Orbit Post Sitemap
#闪迪财报前夕, HBM and storage shortages continue to be market focal points
$SNDK This SanDisk grid strategy has been running for most of the day, with a position at 1391, and the current price is at 1465, with an unrealized profit of about 17%. So far, Grid has traded 509 times, and I plan to finish this strategy before 11 PM tonight.
SanDisk will release its earnings report early tomorrow morning. Continuing to hold positions now is essentially no longer a grid strategy, but betting on the direction of the financial report. Grid strategies are suitable for volatile markets, but not for such high-volatility events. Rather than betting on the outcome, it's better to lock in profits first, wait for financial reports to be delivered and direction clear, and then revise the plan.
Recently, the market has repeatedly played out the same scenario—earnings reports beating expectations, but stock prices falling. $AMD and SpaceX are typical examples. Even if SanDisk delivers impressive results, if the market had already priced in the positive news in advance, after-hours selling pressure could still have occurred; But once the financial report falls short of expectations, the decline tends to be even more pronounced.
In either case, continuing to hold grid strategies during earnings reporting means taking on additional and hard-to-control risks.
Many times, locking in existing profits is more important than betting on the next candlestick.
#DailyOrbit #意大利大行减IBIT普通股94%，加仓质押ETH
🔥意大利最大银行"抛弃"IBIT转投ETH质押，机构调仓逻辑变了
Intesa Sanpaolo Q2砍了IBIT普通股94%，但ETHB持仓翻三倍——这不是看空加密，是机构从"买β"转向"吃α"。
---
▸ IBIT普通股：646,809股 → 40,723股（-93.7%），市值从2485万刀缩到136万刀
▸ 新增IBIT看跌期权：50万股put，价值1665万刀——对冲或表达短期谨慎
▸ ETHB（质押ETH ETF）：116,200股 → 349,600股（+200%），市值710万刀
▸ ARKB：基本没动，仍持有347万股（6763万刀），仍是最大加密持仓
▸ SOL质押ETF：2817股 → 7股，基本清零
---
关键洞察：不是逃离BTC，是优化配置结构
Intesa没有全面撤退。ARKB几乎没卖，说明对BTC长期敞口还在。但把IBIT普通股换成ETHB，逻辑很清晰：
BTC ETF = 纯价格敞口（β）
ETHB = 价格敞口 + 质押收益（β+α）
在ETH年化质押收益约2-3%的环境下，机构同样承担加密波动，ETHB还能额外吃息。对于银行资金来说，"有收益的波动资产"比"纯波动资产"更香。
而且IBIT的call期权从250万股砍到1.8万股，同时新开50万股put，说明短期对BTC价格偏谨慎，但不愿意彻底下车。
---
对市场的两个信号
1. ETH质押叙事正在机构端升温
BlackRock的ETHB今年2月才在纳斯达克上市，AUM约5.6亿刀。Intesa作为意大利最大银行（总资产9927亿欧元）三倍加仓，是传统金融对"生息加密资产"投票。
2. BTC ETF的"无脑买入"阶段可能结束
IBIT此前是机构配置BTC的首选，现在出现大行调仓到竞品（ARKB）+衍生品对冲。说明机构不再满足于"买了放着"，开始精细化管理和战术性调整。
---
一句话
Intesa的调仓不是加密寒冬的信号，是机构加密配置从"粗放式"进入"精细化"的标志。ETH质押ETF正在成为银行资金的新宠。
---
你更看好BTC的纯β敞口，还是ETH质押的β+收益模式？👇06｜真正值得担心的是信用，而不是资本开支本身
这轮下跌并非完全没有基本面依据。
真实利率上升、企业债利差扩大、CDS价格上涨，说明信用市场确实开始担心AI建设的融资压力。
如果未来的数据中心建设主要依赖债务，那么AI周期会越来越像互联网泡沫时期：
企业依靠债务扩张，一旦算力供需失衡或项目回报下降，融资立即收紧，资本开支随之快速坍塌。
所以真正需要观察的，不是云厂资本开支数字有多大，而是这些资本开支能否转化为经营现金流。
如果算力合同持续提价，GPU利用率保持高位，云厂经营现金流继续加速，那么大部分建设仍然可以依靠内部现金流完成。
只有当经营现金流停止改善，同时企业不得不大量举债维持建设，AI资本周期才真正开始变得危险。
07｜HBM和长期供应协议正在改变存储周期
AI竞争不仅取决于谁拥有最多GPU，还取决于谁能够拿到足够多的HBM。
对于固定计算能力来说，更多、更快的内存通常意味着更高的Token输出。因此，HBM已经成为Amazon、Google、AMD和英伟达竞争中的关键资源。
云厂正在与内存厂签署长期供应协议，也就是LTA。
过去的存储周期中，价格下降后，客户往往会重新谈判甚至撕毁合同。但现在撕毁LTA的风险明显提高。
因为一旦未来HBM重新短缺，曾经违约的客户可能拿不到足够的产能分配。对于正在竞争AI市场份额的云厂来说，失去HBM供应可能比多支付一些采购成本严重得多。
这意味着内存行业的商业模式可能正在发生变化：
客户愿意牺牲短期低价，换取长期供应安全；存储厂牺牲部分现货涨价收益，换取更稳定的收入和产能利用率。
传统存储周期未必消失，但波动方式可能与过去不同。
08｜为什么这套环境对英伟达尤其有利
市场目前给英伟达的定价，实际上隐含了一个判断：
当前利润处于不可持续的周期高点，未来GPU价格、销量和毛利率都会明显回落。
但英伟达的优势已经不只是芯片性能。
英伟达GPU拥有最成熟的软件生态、最活跃的二级租赁市场和最容易评估的残值。因此，当数据中心项目需要融资时，债权人通常更愿意接受英伟达GPU作为底层资产。
英伟达也正在参与更多数据中心项目、云厂融资、土地电力匹配和AI公司的股权投资。
它正在从单纯的芯片供应商，变成AI基础设施生态的组织者。
只要融资环境没有彻底关闭，英伟达最容易帮助客户获得资金；如果融资环境真的收紧，新增算力建设放缓，现有英伟达GPU的稀缺性和租赁价值反而可能进一步提升。
这也是为什么“信用收紧必然利空英伟达”的逻辑，并没有表面上那么简单。
09｜AI模型公司也很难主动踩刹车
前沿模型竞争是一场典型的军备竞赛。
对OpenAI、Anthropic、Grok等公司来说，买太多算力意味着烧钱过度，但买少了可能意味着模型能力落后。
一旦竞争对手通过更多算力完成更大规模训练，技术领先优势可能在几个月内被逆转。
所以只要资金允许，模型公司很难主动停止购买GPU。
这并不是因为每家公司都确定算力能够立即产生回报，而是因为算力不足可能直接导致公司失去产品、用户和融资市场的领先位置。
AI公司购买算力，既是在追求增长，也是在购买生存权。
10｜真正的风险信号其实很明确
AI基础设施逻辑并不是无法证伪。
未来最需要观察的信号包括：
第一，GPU租赁价格是否持续大幅下降。
第二，企业是否开始公开表示GPU过多或利用率不足。
第三，云厂旧合同续约时，新价格是否低于原合同。
第四，OpenAI、Anthropic、Grok、Cursor和开源推理云的总体Token需求是否出现停滞。
第五，云厂经营现金流是否停止改善，资本开支是否开始严重依赖债务。
第六，持续学习和高样本效率技术是否显著减少大模型训练所需算力。
第七，数据中心建设是否受到更严格的电力、用水和监管限制。
这些指标一旦同时转弱，AI资本周期才可能真正进入拐点。
目前的问题是，股价已经提前交易了这个结局，但多数产业数据还没有走到那里。
11｜市场周期本身也被AI压缩了
现在大量投资者会把新闻、财报和产业消息输入Claude或其他模型，让AI快速判断利好与利空。
同一个模型对同一条消息的解释通常不会相差太大，结果就是大量资金在极短时间内形成相同结论。
过去需要两三年完成的产业预期交易，现在可能六周就经历上涨、拥挤、兑现和崩盘。
基本面周期未必变快，叙事和价格周期却大幅缩短了。
这也是当前AI硬件股波动如此剧烈的原因之一：
股价可以在真实订单尚未落地前就完成整轮上涨，也可以在基本面尚未恶化前提前完成整轮杀估值。
短期股价越来越难直接代表产业趋势。
$SPCX $SNDK $BTC
#SpaceX首份财报超预期，解禁仍是关键变量
#AMD财报超预期，增长已被透支？ The wet soil pressed against my profile, and through the optical sight three thousand meters away, the heavy tactical target codenamed SpaceX had silently completed its secondary charge.
After lurking in the snowy shelter for two whole quarters, my anemometer finally detected the violent movement of the airflow. This post-IPO giant delivered $7.81 billion in second-quarter output, a year-on-year surge of 92%; More importantly, blood loss control — operating losses narrowed sharply from 970 million yuan in the previous cycle to 143 million yuan. In the sniper's eyes, this wasn't just a dull ups and downs; it was a signal to kill the target's arterial bleeding with hemostatic forceps, and vital signs returning to peak condition.
A more dangerous killer move is hidden above the stratosphere. Its "Starmind" space-based micro computing payload tied to NVIDIA effectively elevates a military-grade fire control computing array to low Earth orbit. On the low-orbit battlefield, traditional ground computing power shelters are like open-air trenches, facing the all-around scanning of the space-based Sky Eye system, leaving no line of defense.
My secondary scope is synchronously locked onto the $XGOOGL of the linked target. As a multi-axis mirror on the capital alliance front, $XGOOGL's trading spots are shaking violently. But I still held onto the safety guard, keeping my pulse at 45 beats per minute.
Those who truly understand the industry never leave the environmental parameters on the dashboard. Take a closer look: August 6 is an extremely dangerous crosswind window—the unlocking window for the first issue of restricted shares. Up to 20% of restricted chips will grant free throw rights. The earliest primitive capital lurking in the trenches had their rifles filled with doubled profit bullets. At this moment, a wave of intense dumping pressure barrage sweeping the position is sure to strike.
In a hail of bullets, blind charging is the worst thing. Before the profit-loss ratio widens to 1:5, any rash pull of the trigger is just a sitting duck for the opponent. Veterans lurking in the grass waited for the pressure storm on August 6 to wash away the mud and sand, to see who was retreating and handing over their chips, and who was quietly resupplying ammunition under the cover of craters.
The bullet was loaded, wind speed level three, and the target was stationary at the center of the crosshair.
#SpaceXBeatEstimates This round of AI crashes is not supported by data
Over the past month, many AI hardware stocks have retreated 40% or more from their peaks. The market is trading a very pessimistic story:
AI capital spending has already exceeded, GPUs are about to become oversupply, open-source models are destroying business models, cloud factories will ultimately have to rely on borrowing to sustain construction, and NVIDIA's current profitability is at the peak of an unsustainable cycle.
The problem is, at least from the industry data currently available, this story has yet to be verified.
On the contrary, GPU rental prices, token usage, difficulty in acquiring computing power, and cloud factory operating cash flow still point to one conclusion:
The real demand for AI infrastructure has not slowed down, and market prices are clearly diverging from industry fundamentals.
