebound, but miners lose due to expansion costs and debt pressures.
AI is not a get-out-of-jail-free card for all miners. What can truly be revalued are companies with low-cost electricity, scalable land, grid-connected resources, and execution capabilities. Those miners who only say "We can switch to AI too" will ultimately be restored to their original state by financial reports.
This line is best used to observe market risk appetite: if funds are willing to buy miners' AI transformation stories, it indicates that the boundaries between crypto and tech growth are remerging; if funds only buy BTC ETFs and avoid miners, it indicates the market is more defensive.
BTC is an asset, miners are a business. AI amplifies this distinction, making mining no longer just a shadow of Bitcoin's price.
This also reminds many people not to treat miner stocks as simple leverage against BTC. In the past, this approximate relationship was somewhat useful because miners' income was almost entirely tied to mining; Now, once AI hosting, data center leasing, or electricity contracts appear in financial reports, mining companies become hybrid assets. They may rise when BTC is weak because of AI stories, or lose when BTC is strong due to expansion costs, debt, or underperformance in computing power upgrades.
So miner lines are best suited for "industry chain screening," not for blindly buying a basket. Whoever has real customers, who has cheap electricity, who can convert mining farms into commercial-ready data centers, deserves to pay AI premiums. Only talking about concepts without submitting contracts will ultimately come back to the mining cost list.
For BTC holders, there's another reminder from this line: don't automatically interpret miners' rise as BTC must rise. When miners rise, it may mean the market is buying AI infrastructure; When BTC rises, it may be that the market is buying non-sovereign scarce assets. The two intersect, but they are not the same transaction. Understanding this clearly helps avoid misjudging BTC's own capital flow when mining stocks are active.
Truly strong industry trends often lead to two valuation methods for the same company. Miners are now at this crossroads: one is the Bitcoin cycle, the other is the AI data center cycle. Whoever can turn both sides into financial reports is the one who is not just a concept.Introduction: Breaking the "V-Reversal Myth" and the Cyclical Journey In traditional financial markets (such as US stocks and forex), investors are accustomed to V-shaped recovery triggered by central bank liquidity "put options" or strong market rescues: liquidity crises trigger panic crashes, followed by sudden policy boosts, and asset prices quickly rally and bottom out. However, observing Bitcoin's complete macro cycles over the past decade (especially the 2018-2019 and 2022-2023 cycles), we find a very distinct and recurring pattern: Bitcoin's absolute bottom at the end of a bear market is never an instant "V-shaped reversal" that breaks through instantly, but rather after experiencing the final capitulation, it falls into an extremely low volatility and very low trading volume for 2-3 months. Flat Bottom / Sideways Base." Even during early bull market rebounds and pullbacks (such as in the second half of 2019 and the second half of 2023), bottoming patterns still showed strong stable sideways consolidation characteristics. Why can U.S. stocks reverse like a V-reverse, while Bitcoin's "flat bottom" pattern at the bottom is a structural and game-theoretic inevitability? This article will logically and rigorously break down the underlying factors behind this phenomenon from five dimensions: differences in market rescue mechanisms, principles of chip clearing, chip concentration and on-chain cost lines, derivatives/liquidity games, and psychological "dead silence periods."$SNDK 在涨，韩股在涨，我在看，闪迪站上1600了。投资者日那天涨13.7%，第二天没跌。韩股从低点反弹22%，三星、SK海力士在拉。
闪迪涨是因为AI存储需求、140亿回购、长期目标。韩股涨是因为全球AI资本开支在持续，存储和光通信板块在回暖。同一个逻辑——AI硬件在支撑估值。
但韩股10天反弹22%，闪迪2天涨13.7%。为什么韩股反弹更猛？是仓位回补还是新资金在进？韩股之前跌得太深，杠杆爆完、筹码出清后，回补的速度自然更快。闪迪涨得有基本面在撑，韩股更多是仓位修复。一个是重估，一个是回补。性质不一样，持续性也不一样。
两个市场有同一个问题：供给在扩，需求能不能跟？闪迪FY2028中高双位数增长，毛利率80%。$SKHYNIX 产线2026年下半年才投产。预期全部打在前面，兑现全部还在路上。十天的量把之前的下跌全吃了，剩下的空间是留给故事，还是留给业绩？投资者日之后股价暴拉：闪迪这波大涨，到底是在交易未来还是在给PPT买单？
闪迪在投资者日画出了一张极其夸张的长期增长大饼之后，二级市场的股价就像打了鸡血一样连着往上拔。
华尔街的分析师们突然间全都在重新给它算估值：什么 AI 服务器对大容量 eSSD 的饥渴需求、什么极高的自由现金流回报率、什么未来几年的净利润翻倍空间……
行情火爆的时候，大家永远只看最顺眼的故事。但我个人觉得，在这个位置去盲目追高闪迪，性价比已经低到了极点。
为什么？因为现在的股价走势，是在拿未来三到五年的“最好情况”在给当下定价。
市场现在假设的是一个毫无瑕疵的完美剧本：假设企业级 SSD 需求能持续指数级爆发、假设高毛利产品良率一路顺畅、假设同行绝对不打价格战。但在硬核存储行业混过的人都知道，存储厂的长期规划，哪一次不是在晶圆良率波动、同行大降价和下游客户去库存的毒打中打折兑现的？
对于 AI 存储这个板块，如果你现在手里没有低位成本的底仓，最明智的策略就是管住手，等财报给出的硬指标来做右侧验证。
我只盯两个最核心的数据：
第一，企业级大容量 SSD 的实际出货均价（ASP）和毛利率，到底能不能在接下来的一两个季度里实现扎实的环比提升；
第二，渠道和厂库里的存货周转天数有没有出现被动拉长。
在情绪最热的时候给别人的 PPT 买单，往往是套在山顶的最快路径；等数字在季报里被实打实确认、股价经历过一轮洗盘降温后，再找机会切入，胜率要大得多。
面对闪迪这轮连阳逼空，你觉得它是真的迎来了基本面质变，还是主力在借利好出货？
---
以上内容仅代表个人观点，不构成任何投资建议。DYOR，NFA。
#闪迪投资者日后股价大涨，长期目标待验证 $OKB Recently, OKB has shown clear independent movement, now around $107, up about 15% in 7 days and over 30% in the past month. During the same period, the entire crypto market continued to decline for a week, indicating that this wave is not simply following BTC but actively investing in OKB.
✔ After last year's massive burn, the total supply of OKB was permanently fixed at 21 million. Although not a recent update, fixed supply amplifies the impact of new buying on price.
✔ OKB is the only gas token on X Layer. This year, OKX launched Exchange OS. Before creating a trading marketplace, project teams need to stake OKB, and the market begins to re-trade its future ecosystem needs.
✔ After breaking through the $100 mark, the 24-hour trading volume surged by about 72%, attracting many trending funds and chasers.
✔ The CPI easing only improved market sentiment, but BTC's rebound quickly gave it back, so macro factors are only supporting and not the core reason for OKB's rise.
Next, I will focus on $109–$111. If the market holds steady with increased volume, there is still room for further upward movement; If it surges and then falls below $100, it could easily pull back to $92–$95 in the short term.
My view is that OKB is indeed very strong now, but it's no longer at a comfortable low level. Being strong doesn't mean blindly chasing highs; waiting for a pullback to confirm is safer.