01|The market first misunderstood Meta's rental computing power
After Meta prepared to lease out some GPU computing power, the market quickly concluded that Meta had idle capacity, AI demand was below expectations, and capital expenditures may be cut in the future.
However, Meta's subsequent disclosures did not support this judgment; it did not cut capital expenditures, nor did it show any obvious contraction on the supply chain.
A more reasonable explanation is that Meta found that early low-priced hashrate contracts had a huge price gap with current GPU spot rental prices.
When the market is willing to pay far more than the contract cost for the same GPU, renting out part of the computing power isn't because it's unused, but because the return on investment is too high.
Meta's rental of computing power may not indicate a lack of demand, but rather that computing power has become an asset that can be independently monetized.
02 | GPU prices have not dropped as the market expected
Over the past two years, even the most optimistic investors generally believed that as supply increased and new generation chips were released, old GPU rental prices should gradually decline.
But reality is completely different.
The interview mentioned that a popular AI startup rented thousands of B200s a few months ago, priced at about $2 per GPU hour. About seven months later, it was ready to rent nearly identical clusters, with the offer approaching $4 per GPU hour.
For the same GPU and cluster size, rental prices have risen by about 50% to 60% in just over half a year.
Some inference cloud companies even estimate that the cost of renewing Blackwell computing power may nearly double after the existing contract expires.
This is completely different from the traditional semiconductor cycle.
Normally, the longer a chip has been on the market, the lower the unit price of computing power should be. Now, both old and new computing power are in short supply. Companies are not discussing GPU surplus, but rather how to secure enough GPUs.
03|Cloud companies may not be overly profitable, but are seriously underestimating their own computing power
Many cloud providers and Neo Cloud sign three- or even five-year computing power contracts in advance to secure financing.
When these contracts were signed, GPU rental prices were far lower than they are now. This creates a very critical price difference:
The computing power revenue currently recognized on the cloud factory still follows past low-price contracts, but the same computing power is already much more expensive in the spot market.
Suppose an old contract was priced at $2 per GPU hour, and the current spot price is $4. Even if the spot price falls from $4 to $3.30 in the future, the new contract will still be priced 65% higher than the old one.
Therefore, when judging cloud factories' future revenue, you shouldn't just compare future spot prices with today's spot prices. Instead, you should compare:
How much has the price of the new contract increased compared to the old contract that is about to expire?
As long as contract repricing continues, cloud factories' operating cash flow may keep improving. Even if GPU spot prices fall in the future, it does not mean cloud business revenue will decline.
04|Open-source models target model layer profits, not computing power demands
There is another common misjudgment in the market about open-source models:
Open-source models are getting cheaper, so AI revenue declines, and ultimately, GPU demand will also decline.
Here we confuse two completely different issues:
Who earns profits from a token and how much computing power is required to generate it are not the same thing.
Open-source models may indeed squeeze the gross margins of leading model companies like OpenAI and Anthropic, but generating tokens still requires GPUs, video memory, power, networking, and data centers.
Token gross margins for frontier closed-source models may be very high, while open-source inference services have much lower profit margins. After companies adopt open-source models, the price they pay decreases, and the main savings are the model layer profits, not the underlying computational costs.
The outcome is likely to be:
Individual tokens are cheaper, and companies are starting to use more tokens; Frontier model companies earn less, but inference cloud, GPU, HBM, and data center demand increases.
So the real impact of open-source models may be shifting profits from the model layer to the infrastructure layer, while further expanding the entire market through price reductions.
This may not be good news for model companies, but not necessarily bad news for AI hardware.
05|The future is more likely to be multi-model architectures
Enterprises do not use just one powerful model to solve all tasks.
A more realistic architecture is:
Ordinary tasks are first assigned to open-source models fine-tuned by enterprise data, complex tasks are assigned to Claude, Grok, or other cutting-edge models, and finally, the strongest model is responsible for planning, reviewing, and correcting errors.
This approach can significantly reduce the average cost of use while maintaining the quality of the results.
However, a decrease in cost does not mean a decrease in computing power usage.
A company's original budget could only generate 100 million expensive tokens, but after switching to open-source models and model routing, it could generate 300 million or even 500 million tokens with the same budget.
Enterprise AI spending may stabilize, but underlying GPU computing hours are still increasing.
Therefore, judging AI demand should not be based solely on how much a company spends, but also at the number of tokens, GPU hours, call frequency, concurrent tasks, and the number of agents.
$SNDK $SPCX $QQQ
#SpaceX首份财报超预期, unlocking remains a key variable
#AMD财报超预期 has growth been overdrawn? Storage dropped 30%, optical modules dropped 60%, and orders sold until 2027—who is lying?
Recently, a very contradictory picture has emerged in the market.
On one side:
Shares of storage and optical modules have seen significant corrections.
Micron has fallen more than 30% from its peak, and some optical module companies have even seen halved adjustments.
Meanwhile:
On the industrial side, however, they are aggressively cutting production capacity.
DRAM and HBM capacity from Samsung, SK Hynix, Micron, and others continue to be booked in advance, with some orders already scheduled through 2027.
Although NAND is not as tight as HBM, the market still expects tighter supply.
This led to a seemingly contradictory problem:
If future demand is really this strong, why have stocks dropped so much?
If the industry has already peaked, why are downstream companies still spending large sums to lock up inventory in advance?
This actually reflects the logic of two different markets.
From the industry side, the focus is:
Demand for the coming years;
Supply capacity;
Long-term orders;
Capital investment.
The secondary market trades the following:
Expected changes;
Capital flows;
Valuation pressure;
Short-term sentiment.
The two are not always synchronized.
Many cyclical industries experience similar situations:
When the industry is at its peak, stock prices may reflect the future in advance;
During short-term corrections, the market may quickly plunge due to concerns about peaking or overvaluations.
For example, the AI industry chain.
Enterprises may still be increasing capital expenditures, cloud providers still need more computing power, but investors will start to ask:
"Can future growth continue to exceed expectations?"
As a result, the stock price experienced sharp fluctuations.
Optical modules are similar.
If AI data center construction continues to advance in the future, the demand for high-speed optical communication will still exist.
At the same time, some policy changes and supply chain adjustments may also bring new market opportunities.
But investors also need to note:
Strong long-term industry demand does not mean all companies will see sustained price increases.
The biggest risk of cyclical stocks is that the market trades for the future ahead of time.
Even in a good industry, buying at the wrong price can still cause losses.
So what really needs to be judged is not just:
"Do you have any needs?"
Instead:
"Has the demand growth already been fully priced in by the stock price?"
Back to the original question:
Which is real: the huge drop in storage and optical modules, or the continuously growing industry orders?
The answer may not be who is lying.
Instead:
The industry sees the next few years;
The market is trading for the coming months.
When the two deviate significantly, opportunities and risks often appear simultaneously.
Ultimately, the outcome is not whose voice is louder, but whose data ultimately delivers on the score. #AMD财报超预期 has growth been overdrawn? $BICO Unitree IPO: A-shares valued at 40 billion yuan versus on-chain 30 billion USD in "pricing war"
1. Core milestones in A-share IPOs
Unitree Technology (688836. SH) today (August 5) officially entered the preliminary inquiry phase for its STAR Market IPO:
· Issuance Scale: Plans to publicly issue 40.4464 million shares, accounting for 10% of total share capital after issuance, raising 4.202 billion yuan
· Estimated issue price: about 104 yuan per share, corresponding to a market value exceeding 40 billion yuan, with a single lot (500 shares) requiring payment of about 52,000 yuan
· Key timeline: Preliminary inquiry on August 5 → Issue price confirmed on August 6 → Online and offline subscription on August 10 → Payment deadline on August 12
· Equity structure: Founder Wang Xingxing controls a total of 68.78% of voting rights through special voting rights arrangements
2. On-chain Derivatives Market: A Fierce "Code War"
Before the official pricing of A-shares, the on-chain derivatives market had already engaged in fierce competition around Unitree Technology:
🏆 HIP-3 Code Auction: Two Strong Competitors
HIP-3 is an architecture on Hyperliquid that allows external deployers to create their own perpetual contract marketplaces. On August 3, decentralized exchange Paragon acquired UNITREE's code for 577.66 HYPE, becoming the first to list pre-IPO perpetual contracts. Subsequently trade.xyz purchased the same code for 565.54 HYPE (about $31,000), becoming another popular pre-IPO target after Changxin Memory and SK Hynix.
📈 Nearly 5x premium: The "madness" of on-chain pricing
Hyperliquid has launched a Pre-IPO perpetual contract for Unitree Technology, initially priced at $73.33, corresponding to a total share capital of about 404 million shares, with an implied market value of about $29.7 billion (about RMB 212.9 billion). As of press time, the contract is quoted at about $74-75, corresponding to a market value of approximately $30 billion. Based on the expected A-share issue price, the on-chain price is about 4.88 times higher than the expected A-share price—the market is betting real money on Unitree's "robot future."
⚠️ Risk warning: Pre-IPO perpetual contracts are derivatives tracking the price of specific assets and do not provide dividends or voting rights from the actual shares. Previously trade.xyz SK Hynix contract plunged 19% due to pre-market volatility in South Korea, triggering large-scale forced liquidations.
🏦 More platforms are entering the market
Gate.io and Bybit have also launched UNITREEUSDT Pre-IPO perpetual contracts, supporting up to 10x leverage.
3. FCC Ban: The "Gray Rhino" Before IPO
Just before the IPO, the U.S. Federal Communications Commission (FCC) announced new regulations on July 28, officially banning the import of new Chinese-made humanoid robots and quadruped robots. Unitree Technology, as a global leader in civilian robotics, has become one of the core targets affected by this ban.
Impact on Unitree:
· The revenue share of the U.S. market has dropped from its peak to 13.30%, with domestic revenue now surpassing that of overseas markets
· The company stated that the impact is limited, and domestic scenarios are sufficient to support large-scale implementation
· Being targeted by special legislation is, to some extent, evidence of the industry's status
4. Market Observation
Unitree Technology is staging a dual pricing battle between A-shares and on-chain platforms:
· A-share market: valuation of 40 billion yuan, 104 yuan per share, IPO subscription on August 10
· On-chain market: $30 billion in implied market value (about 212.9 billion yuan), nearly five times higher than A-shares
As the pricing milestone for A-share IPOs approaches, the pricing competition among unlisted tech giants on-chain is significantly heating up. trade.xyz's follow-up intervention will provide on-chain traders with deeper market-making liquidity. The concept of embodied intelligence is becoming a key gold-attracting target in the crypto derivatives market recently.
$HYPE The whale chasing the SPCX eventually accepted a loss and left. Starting with 0xb37 yesterday, the address had cumulatively bought 229,600 shares of SPCX, with a turnover of about $27.064 million, and a weighted average price of $117.9. This morning, all positions were sold at an average price of $114.8—a loss of about $708,000, and the net asset value of the account has been reset to zero. This rally-chasing operation lasted only about 24 hours from entry to exit. SPCX is currently at $112.39, down 3.52% in 24 hours, with a trading volume of about $874 million and open interest of $174 million. Currently, addresses with millions of dollars have placed about $24.278 million in orders on both the upper and lower ends. Whale order structure: Below support: $100 to $106.1 range, largest short cover (about $5.136 million), not new long positions; $83.8 to $96.2 with clear long capital (about $4.39 million), planning to open 50,000 new long positions. Selling pressure above: Whale sell orders accumulated in the $133.5 to $153 range about $10.761 million, with long positions taking profit about $7.617 million, new short positions about $3.144 million, and about $2.732 million short positions between $155.4 and $180. The long-short battle in SPCX is far from over. After the bullish whales cleared positions, there is a clear support structure below, with whale take-profit and new short positions above. $BTC $ETH $SPCX #SpaceX首份财报超预期, unlocking remains a key variable #财The temporary navigation agreement is yet to be implemented, and oil price risks have yet to reverse
Trump spoke with the Emir of Qatar, and Qatar said a proposal focused on the Strait of Hormuz had been drafted. US and Iranian officials were optimistic about the provisional agreement, and Iran even relented to consider allowing European countries to enter the field to clear mines. The market reaction was very direct: Brent dropped over 5% in settlement the previous day, falling below $80 for the first time in three weeks, and continued to weaken today. Brent was at $78.44, WTI at $74.63, marking three consecutive days of decline.
But a decline does not mean a risk reversal. The market is currently trading "possible signs," not "already opened." The temporary agreement only solves physical channels, but can't solve channel credit. Mines haven't been cleared, war risk rates remain high, oil tankers loading empty in the Persian Gulf are still low, and global inventories are thin. What's even more troublesome is that the agreement bypasses the tough issues of Iran's nuclear program and regional proxies, allowing either party to flip the table during the window period. Oil prices have only switched from a surge pattern back to a wide range, with support around $70, and the rebound is just as fragile.
On the crypto side, $BTC is fluctuating around $64,000, around $64,100 today, with an intraday range of 63,580 to 64,494. It's still down about 27% this year, far from the previous high of $126,000. A drop in oil prices means inflationary pressure is eased, rate cut expectations and risk appetite can ease, which is a tailwind for liquidity-hungry assets like BTC and the Nasdaq; But once the protocol fails, oil prices rebound in a V-shape, inflation expectations return, and crypto will take the hit along with risk assets. Don't mistake the drop in oil prices for a strengthening crypto fundamentals; that's just macro beta not strangling for now.
The core contradiction is simple: the market is eager to treat the expectation of general aviation as the reality of general aviation and the temporary ceasefire as permanent peace. Next, focus on three things: whether the mine-clearing and AIS signals have returned to normal, whether the empty tanker dares to return, and whether the U.S. has officially confirmed the agreement text. Before the boot hit, oil prices couldn't be called a reversal, and Bitcoin only managed to hold above 64,000 for risk appetite to truly recover. If it fell back below 62,000, it means funds are starting to requote geopolitical risks.Daily Market Watch | Asian Closing and U.S. Stock Outlook
— US stocks hit new highs; can BTC keep surging?
Today's Key Takeaway: The expectations of the US-Iran agreement have triggered global risk assets, with both the Dow Jones and S&P 500 hitting record highs, and Asian stock markets surging across the board. However, BTC repeatedly hovered above 64,000, and the sharp drop after AMD's earnings report cast a shadow over sentiment among tech stocks.
Asian review: Tech stocks collectively counterattack
Asian stock markets surged across the board today. The Nikkei 225 closed up 3.66% at 66,300 points, South Korea's KOSPI rose 3.76% to 6,598 points, SK Hynix rose 5.8%, and Samsung Electronics gained 2.5%. The A-share Shanghai Composite rose 1.47%, and the STAR 50 rose 4.78%.
The core driving forces come from two directions: first, the expectation of the US-Iran agreement to drive oil prices to plummet, and Asia, as a major energy import region, is particularly sensitive to price changes, with falling energy costs directly boosting risk appetite; Second, U.S. stocks surged overnight and spilled over into the Asia-Pacific market. Palantir's 29% surge proves that a turning point for AI commercialization is approaching.
Hot Focus: Palantir ignites AI sentiment, AMD pours cold water on it
Palantir's earnings exceeded expectations across the board, with Q2 revenue of $1.94 billion. The full-year revenue guidance was sharply raised to $8.15–8.16 billion, with a 29% surge in after-hours trading—the largest single-day gain since February 2024.
The Philadelphia Semiconductor Index surged over 7%, Coherent and Mywell Technologies rose over 12%, Intel and SanDisk rose over 10%, and Micron Technology and AMD gained over 7%.
But today, the pre-market drama took a turn of events. AMD's Q2 revenue was $11.54 billion, up 50% year-on-year, and data center revenue was $6.7 billion, up 107% year-on-year. However, Q3 revenue guidance was about $13 billion (analyst expectation $12.52 billion), which exceeded expectations but did not meet more aggressive investor expectations, with the stock dropping more than 8% before market opening.
SpaceX's first earnings report showed revenue of $7.814 billion, capital expenditures soared to $18 billion, down more than 10% in pre-market trading.
The market's pricing logic for AI stocks has shifted from "whether they can make money" to "how much they can earn." Palantir has proven the explosive power of AI commercialization, but the high investments by AMD and SpaceX have made the market worry about the pace of return on investment.
US stock market pre-open: futures edged higher, semiconductor stocks under pressure before the market
Futures for the three major indices edged higher, with Dow futures up 0.29%, Nasdaq futures up 0.20%, and S&P 500 futures up 0.39%.
The memory chip sector showed mixed results in pre-market trading: Micron fell 1.71%, SK Hynix dropped 2.6%, and SanDisk declined 0.55%. Nvidia rose 2%, Apple and Google rose 0.6%.
Tonight's focus:
(1) 20:15 Beijing time, US July ADP employment data (expected 68,000)
(2) At 21:45 Beijing time, the US July Services PMI
(3) At 22:00 Beijing time, the US July ISM Non-Manufacturing PMI
(4) SanDisk and Western Digital Tonight's Earnings Reports (Memory Chip Sentiment Barometer)
Observation point tomorrow
1. Will AMD's after-hours drop drag down the semiconductor sector tonight, thereby affecting BTC risk appetite?
2. Whether ADP and non-manufacturing PMI data strengthen expectations for rate cuts;
3. After today's sharp rise in South Korea's KOSPI, can the rebound continue?
"Which side do you choose to stand on today?" 》
BTC is repeatedly oscillating above 64,000; what do you think:
A: Riding on US stock gains + expectations of the US-Iran agreement, this week it broke through 65,000
B: Tech stock sentiment was dragged down by AMD's sharp drop, with BTC pushing back to 63,000
C: Continue to fluctuate in the 63-64K range, waiting for the non-farm payrolls
I'll vote for A first. The rallies in Palantir and tech stocks provide indirect sentiment support for BTC, and if the U.S.-Iran agreement materializes, it will directly improve inflation expectations.
But AMD's earnings report also reminds us that the threshold for "exceeding expectations" in the AI sector is continuously rising.
BTC's rebound potential is limited, so don't blindly chase highs.
Which one would you choose? See you in the comments.MARA transferred 6,000 BTC, and asset management by mining companies is becoming a new market focus
On August 5, according to on-chain monitoring data, Bitcoin mining company MARA Holdings transferred 6,000 BTC to the Two Prime address in the past five hours, valued at approximately $384.6 million at current prices. Data shows that MARA previously held about 36,300 BTC, valued at approximately $2.34 billion, with this transfer accounting for approximately 16.5% of its Bitcoin reserves.
However, it should be noted that on-chain transfers do not necessarily mean a sale. This BTC flow to Two Prime is more likely related to asset management, institutional custody, or adjustments in capital strategies. The market's real focus is on how large mining companies are managing their Bitcoin assets.
Judging from the BTC/USDT perpetual hourly trend, the current market has not shown obvious panic due to this large transfer.
Currently, BTC is priced around $64,070, with a 24-hour high of $64,534 and a pullback of $63,581. Intraday fluctuations are about $950, with an overall range of about 1.5%. Previously, BTC pulled back from around $65,390, dropping to a low of around $62,227, then quickly rebounded and is now consolidating near $64,000.
From a technical perspective, BTC is in a short-term phase of oscillating recovery.
Hourly moving average data shows:
MA5 is about $64,078, MA10 is about $64,137, and MA20 is about $64,160.
The three moving averages are currently close to each other, indicating the market is waiting for a new direction. Prices are currently trading around the $64,000 area, and a breakout above the $64,500–$65,000 range is needed to further open up space; If pressure persists, attention should be paid to support near $63,500, and further downwards, the previous low around $62,200 should be watched.
In terms of trading volume, the current hourly volume is about $7.19 million, a significant decrease compared to previous rapid volatility phases, indicating that the market has not experienced concentrated selling pressure due to MARA transfer news. Large on-chain transfers are more like institutional funds being mobilized, rather than simple market sell signals.
In fact, from recent trends, the way mining companies and institutions hold BTC is changing.
In the past, mining companies mainly relied on selling mined BTC to sustain operations, but as the value of Bitcoin assets has increased, more and more companies are beginning to include BTC as part of their corporate balance sheets. MARA holds over 36,000 BTC, essentially approaching the level of a major Bitcoin reserve institution.
But holding large amounts of BTC also brings new problems.
When a company holds tens of thousands of BTC, how to improve capital efficiency, reduce risk, and manage assets becomes a new challenge. Transferring to institutional service providers like Two Prime may indicate that mining companies are exploring more ways to financialize BTC, rather than simply holding them long-term.
For the crypto market, this is also a sign of the maturity of the Bitcoin ecosystem.
The market is no longer just focused on "who is buying BTC," but has also shifted to "how BTC holders manage BTC." From ETF inflows, to corporate reserves, and then to mining enterprise asset strategies, Bitcoin is gradually evolving from a trading asset to an institutional-grade asset.
In the short term, BTC remains oscillating around $64,000, with the market waiting for more macro and capital signals to confirm its direction. This MARA transfer is more of an asset management move, but it also reminds the market that every move of large-scale BTC holders is becoming a key variable influencing market sentiment.
This is for personal market observation only and does not constitute investment advice. DYOR.
$BTC Terra peaked at 43% LUNA stake at something like 15% APR & 10,000 airdrop and liquidity mining schemes layered on top of staking while being the hottest ponzu in crypto...also having in-protocol delegation and in-protocol LST so less risk...also being in a low-interest-rate QE/covid subsidy macro environment that we'll probably never get again in our lifetimes...
there is basically no reason to think ETH staking will ever get to 50% no less 100%, just a bunch of hysterical bullshit....$ETH
#EarningsRealityCheck 🔥 WHEN MONEY IS ON THE TABLE, DECENTRALIZATION BECOMES A FAIRY TALE.
Ethereum's latest proposal (EIP-8363) proves one thing:
Everyone supports "what's good for Ethereum"... until it hurts their own yield.
The proposal aims to gradually reduce validator issuance, pushing ETH toward net-zero issuance as staking reaches 50% of supply.
Sounds bullish.
Less inflation.
More scarcity.
Better tokenomics.
Then why did the community explode?
Because someone has to pay the bill.
Validators would see annual rewards fall from 2.86% to 1.48%.
Liquid staking protocols.
Institutional staking products.
DeFi lending strategies built on staking yield.
All suddenly become less attractive.
This is where crypto stops being ideology and starts becoming economics.
People don't fight because they hate Ethereum.
People fight because their cash flow is under attack.
Every tokenomics change creates winners...
...and losers.
The moment incentives shift, decentralization instantly splits into factions.
In crypto, governance isn't about consensus.
It's about discovering who loses money.
💬 If EIP-8363 really makes ETH scarcer but cuts validator income in half... would you still vote YES?🤗 Breaking news: Circle has announced the list of founding validators for the Arc blockchain—11 financial giants including BlackRock, Visa, Mastercard, DTCC, and others (including Galaxy, Standard Chartered, SBI, ICE, MoneyGram, Sumitomo, Global Payments).
Arc is Circle's L1 for institutional development. It launched on the public mainnet on September 16, natively embedding USDC, and runs stablecoin + tokenized asset settlement.
Is this good news for cryptocurrencies?
In the long run, yes. In the past, institutions talked about on-chain integration, but their feet were still outside—because they feared the SEC would crack down on securities and worry about which nodes were running smoothly.
Now we just have to wait for the CLARITY Act to pass. Once it passes, the drama of BlackRock running nodes and Visa handling liquidations will go from news to everyday life.
But don't get carried away: Arc isn't for retail investors—it's for tokenizing assets managed by BUIDL funds and DTCC. The positive news is that crypto is entering Wall Street at the bottom of its umbrella, not that it will explode tomorrow $BTC $ETH $SOL.
BlackRock is playing both the role of father and mother 😂. Do you think this is a victory for crypto, or has Wall Street taken over the crypto world?SanDisk surged 10.84%, hitting a new stage high. Why is Bitcoin still stuck in the narrow 63,800 range?
On the evening of August 4, the storage sector experienced an epic surge, with SanDisk closing at $1,427.62, a single-day increase of 10.84%, and a total turnover of $9.68 billion for the day.
SK-SK Hynix rose 8.17%, Micron Technology gained 7.62%, and the Philadelphia Semiconductor Index surged 6.5%, closing stronger for four consecutive trading days.
In contrast, Bitcoin was locked in the $63,350-63,817 range throughout the day, with a 24-hour overall fluctuation of only 0.76%, showing almost no volatility.
In July, single-layer NAND flash spot prices surged 35% month-on-month, and storage manufacturers' 2027 production capacity has been fully booked by AI server customers, providing hardware stocks with tangible supply-demand dividends.
Bitcoin spot ETFs saw $187 million in capital flight in a single day, as institutional funds flocked to the recovering storage sector, unwilling to touch crypto assets with higher regulatory risks.
Once the current wave of spot spot deposit price hikes cools down, will there be a chance for institutional funds to flow into the crypto world? #财报观察员: Mixed results, lifting restrictions imminent! What do you think about SpaceX's future? #SpaceX首份财报超预期, unlocking remains a key variable #AMD财报超预期 has growth been overdrawn? Today's financial report from Fangwai
Today is August 5th, and just like yesterday, it is one of the most densely packed days for US stock earnings reports.
Before the market opened, we focused on consumption, pharmaceuticals, and platform economy; after the close, storage, AI applications, and high-valuation growth stocks concentrated in the market.
Based on Beijing time, US pre-market earnings reports are mainly released tonight from 18:00 to 21:00; Post-market earnings reports will be concentrated tomorrow from 4:00 to 4:30 a.m.
Tonight's pre-market open: SHOP, UBER, DIS, CRCL
$SHOP
Shopify
The market expects revenue of about $3.43 billion, up about 28% year-over-year, with adjusted earnings per share of about $0.40.
Three main factors to look at: GMV growth, payment business penetration, and revenue growth guidance for next quarter.
SHOP is a platform for e-commerce payments, advertising, and merchant infrastructure, so growth is more important than EPS alone.
$UBER
Uber
Market expectations for revenue of about $14.21 billion and adjusted earnings per share of about $0.83. The company previously set its Q2 total booking guidance at $56.25 billion to $57.75 billion.
The focus is on travel order growth, food delivery profit margins, free cash flow, and autonomous driving cooperation.
Uber's biggest valuation logic now is how much more profit margins can improve.
$DIS
Disney
I went twice just last year—the first time on the day before 1011, and the second on Christmas
There are really a lot of people at Christmas.
Market expectations for revenue of about $25.41 billion and adjusted earnings per share of about $1.88.
The market will focus on the park business, Disney+ profitability, and the long-term strategy of the new management.
Analysts expect the parks and experiences business to remain the core of profit, but changes in oil prices, inflation, and travel demand may affect visitor flow.
$CRCL
Circle
My take: buy at the start of five, increase the buy at the start of four
The market expects revenue to be around $710 million–$745 million, with earnings per share of about $0.16–$0.18.
The most important thing for CRCL is not traditional EPS, but USDC circulating scale, reserve interest income, distribution costs, and how management responds to payment giants like Visa, Mastercard, and Stripe entering the stablecoin market. It directly affects the overall stablecoin concept and the sentiment of certain crypto stocks.
Other pre-market noteworthy items include: $CVS, $KHC, $GFS, $GLXY, $RIOT, $WULF, and $SEDG.
Tomorrow morning after-hours: The main battleground is
$SNDK
$SNDK
SanDisk
This is the most important financial report of the day.
Market expectations vary among institutions, with revenue estimated at approximately $8.3 billion to $8.7 billion, and adjusted earnings per share around $33 to $35.5. Last quarter, the company's official guidance was only $7.75 billion to $8.25 billion and earnings per share of $30 to $33.
In other words, the market has already priced in "exceeding guidance" in its expectations.
This time, besides revenue and EPS, we also need to look at:
Can NAND and enterprise-grade SSD prices continue to rise?
Data center revenue and shipments
Can long-term supply contracts truly reduce storage cycle fluctuations?
Next quarter gross margin and revenue guidance
Is management still optimistic about AI storage demand?
SanDisk signed several long-term supply contracts last quarter, three of which are worth approximately $42 billion.
This indicates that the company is trying to transform the traditional storage "skyrocket cycle" into more stable contract-based revenue.
But the risk is that SNDK has already surged this year. Although it has recently pulled back significantly from its peak, the market's demand for profit growth remains very high.
Even if the earnings report exceeds expectations, as long as next quarter's guidance is not explosive enough, AMD and SPCX may reappear.
And tonight
$WDC
Western Digital will also release its earnings report at the same time, with market expectations of revenue of about $3.7 billion and earnings per share of about $3.32–3.35.
Jointly handing over documents between two companies is equivalent to cross-validating AI storage demands, enterprise-grade SSDs, and the entire data storage cycle.
$APP
AppLovin
Market expectations for revenue of about $1.94 billion and earnings per share of about $3.72.
Focus on the AI advertising engine AXON, client growth, and e-commerce advertising business. The app's profit margins are already extremely high, so the market is more concerned about whether new businesses can continue to expand, rather than just cost control.
$DASH
DoorDash
The market expects revenue of about $4.32 billion and earnings per share of about $0.50, with year-on-year growth expected to exceed 30%.
Mainly focusing on order growth rate, international markets, grocery delivery, and profit margins. Compared to Uber, you can judge whether local U.S. consumption and demand for instant delivery are strong.
AI and high-valuation growth stocks
There are still many high-volatility stocks after tonight's close:
$IONQ: Expected revenue is approximately $66.36 million, with a loss of $0.29 per share
$SOUN: Expected revenue is approximately $52.49 million, with a loss of $0.05 per share
$DUOL: Expected revenue is approximately $297 million, with earnings per share of $0.61
$FIG: Important post-IPO financial report, expected revenue of approximately $351 million
$HUBS, $AXON, $SMR, $JOBY will also release results
In addition, there are $MELI, $XYZ, $OXY, $EBAY, $EXPE, $ELF, and $MCK.
In summary
Tonight's pre-market session is evaluating US consumption and the platform economy: LLY looks at pharmaceutical competition, UBER and DASH look at local consumption, SHOP and CRCL look at platform growth.
But what truly affects the sentiment of tech stocks is SNDK + WDC + APP.
Especially SNDK.
Yesterday, AMD and SPCX told us that the current market is not short of good numbers; what is lacking is a future that can continue to be revised upward.
If SNDK's earnings report is strong but the guidance only meets expectations, the storage sector is unlikely to buy in
Only if price, gross margin, data center demand, and next quarter guidance all exceed expectations can the overall sentiment in the storage sector be boosted. #EarningsObserver: Mixed Results, Unlocking Imminent! What do you think about SpaceX's future? $ETH $BTC $BICO BEAT Market Analysis
Current Price: 2.4226
Overall Pattern: After an initial surge to 11.6576 followed by a long-term continuous decline, there was a recent rebound near 4 which faced resistance again, initiating a new round of decline. The long-term cycle is clearly within a downward channel, dominated by a bearish trend, with short-term slight low-level oscillations and weak rebound strength.
Structure Across Timeframes
1. Daily Level
Price continues to move along the downtrend, with resistance at 7.6509 above, difficult to reach in the short term; key support at 2.2109 below, which is also the recent low defense line. Daily indicators remain bearish; if support breaks, the downside space will further open.
2. 4-Hour Level
All moving averages diverge downward, clearly showing a bearish alignment. The first resistance above is in the 2.68~2.70 range, with support near 2.20 below. Price is under continuous downward pressure; rebounds are considered mere corrections during the decline and should not be viewed as bullish unless resistance is broken.
Trading Ideas Reference
✅ Bullish Idea
Only consider light position trades if price stops falling and stabilizes at the 2.20~2.21 support zone, aiming to capitalize on oversold rebounds. Stop loss should be set below 2.18, with a short-term target near 2.65.
❌ Bearish Idea
If rebound faces resistance and fails to break through the 2.65~2.69 range, consider shorting with the first target at 2.21;
If the 2.20 support is effectively broken, a new round of decline will begin, expanding the downside space further.
Content is for technical communication only and does not constitute investment advice$BEAT I am Cige. AMD's earnings exceeded expectations, but the stock fell 8% after hours. An old script is playing out again: good performance does not necessarily mean the stock price will rise.
AMD's earnings figures
Revenue was $11.536 billion, up 50% year-over-year, with adjusted EPS of $1.66, both higher than market expectations. Data center business revenue was $6.7 billion, up 107% year-over-year, accounting for 58% of the company's total revenue, with a gross margin maintained at 56%. The Q3 revenue guidance is about $13 billion, higher than the market average expectation but below some institutions' higher forecasts. After the earnings release, the stock fell more than 8% after hours.
Why did it fall despite exceeding expectations?
Market expectations had been pulled up too high in advance, and the guidance did not further surpass them, so the market was already pricing in higher expectations. When the "exceeding expectations" margin is not large enough, it is interpreted as "not good enough." AMD's problem is that high growth has already been priced in.
Impact on SanDisk
AMD's 8% drop after hours directly pressures SanDisk's earnings tonight. The storage sector has seen huge recent gains; SanDisk rose from 998 to 1468, with expectations fully priced in. AMD's movement is a signal: if the earnings just meet expectations rather than greatly exceed them, profit-taking will be very intense.
How to handle the 1331.77 short position
Set the stop loss above ？ #EarningsRealityCheck #SpaceXBeatEstimates #AMDBeatsButDrops #标普500首次站上7700点, setting a new all-time high
The S&P 500 officially broke above 7,700 last night, closing at 7,736.52, up 1.79%. The Dow also broke below 54,000, up more than 900 points. The Philadelphia Semiconductor Index was even stronger, rising 6.55% in a single day. The total market capitalization of the S&P 500 also surpassed $70 trillion for the first time.
On the surface, it looks like a broad rise, but what truly drives this big bullish candlestick is the simultaneous occurrence of three things.
First, the earnings season ended beautifully. Palantir's earnings exceeded expectations, rising 13% in after-hours trading. Kaituo Heavy Industries' earnings also exceeded expectations. Microsoft and Amazon already proved last week that spending money on AI yields returns. The overall narrative of this earnings season shifted from an "AI bubble" to "AI is delivering revenue."
Second, oil prices have crashed. Bescent said the US and Iran might reach an agreement on Tuesday or Wednesday to restart the Strait of Hormuz, causing oil prices to plunge in a single day. When oil prices fall, inflation expectations cool down, and long-term bond yields also fall. Simultaneous easing of inflation and interest rates is a double benefit for risk assets.
Third, semiconductors and storage surged across the board. Philadelphia Semiconductor rose 6.55%, Micron rose 7.6%. Intel and SanDisk surged over 10%. Memory stocks fell 40% in July; this rebound is the result of oversold recovery and sentiment reversal.
But the problem lies — while the S&P hit new highs, AMD and SpaceX both plunged in after-hours trading. AMD fell nearly 9% in after-hours trading, earnings beating expectations but lacking impressive guidance. SpaceX's first earnings report exceeded revenue expectations, but AI spending surged, causing it to fall more than 7% in after-hours trading. The overall market is rising, but individual stocks are diverging. Funds are pouring into giants that have already proven themselves, while abandoning any "not good enough" stocks.
Bespoke's Hickey said the Nasdaq has risen more than 1% for four consecutive days, and this kind of continuous buying usually indicates genuine capital inflow, not a short-term rebound. But the question is, how long this round of inflow can last depends on whether the US-Iran agreement can actually materialize, whether oil prices can hold steady, and whether next week's CPI data will give some respect. The market is indeed hitting new highs, but I don't think this is a blindly all-in signal. The S&P has rebounded from its June low to now, has risen a lot, and its valuation remains high. Any small movement could trigger profit-taking.
At this level, I prefer to keep positions flexible, not chasing highs, and adding more on pullbacks. The significance of new highs is not chasing them, but confirming the trend. After trend confirmation, there are plenty of opportunities to get in on board.📢 Market News | After SPCX released its latest earnings report, its stock price saw a noticeable pullback, and market attention quickly shifted to the upcoming large-scale lock-up. Many people are asking: With record-breaking performance, why has the stock price fallen? 🤔 The answer is actually quite simple—the market trades the future, not the past. 📊 The latest data shows that the company's revenue continues to grow rapidly, with AI-related businesses remaining the biggest highlight. Core business revenue keeps rising, with multiple indicators exceeding market expectations. However, what truly worries investors is that high-intensity AI capital investment will continue in the coming quarters, meaning short-term profit margins will remain under pressure. 💰⚠️ 🔥 The bigger highlight is not the financial report, but the upcoming unlocking event. A large number of restricted shares entered the circulating market, raising concerns among investors about increased selling pressure, which significantly cooled market sentiment. Many short-term funds chose to exit early, putting pressure on the stock price. 📉 On the other hand, short positions in the market remain at a high level. When the market is almost unanimously bearish, the real situation to be wary of is another scenario—if there is no panic selling on the day of the lock-up, short covering could quickly push the stock price higher, forming a short squeeze. 🚀 🎯 Current market focus: ✅ Future profitability of AI business ✅, whether actual selling pressure after unlocking is lower than expected ✅, whether institutional funds continue to increase holdings ✅, and whether bears are starting to concentrate their holdings ⚡. The news has been released, but what truly determines the trend will be the choice of funds. The market is always like this: 📉 negative news is unevenly realized$BEAT 币的“吹牛”经济学
当“吹牛”被写进代币经济模型，故事与泡沫之间只隔着一层流动性的窗户纸。BEAT币，这个2026年夏天月涨超1400%的加密项目，正是这样的观察样本。
三条故事线，撑起一个“牛”
BEAT的叙事逻辑堪称加密项目“吹牛”的标准范本：一是绑定经典音乐游戏《劲舞团》6亿注册用户的IP故事；二是包装“AI智能体拥有独立钱包、与人类平等参与链上经济”的AI叙事；三是打出“每周收入几乎100%回购销毁代币”的通缩牌。三条线相互嵌套，把币价从不到1美元一路“吹”到11美元的历史高点。
光鲜叙事下的三重暗流
然而链上数据揭示了另一面：第一，大部分供应量集中在少数钱包手中，筹码极度集中，暴涨更像精密操纵而非市场行为；第二，价格创历史新高时，链上活跃地址数反而下滑，真实用户并未同步增长，这轮上涨主要由杠杆衍生品驱动；第三，“发布路线图→拉盘吸引FOMO→庄家出货→散户站岗”的剧本反复上演，利好发布即暴跌，三天跌去60%。
谁在为故事买单？
对于项目方和早期参与者，“吹牛”是稳赚的生意——高位派发筹码给追涨的散户。而对于散户，这是一场信息不对称的赌局：他们听到的是“6亿用户”“AI平权”“每周百万销毁”，却忽略了更关键的事实——完全稀释估值一度高达90亿美元，而流通量仅为总量的三分之一，数亿枚代币还在等待解锁抛向市场。
一位社区评论者说得直白：“BEAT现在不是在讲基本面，而是在考验市场的贪婪程度。”当叙事成为资产，真相便不再重要。重要的是，在鼓声停止之前，谁在“吹牛”的游戏中占据主动，谁又成了那个为故事买单的人
$BTC #SpaceX首份财报超预期，解禁仍是关键变量 $BTC $ETH $SNDK The computing power and storage sector attracted $32 billion in a single day, making a huge splash. When will the crypto world see new growth and vitality?
On August 4, the total single-day trading volume of the US storage sector exceeded $32 billion, with global short-term speculative funds flocking to storage leaders like SanDisk and Micron, competing for the recovery of the flash memory cycle.
SanDisk closed at $1,407.66, up 9.02% in a single day, with a price-to-earnings ratio (TTM) of 46.25, fully unlocking the profit-making effect in the storage sector.
After all incremental risk funds were absorbed by the US hardware sector, the crypto market lost its off-exchange lively supply.
Bitcoin spot ETF funds experienced sharp fluctuations, with a single-day outflow of $265 million, and Ethereum ETFs seeing capital outflows for several consecutive days.
US hardware stocks have solid confidence due to AI computing power demand and spot flash memory price increases; Bitcoin can only passively wait for ETF funds to flow back and the positive news of U.S. crypto regulation being implemented.
After prolonged box swings worn down patience for positions, when will the crypto world welcome the long-awaited arrival of off-exchange incremental funds? #财报观察员: Mixed results, lifting restrictions imminent! What do you think about SpaceX's future? #SpaceX首份财报超预期, unlocking remains a key variable #AMD财报超预期 has growth been overdrawn? #闪迪财报前夕, HBF and storage shortages have sparked heated discussion
SanDisk's Financial Report: HBF Technology Nuclear Bomb and Storage Shortage, the "Key Battle" for AI Storage Leader
In the early hours of August 6 Beijing time (after the US East Market closed on August 5), SanDisk (SNDK) will release its financial report for the fourth quarter of fiscal year 2026. Prior to this, the storage market had already been ignited by two major topics: the release of HBF's first standard specification and the full exhaustion of storage capacity by 2027. After plunging 47% from the June high of $2,354 to $998, the stock price rebounded violently over 40% to $1,427. Is this financial report "rocket fuel" or a "guillotine"?
1. HBF: The "New Species" of AI Storage
The biggest technological catalyst on the eve of the earnings report was the official release of the first high-bandwidth flash memory (HBF) standard specification on August 4 by SK Hynix and SanDisk at the FMS 2026 Summit.
HBF is positioned as a new storage layer between HBM and SSD. HBM is fast but has limited capacity and is costly, while SSD has large capacity but is slow. HBF combines the advantages of both—up to 512GB capacity, maximum 3TB/s transfer speed, and uses the UCIe open interconnection standard.
This specification was officially released by OCP, the world's largest open data center technology organization, and has become a universal open industry standard. Google and Tenstorrent have joined the HBF alliance. SanDisk's CTO previously stated: "The global AI race depends on memory, not computing power." HBF chip samples will be available by the end of this year, and the full product is expected to launch next year.
On the same day, SanDisk and Kioxia also released the BiCS10 QLC 3D NAND—332-layer stacking, with 60% higher density than the previous generation.
2. Storage Shortage: All Production Capacity Will Be Sold Out by 2027
Even crazier than HBF is the level of supply tension. The three major manufacturers—Samsung, SK Hynix, and Micron—have fully allocated DRAM and HBM production capacities for 2027. NAND is also tightening—Samsung, Micron, and SanDisk have already sold out their full-year 2027 production capacity, and Kioxia and SK Hynix are expected to complete allocation by the end of August at the latest.
Downstream customers have gone almost crazy to compete for capacity—some even "humble themselves to beg for supplies," with the original factory offering only 60 to 70% of the customer's expected capacity. SK Hynix's CEO previously stated: "2027 will be the tightest supply year in the history of the storage industry, and supply shortages may continue beyond 2030." TrendForce predicts that by 2026, the NAND Flash market will experience a 4% to 5% supply gap, with shortages likely to continue at least until the first half of 2027.
3. Earnings Eve: A showdown between bulls and bears
Bullish Cards:
· Q3 data center business revenue was $1.47 billion, up 233% quarter-on-quarter and 645% year-on-year
· Five multi-year supply agreements have been locked, totaling at least $42 billion
· Zero debt + $6 billion buyback plan
· The average target price of 21 analysts is $2,095, representing a 47% upside from the current price
Short Chips:
· In July, the stock price plunged 47%, with extremely unstable holdings
· Market expectations were fully met: revenue reached $8.395 billion (+341% year-on-year), and net profit surged more than 229.6 times year-on-year
· Before earnings reports, storage stocks fell broadly in pre-market trading—SanDisk fell 0.43%-1%, SK Hynix dropped over 2%
The core market disagreement is: Is SanDisk a "cyclical stock" or a "growth stock"? If it's a cyclical stock, a PE of about 6.5 times now might be reasonable; but if it's a growth stock with annual growth over 300%, gross margin of 80%, and zero debt—giving a 6.5 PE is "massacre."
4. Summary
SanDisk stands at a historic crossroads. HBF technology has opened up a new frontier for AI storage; All capacity sold out by 2027 locks in performance visibility for at least the next 12-18 months. The key to tonight's earnings report is not "whether it meets expectations," but how long the high prosperity of AI storage can continue. If the earnings report only "meets expectations" rather than "significantly beats expectations," this AI storage leader, which soared from $28 to $1,427, could repeat the July crash. If the data fully crushes — revenue hits $8.5 billion, EPS breaks through $35, gross margin stays around 81%, and the 2027 fiscal year guidance continues to be raised, that would be the strongest counterattack to the "cycle peak" theory.
$SNDK $UNI rose to around $3.9 today, but policy has just added a layer of pressure.
The probability that the U.S. Clear Act will pass before adjournment is only about 27%. This bill will determine under what conditions DeFi protocols like Uniswap qualify as truly decentralized, and will also determine whether the SEC or CFTC will govern the future.
Prices can still rise now, mostly due to market recovery; policy incentives have not materialized. The bill is delayed until September, and UNI can still rebound in the short term, but the upside space is easily suppressed by regulatory news.
I'll see if $4 can hold steadily; if not, first guard around 3.7 yuan. What truly changes valuations is the bill re-entering the vote.
Personal analysis and does not constitute investment advice.📌 [Market Summary] BTC is moving sideways while waiting for the wind to arrive, but macro positive news cannot drive it
On August 5, BTC was still consolidating near 64,300, with a 24-hour gain of less than 1%, and had basically remained unchanged over the past week. Global stock markets have repeatedly hit new highs, yet the crypto world has chosen to hold steady.
ETH is at 1,864, down about 2% during the week, making it one of the few mainstream coins to close lower.
The Panic and Greed Index is 27, still in the "panic" range, but has slightly warmed up compared to yesterday's "extreme panic" (25).
---
🧐 What is the market waiting for?
The US-Iran Strait of Hormuz Agreement may have a chance to be implemented today. Oil prices fall, gold is rising, Treasuries are climbing—the macro is almost entirely positive, but BTC just won't keep up. This happened for three consecutive days, indicating the problem was not external but internal.
USDT liquidity continues to shrink: market cap has decreased by about $4 billion in the past 60 days, with another $870 million shrinking in the past 11 days. With fewer stablecoins, buying interest naturally weakens.
---
📌 A noteworthy signal
Well-known KOL Ansem pointed out that the market is showing signs of a bottom—prices are beginning to show resilience to bearish news. Institutions are slowly acquiring goods, while retail investors remain on the look-and-see stance.
---
📊 Summary
The stock market is rising, gold is rising, and Bitcoin is at 64,000 for playing its own way.
Without increased volume, the pattern will remain volatile. Wait for the wind to come, don't chase the rise.
⚠️ The above is a market observation and opinion sharing and does not constitute investment advice. The market carries risks, and decisions must be made cautiously.
#BTC #ETH #DOGE #市场观察 #非农数据 #64,000 is volatileAs high-risk assets, why do U.S. storage stocks crush the currently sluggish cryptocurrencies?
SanDisk rose 9.02% in a single day, Micron rose 7.37%, and storage chip stocks saw explosive gains; Bitcoin rose only 0.98% in 24 hours, while Ethereum rose 0.57%, indicating a sluggish crypto market.
Both are classified as high-risk growth assets, but their recent returns have been vastly different.
The fundamental gap is fundamental support. In June, global memory chip sales reached $74.6 billion, up 31.7% month-on-month. Flash spot prices have risen, AI giants are frantically purchasing HBM memory—all these positive factors can be supported by industry data.
Bitcoin lacks revenue, physical projects, or orders as support; price fluctuations depend solely on institutional ETF fund flows and market sentiment.
Institutions prioritize hardware stocks with measurable returns when allocating risk assets, naturally abandoning the crypto track full of policy variables.$BTC $ETH $SNDK In the past, stocks and cryptocurrencies rose and fell together, but now the Nasdaq is up 2.59%. The big cake remains unmoved—who caused the market disconnect?
Last year, whenever the Nasdaq index rose more than 1%, Bitcoin was very likely to rebound simultaneously, with both types of risk assets linked to rise and fall.
On August 4, the Nasdaq surged 2.59%, hitting a new all-time high, with all three major indices closing at new highs. The Dow Jones held above 54,085 points, and the S&P 500 climbed above 7,736 points.
Bitcoin rose only 0.98% slightly throughout the day, with a narrow range of volatility, completely unaffected by the rally in the US stock market.
US stocks are experiencing a cyclical bull market driven by AI storage hardware, with financial reports from various storage manufacturers multiplying in numbers; The crypto world has been repeatedly hit by ETF fund outflows and negative crypto regulatory pressures.
US stocks rely on the supply and demand cycle of the real industry, while the crypto world is trapped in a game of existing contracts. Going forward, we need to view the rise and fall cycles of the two markets separately. #财报观察员: Mixed results, lifting restrictions imminent! What do you think about SpaceX's future? #SpaceX首份财报超预期, unlocking remains a key variable #AMD财报超预期 has growth been overdrawn? $BTC $ETH $SNDK The US stock sector switched in succession, with the Nasdaq and semiconductors surging consecutively, and the crypto world failed to capitalize on the trend
Recently, the rotation of the US stock market has been rapid: first the AI software sector surged, then computing chips strengthened. On August 4, memory chips surged again, with the Philadelphia Semiconductor Index rising 6.5% and ARM surging 17.36% in a single day, continuously producing stocks with big gains.
#财报观察员: Mixed results, the lifting of restrictions is approaching! What do you think about SpaceX's future?
After the rotation of various US stock tracks, profit-driven capital remains in the US market, preparing for the next hardware boom.
#SpaceX首份财报超预期, unlocking remains a key variable
Even though the storage sector's single-day turnover exceeded 30 billion, there was no capital diverting into the crypto market.
#AMD财报超预期 has growth been overdrawn?
Ethereum ETFs have seen net capital outflows for several consecutive days. On August 4, $12.3 million was a single-day outflow, and the altcoin market fell 0.62% overall in 24 hours.
Regulatory risks, lack of revenue fundamentals, and unstable ETF funds are trapping the crypto sector's upward potential. What kind of major positive factors could guide U.S. stock market trends into the crypto world?Institutions invested 211 million yuan heavily in storage stocks—who else would be willing to enter the crypto market and ambush the crypto world?
On August 4, the total single-day turnover in the US storage sector exceeded $32 billion, with a large amount of institutional funds entering the market to increase positions in leading stocks like SanDisk and Micron Technology. The storage sector's single-day fundraising scale hit a one-month high.
Nvidia maintained a 2.56% increase, and AI computing hardware remains the base allocation direction for long-term institutions.
After all funds concentrated in the US hardware sector, the crypto sector's incremental vitality was nearly depleted, with Bitcoin spot ETF funds fluctuating back and forth, with a single-day peak of $265 million fleeing.
Every time Bitcoin rebounds slightly, it triggers institutional share reductions, leaving only retail investors with their existing funds engaged in short-term trading.
Once the short-term hype around memory chips fades, will the hot money that left the hardware sector be willing to return to the crypto sector?#临时通航协议待落地，油价风险尚未反转 美伊借阿曼斡旋推进霍尔木兹海峡60天临时通航协议谈判，缓和预期带动原油大幅跳水，WTI单日跌幅近6%，市场集中出清此前地缘冲突推升的风险溢价，但多重变量决定油价下行趋势并未确立，多头风险依旧潜伏。
谈判存在根本性分歧：美方宣称短期内即可官宣协议，伊朗却否认与美国直接对话，双方在海峡管控权、通行规则、海上封锁解除条件上难以达成统一，即便短期签署临时协议，也仅为60天短期安排，无法彻底化解长期地缘矛盾。同时红海商船袭扰事件未完全平息，中东局部冲突隐患仍在。
供需端也持续给油价提供底部支撑，当前处于全球原油夏季需求旺季，OPEC+维持减产策略，海湾原油出口量较战前仍有显著缺口，全球原油库存处于近年低位，单纯情绪降温难以扭转供给偏紧格局。
本轮下跌仅为预期驱动的阶段性回调，并非趋势反转。若谈判破裂、美伊海上摩擦再度升级，油价将快速修复溢价；即便协议落地，市场重心也会切换至非农就业、美联储政策与OPEC产量指引，原油宽幅震荡行情延续，不宜盲目追空。$BTC $ETH $SNDK $BTC ChainCatcher reports, according to Jin Shi, French President Macron stated that the EU and its partners will continue to increase pressure on Russia and continue to advance various sanctions.
This message needs to be viewed in layers, distinguishing between verbal statements and concrete actions.
At this stage, it is only a stance statement; the new round of sanctions lists and specific restrictions in energy, finance, and crypto sectors have not yet been announced.
Looking back at historical patterns:
Simply emphasizing tough words is unlikely to sustain a sustained increase in geopolitical risk premiums in the short term; What truly shakes up the market is a new round of sanctions that can be implemented and implemented.
Two conduction pathways closely tracked
1. Energy side
If subsequent sanctions tighten Russian oil trade, it could once again disrupt global crude oil supply expectations and push oil prices higher. Rising oil prices will reignite inflation concerns, indirectly limiting the Fed's room to cut rates and suppressing valuations of risk assets (BTC, US stocks).
Conversely, if only verbal deterrence is used, the energy market is very likely to continue its current volatile pattern.
2. Potential risks in the crypto sector
Recent rounds of EU sanctions have continuously targeted crypto trading channels that help Russia evade regulation. Once a new round of sanctions is implemented, it is highly likely that compliance constraints on overseas exchanges and cross-border crypto transfers will be expanded, causing short-term emotional shocks.
Thinking from a trading perspective
Currently, the market has just digested the positive news of easing US-Iran negotiations and falling oil prices, and local sentiment is at a low level.
Macron's remarks are a risk warning signal, but not enough to serve as a sole basis for opening positions.
Key points to watch next:
✅ Whether detailed sanctions drafts have been issued;
✅ Whether there is sustained linkage among crude oil, US Treasuries, and risk assets.
BTC is currently fluctuating within a range, waiting for direction, and geopolitical news only brings short-term impulse fluctuations.
Before substantive policies are implemented, the original range-bound strategy will still be followed, avoiding preemptive geopolitical black swan competition.
⚠️ Risk Warning: This content is for idea exchange only and does not constitute investment advice.
$BTC $ETH$SNDK $DOGE Any curve that brings staking yield to zero earlier than at 100% staked is essentially a centralization tool.
It is simple.
Reaching 100% staked is practically infeasible, but if it does, there is no point in paying for staking anymore. Hence, a curve that goes to zero at 100% staked could be at least considered reasonable.
What about any curve that zeroes out staking rewards for all stakers at N < 100%? Simple again!
What we will see is a survival game. If we assume there is demand to stake more than 50% of $ETH (as all issuance proposals do), then the staking ratio will get as close as possible to 50% or even reach 50%. And then we wait. We wait for the small players to die out or just quit staking. While institutions like BitMine will be able to afford zero staking rewards for long enough so that others will leave and staking rewards will become positive again.
Now ask yourself, is it what we really want?
As a person who devoted the last 5 years of my life to Ethereum decentralization improvement and strengthening, I take this proposal as a clear signal that its authors are against decentralized Ethereum!This is the kind of positioning I like to watch.
SolsticeFi Season 2 distributed a massive 185.1B flares.
At today's numbers, that works out to roughly:
175 $SLX per 1M flares
Around $13.4 in value per 1M flares.
But here's the interesting part...
One participant managed to accumulate 1.35% of the entire Season 2 airdrop with an estimated cost basis of just ~$8K by aggressively buying YT.
According to their math, the trade becomes profitable if the project reaches a fully diluted valuation above $20M.
Will it happen?
Nobody knows.
But I always pay attention when someone is willing to size up before the crowd arrives.
Sometimes that's where the best asymmetric bets are found. Storage chips are entering a super-cycle bull market, so why has the crypto market become a discarded asset for institutional funds?
UBS data shows that global total storage chip sales in June reached $74.6 billion, a month-on-month surge of 31.7%. Flash spot prices have risen for several consecutive months, marking the official start of a cyclical reversal in the industry.
On August 4, SanDisk surged 9.02%, with a price-to-earnings ratio (TTM) reaching 46.25. The market prematurely exhausted the earnings dividends, with Micron and SK Hynix both rising, making the storage sector the strongest main trend in the US stock market today.
In contrast, the crypto market as a whole is sluggish, with Ethereum ETFs seeing a single-day net outflow of $12.3 million, and more than 70% of altcoins falling within 24 hours.
Institutional funds prefer hardware stocks that quantify revenue and order data, while Bitcoin lacks the backing of real business revenue, coupled with the slow implementation of U.S. crypto regulations.
As long as policy uncertainty persists, large institutions will not enter the crypto space on a large scale. When will the crypto sector escape the predicament of capital abandonment?Saylor不是“永不卖币”吗？Strategy一周三卖BTC，累计甩出近3000枚！
曾经把“Never Sell Your Bitcoin”挂嘴边的 Michael Saylor，亲手把这句话拆成了两半——“我个人一聪不卖，但 Strategy 是上市公司，该卖就得卖。”
最新链上+SEC文件拼出来的事实：
• 7/27–8/2：卖 1,638 BTC，均价约 63,957 U，套现 ~1.05 亿 U（低于其 75,419 U 的平均成本，账面亏着卖）
• 紧随其后：再卖 ~300 BTC（~1,890 万 U）
• 8/5 凌晨：链上转出/卖出 1,030 BTC（~6,600 万 U）
• 一周合计：≈2,968 枚 BTC，离“近3000枚”就差临门一脚，当前持仓仍高达 842,138 BTC（占全网 4%）
钱去哪了？不是看空 BTC，是救资本结构：
• 一半付 STRC 优先股股息
• 一半回购折价的 STRC（面值100，市场曾跌到89）
• 顺手增发 301 万股 MSTR，募 2.9 亿 U 充美元储备到 40 亿 U
所以别再信“超级多头叛变”的标题党。真实剧本是：
BTC 溢价飞轮（mNAV>1 发股买币）转不动了 → 优先股利息压身 → 用 BTC 当流动性水龙头，保住融资壳子。
但市场情绪挡不住：
• 最大企业多头都亏本卖，短线散户慌的是“下一个卖家是谁”
• MSTR 年内跌超 40%，BTC 探到 62k 区间，STRC 不回面值，卖币就不会停
• 董事会已授权最高 50 亿 U 的 BTC 变现额度，现在才扔了零头
我的判断：
对 BTC 现货是情绪利空、实质抛压可控（3000枚 vs 842k持仓）；
对 MSTR 是“去神话化”定价，每股含币量增速从13.3%砸到3.5%；
对散户的提醒是——别把创始人的 meme 当公司资产负债表读。Italian major banks cut 90% of their $BTC ETFs and instead tripled their holdings by staking $ETH, and institutional funds began to "pick on the moves."
Italy's largest bank, Intesa Sanpaolo, stunned the market in its latest 13F filing: in the second quarter, it cut BlackRock's IBIT common shares from 646,800 to just 40,700, a 93.7% decrease, cut call options by 99.3%, and even opened 500,000 new IBIT bear positions. But on the other hand, it tripled its staked Ethereum ETF ETHB from 116,200 shares to 349,600 shares. On the surface, it looks like a bearish stance on Bitcoin, but it's not that simple—it still holds 3.47 million ARKB shares, worth over $67 million, which remains its largest crypto ETF position.
The real highlight of this isn't that banks don't favor BTC, but that institutions' criteria for selecting assets have changed. Spot Bitcoin ETFs hold coins that just sit idle and don't generate any profit; Staking ETH can earn an additional 3% to 4% protocol reward each year. For a bank vault that must report to the investment committee, an asset that "holds interest on" is clearly easier to tell a story than a pure gambling price. That's why it was cutting naked long positions while buying bearish protection—not because it's bearish on the industry, but because pure long positions offer too little cost-effectiveness.
Looking at the market, BTC is now near $64,000, up slightly intraday, but still hovering below the 200-day moving average, with support at $62,800 below and resistance at $64,300 and $65,000 above. The RSI is neutral to weak, typical of a top above and a bottom below. ETH is trading near $1860, stuck at the upper edge of the 1800-1820 support band, with hard resistance at 1950-2000. SOL is trading near $73, still struggling within a converging triangle. $69 is the bottom, and unless it breaks $79, there's no chance. The Panic and Greed Index is 36, the fear zone, perfectly confirming the saying—institutions quietly switch positions during weakness, not chase buying gains.
The core contradiction is that ETF capital flows are no longer a matter of "whether to enter," but "where." Bitcoin ETFs flowed out just two months after the end and recovered $170 million in July, indicating demand hasn't died down, just become more selective. Next, funds focus on targets that generate cash flow logic—staking yields, on-chain activities, and real users. The Intesa document sets the trend: institutional crypto allocations are shifting from "buying digital gold to hold on" to "buying assets that can work." #意大利大行减IBIT普通股94%, increased staked ETH The thing I don't like most about cutting issuance discource is lack of positive vision for Ethereum validator set. Original design was not prescient enough, but has clear design goals: hundreds of thousands of individual stakers, running Ethereum all across the world.
Status quo proponents also have a version of positive vision - e.g. Lido has built community staking, is decentralizing and geographically distributing validator set. We're successfully making a staking protocol that can bring Ethereum closer to original goals via opinionated, non-neutral decisions that base protocol can't adopt.
Cutting is just "do it or it's going to get worse", no clear vision of good validator set we're solving for. They're given up on original vision, but don't want to clearly admit it and offer a new one.Everyone is celebrating the massive $UNI outflows from Binance.
The narrative is simple:
"Coins are leaving exchanges... bullish."
Maybe.
But that's not the question I'm asking.
I'm asking what investors are actually buying into.
Right now, more than half of Uniswap's daily revenue comes from Robinhood Chain—a network that's only been live for a few weeks.
That's impressive...
But it's also a concentration risk.
If Uniswap v4 expands revenue across multiple ecosystems, great.
If not, the market could eventually start valuing $UNI as a Robinhood proxy rather than a standalone DeFi giant.
I'm still bullish on Uniswap.
I just think this is the risk almost nobody is talking about. 👀Currently, about 41.5 million ETH are staked on Ethereum, accounting for 34.03% of the total supply. There are still 2,483,600 ETH waiting to enter the validator queue, expected to take 43 days; only 128 ETH have left the queue.
Just as staking demand remained clearly strong, EIP-8363 proposed a completely different reward logic.
This draft sets a "saturation line" of 60.25 million ETH, which corresponds to about 50% of the total supply. As staking volume approaches this line, the proportion of validator consensus layer rewards burned will gradually increase; When 60.25 million tokens are reached, all protocol issuances received by normal validators will be fully offset.
Transaction priority fees and MEV income are not included in this burn, so the "staking rewards reset" is not entirely accurate. More precisely, the consensus layer bonus rewards will gradually reset, and validators may still receive execution layer revenue.
Currently, the staked amount is still 18.75 million short of 60.25 million tokens, requiring an additional growth of about 45.2%. Even if all 2.4836 million tokens in the queue are brought in, the staking ratio will only rise to about 36.1%, leaving a significant gap to the saturation line.
The controversy comes before the 50%.
The draft plans to complete the transition in 18 months. According to the community's estimate based on the proposal formula, when the staked amount is around 42 million, validators' total returns may drop from the current 2.65% to around 1.26%, nearly halving.
Based on a rough estimate for 32 ETH validators, the annual yield drops from about 0.85 ETH to around 0.40 ETH, not before deducting losses from hardware, electricity, operations, taxes, and downtime.
Large staking institutions can spread server, monitoring, and personnel costs among thousands of validators, while individual validators bear the fixed costs alone. The proposal originally aimed to avoid staking concentration in large custodians, but as yields declined, higher-cost individual validators might exit first, which is precisely where opponents most criticize.
DeFi will also be repriced accordingly. Staking yields are an important base interest rate for stETH, LST, LRT, ETH LL lending, and recurring staking strategies. Forum participants gave an empirical tip: staking returns need to be at least 0.3 to 0.5 percentage points higher than the borrowing cost for leverage to have enough room for circulation. Once yields are compressed, ETH borrowing demand, staking derivative income, and related protocol TVL may all be affected.
Supporters are concerned about another set of figures: Ethereum currently issues about 1.056 million ETH annually, with a supply growth rate of about 0.85% annually. If validator inflows remain fully loaded and exit volumes remain low for a long time, by early 2028, staking may exceed 70 million tokens, accounting for more than 55% of supply.
They argue that continuously increasing the staking ratio does not proportionally enhance security, but rather increases custodians and liquid staking platforms' control over ETH, while continuously diluting holders who do not participate in staking.
Currently, EIP-8363 remains an open draft core protocol, not approved nor officially included in the Hegotá upgrade. August 6 is only the deadline for submitting candidate proposals, not the final deadline for deciding on implementation.
This debate cannot yet be simply summarized as "positive deflation" or "negative staking." What really needs to be answered is: can reducing issuance lower staking concentration, or will it first eliminate higher-cost individual validators?
#以太坊草案EIP-8363 sparked controversy $BTC $ETH $SNDK Analysis of the current gold and silver rally trend
1. Current market situation
Spot gold prices surged to 4167, with a single-day gain of 2.2%. Silver strengthened in tandem, and precious metals closed higher for three consecutive days, with strong bullish sentiment in the market.
The Eurozone released two PMI figures in the evening, both beating expectations:
July Services PMI final reading: 51.7 (expected, previous: 51.6)
July composite PMI final reading: 52 (expected, previous: 51.9)
The recovery in Eurozone economic data indirectly suppressed the US dollar index, which is a positive external news for gold and silver.
2. Multiple Driving Factors 📊 Behind This Round of Gold Price Increases
1. U.S. inflation data cools down
US June PCE fell month-on-month, easing inflationary pressures, cooling market rate hike expectations, and declining real yields on US Treasuries, reducing the opportunity cost of holding gold, which is positive for precious metals.
2. Global central banks continue to make large-scale purchases of gold
The Bank of Korea resumed gold purchases after 13 years;
China's central bank has been purchasing gold for 20 consecutive months;
In the second quarter, global central banks purchased a total of 289 tons of gold, up 62% year-on-year;
Central banks' gold purchases are essentially a de-dollarization strategy, providing a long-term solid bottom for gold prices.
3. Reduced burden of geopolitical news
The temporary agreement between the U.S. and Iran in the Strait of Hormuz eased crude oil tensions, led to a decline in oil prices, easing inflationary pressures and giving the Federal Reserve room to operate its accommodative policies.
3. Market Characterization: Medium- to long-term slow bull market
This rise in gold and silver is not a short-term sentiment speculation, but is supported by three core logics:
Central banks around the world continue to buy at the bottom;
U.S. Treasury real yields have reached a turning point of decline;
Currency risks brought by overseas fiscal expansion.
Short-term nonfarm payroll and CPI data can cause oscillating and shake-out markets. Compared to high-volatility risk assets like Bitcoin, gold is more stable.
4. Risk Warnings ⚠️ at the Practical Level
Not suitable for a one-time heavy position to enter the market;
Prioritize waiting for price pullbacks before gradually building positions;
The Eurozone PMI has already taken effect, but in the evening, attention should be paid to US data and the short-term pullback caused by sudden fluctuations in the US dollar index.我们存的USDT刚被标普判了最差一档
标普全球这周干了件以前没人干过的事，它给稳定币发了一张某真正的成绩单。周一先给贝莱德新上的代币化货币基金 BRSRV 打了个 AAAm，这是标普能给基金的最高本金稳定性评级。周二又一口气给 11 种稳定币排了座次，结论是 11 个里头只有 6 个具备充足或更强的保住美元锚的本领。
先把这张榜摊开看。USDC、EURC、USDG、USDP 拿的是 2 级，标普用词是强。USDT 在 5 级，标普用词是弱。TUSD 和 USDe 一样待在 5 级。同一个被叫做稳定币的东西，底层质量差出三档，可大家平时买的时候，谁会去翻这一页。
为什么这件事值得你多盯一眼。USDT 是整个加密市场流动性的底座，绝大多数交易对、借贷仓位、跨所搬砖都绑在它上面，一直被当成准现金用。但从储备质量看，标普把它和 USDC 拉开了三档距离。这个评级不是今天才变差，是一直就在那，只是第一次有人白纸黑字写进公开报告。
再看贝莱德那只 BRSRV 在打什么算盘。它不是普通货币基金，是专门冲着 GENIUS Act 去的，目标是让自己的份额成为合格稳定币储备资产。它持有现金、93天以内到期的美国国债、国债担保隔夜回购，加权到期不超过 60 天，把波动压到极低。换句话说，以后稳定币发行方要按美国新规存储备，贝莱德想把自家代币化基金当成那个优质底层资产卖给他们。
这盘棋的意味在于，稳定币的钱生钱生意正在分层。一端是评级偏弱、储备构成常年被追问的头部稳定币，另一端是贝莱德用华尔街评级体系重新定义什么叫合格储备。谁掌握评级话语权，谁就握住了稳定币的底层资产入口。
给做波段的人留个实在观察点：别再把所有稳定币当成同一个东西。以后看链上稳定币净流出，得区分是哪种在跑。如果哪天是评级偏弱的那几只在集中流出，对流动性的冲击会比 USDC 这类强评级的流出大得多，盘面反应会更快更猛。标普这张榜，等于递给你一份提前排雷的清单。
真要说玩法，贝莱德这套评级即门槛的思路很老辣。它不跟你抢稳定币的发行份额，它抢的是你发行稳定币时背后必须配置的储备资产，这生意比发币更稳，因为不管谁发币，储备总得有去处。
你会因为一张评级单，把仓位里的USDT换成USDC吗。The sounds of the AI bubble bursting have not yet faded
But has the US stock market already started to reverse?
What has been the biggest concern in the market over the past few months?
It's not that AI lacks demand.
Instead, it's a more practical question:
So many giants are spending lavishly to build data centers, buy GPUs, and expand computing power.
In the end, can these investments really turn into income?
If not, AI becomes a massive capital black hole.
Therefore, the market recently showed clear divergences regarding the AI industry chain.
Stocks related to NVIDIA, semiconductors, and cloud computing have fluctuated sharply, prompting capital to begin to suspect:
Has AI already overdrawn the future ahead of time?
After Microsoft's earnings report, market expectations have somewhat changed.
Azure growth remains strong, with the company continuing to invest in AI infrastructure, while the market is seeing these investments translate into real cloud business needs. Microsoft's previously released financial report shows that Azure and other cloud services continue to grow rapidly, and the company continues to expand AI-related capital expenditures.
This means an important change:
AI is no longer just a "story of the future."
At least for now, some business models have already begun to take off.
This is also why the market has recently started to refocus on the AI industry chain.
From NVIDIA to Micron, from SanDisk to SK Hynix, funds have returned to the most direct beneficiaries of computing power, storage, and data centers.
Now let's look at the S&P 500.
After the previous round of adjustment, the index has not broken below its key structure.
Instead, it completed consolidation at a low level and then broke through the resistance zone again.
Looking at the candlesticks, after market panic was released, buying interest gradually returned.
If it can continue to hold the breakout level, it would be more like a trend correction than a simple rebound.
In fact, the S&P 500 has recently hit new highs, and market sentiment has warmed up as AI-related earnings expectations improve.
Of course, this does not mean that U.S. stocks are without risk.
Valuations remain very high.
Whether AI investments can sustainably deliver remains to be verified by subsequent financial reports.
But the biggest market change is:
People used to ask:
"Will AI make money?"
Now let's start asking:
"Who can make money from AI?"
These two issues are completely different from the same level.
The capital market's biggest fear is not companies spending money.
The biggest fear is that the money spent doesn't see a return.
And the signal Microsoft is giving this time is telling the market:
The AI investment cycle may be moving from the storytelling stage to the profitability verification phase.
If this logic continues, the reversal after this round of US stock market adjustments may only be beginning.
#标普500首次站上7700点, setting a new all-time high
$SPY Behind the 9-day 100-fold increase in coins stands Binance itself
Let's start with the timing. In the early hours of August 5, half an hour before SpaceX's earnings release, a meme coin called MarsCoin on BSC began to surge, with its market cap surging to $70.53 million, a historic high. The 24-hour increase was about 110%, with a turnover of approximately $25.1 million. Starting from its release on July 27, the cumulative increase over 9 days has exceeded 100 times.
The mechanism of this coin is worth taking apart; it's quite different from those old cats and dogs. When you buy or sell MarsCoin, you pay a 3% transaction tax per transaction. This money goes into the project treasury, is automatically converted into SPCXB, and then distributed to eligible holders according to the rules. SPCXB is the tokenized stock of SpaceX within the bStocks ecosystem. Simply put, it's turning SpaceX's stock into an on-chain token.
So holding this meme coin, in theory, you can passively receive exposure to SpaceX's assets. Sounds pretty nice—a meme coin actually has cash flow.
But who gets the profits depends on who has to settle the score. The SPCXB you receive comes from the 3% toll paid when others buy or sell. This income is entirely supported by trading volume; once volume cools, rewards are cut off. It's not dividends; it's just passing the word and changing the way you keep accounts.
What should be focused on is another set of numbers. It is on PancakeSwap V2 and the SPCXB pool, with total liquidity of about $944,000, containing approximately 8.1 million MarsCoin and 4,040 SPCXB. With a market value of 70.53 million, the actual amount you can actually take over is less than 950,000, a difference of more than 70 times. In 24 hours, the transaction volume reached 25.1 million yuan, meaning the same batch of shares was repeatedly traded within this small pool.
What does this structure mean? A few days ago, CATE already demonstrated it, with transaction volume under 1.5 million yuan pushing the price down to 60%. When the pool is shallow, the rise is fast because it can be pushed, and the drop is faster because it pushes forward. Only when you want to get off do you realize the depth is simply not enough.
Let's see who pushes from behind. On July 30, MarsCoin joined Binance Alpha, marking the first meme project in over four months there. The news briefly surged over 525%. On the same day, CZ publicly expressed support for meme coins. On August 1, Binance Wallet launched a stock meme section, and on August 4, announced reward rules: users of centralized exchanges must hold at least 10,000 coins per month to qualify. In the early hours of August 5, Aster launched perpetual contracts, and on the same day, Binance added 10 new bStocks tokenized stock trading pairs.
The flow of traffic is very clear. Over the past year, the main arena for memes was Solana. Pump.fun's approach closed the chain for token issuance and dissemination, and the aura of token listing on trading platforms has faded. This series of moves essentially aims to reclaim the traffic entry points for themselves.
Here's a practical indicator for swing traders: don't just look at market value and gains—divide market value by pool depth to calculate. The higher this multiplier, the closer the list you hang is to the decoration. Blockworks' statistics are right there: 1,972 tokens outperformed BTC by only 4.1%, with a median loss of 97%, and even fewer survive two years.
In the long run, the direction where memes and RWA are intertwined might not be out of reach. If you can truly turn attention into asset returns, it's a new path. But in this current version, value capture still depends on continuous orders, not a moat.
Would you subscribe to a pool worth less than a million coins just to get a token stock reward? #1 trillion Korean won poured into data centers without a single cent entering the cryptocurrency market
Today, Seoul signed an agreement: the department in charge of technology and communications and the Korea Development Bank will jointly fund a 1 trillion won AI infrastructure fund. Converted to about $700 million, the absolute value isn't shocking, but the direction is very fixed: next-generation AI data centers, plus supporting power and energy facilities.
Follow this money once more, and the route is very clear. The starting point is public funds and policy banks; the middle point is funded to leverage private capital; the endpoint is racks, substations, and cables. The last time these two institutions joined forces was in 2006, when they invested in universities, Incheon Airport Railway, the Sinbundang Line, and power plants. After 20 years, they reunited, and the investment shifted from railways to computing power.
The authorities have stated that whether a country can stably secure data centers and power networks is becoming a matter that determines national competitiveness. To put it simply, what you're fighting over next isn't chips, but sockets.
On the same day, the Korean stock market absorbed over 1.4 trillion won in foreign capital net purchases, once again boosting risk appetite in the AI boom. The flow of these two funds is highly consistent, both pushing toward computing power, with no cent aimed at crypto assets.
And it's not just Korea. In today's Galaxy financial report, the data center division made a profit for the first time by leasing 133 MW of data centers to CoreWeave. TeraWulf is more direct, signing a lease with Anthropic for about 401 MW with a 20-year term and an initial contract value of approximately $19 billion. Those who used to mine coins or run trading platforms are now competing to become landlords.
What about us? BTC is still trading between 63,000 and 64,000, with the 200-week moving average at 63,657 hovering near the current price, and the Panic and Greed Index at 28. CoinDesk's data set was quite striking: both the S&P 500 and Nasdaq, priced in Bitcoin, broke through the 200-week moving average, marking the first time since 2012. To put it plainly, for many years, buying coins was more cost-effective than buying US stocks for the same amount of money, but now that advantage is temporarily gone.
This is the real reason for the bottoming out. It's not that some negative news is dropping; it's that incremental funds simply haven't gone this way. Once you understand this, you need to change your swing trading mindset. In an environment without new money entering the market, the probability of expecting a bullish candlestick to start a one-sided market is very low. The upper and lower boundaries of the range are repeatedly tested, and trading volume keeps shrinking—this is the current norm. The real signals to watch are not in the market itself, but when AI capital spending shows a turning point and when dollar liquidity eases.
Looking further ahead, money is piled up on appliances and cabinets, and once it reaches a certain point, there will inevitably be overcapacity and declining returns, at which point the funds will have to find new sources. What crypto should do during this period is solidify cash flow and compliance, waiting for their turn. Worrying is useless, and blaming the market is useless.
How long do you think it will take for this round of funds to remember crypto streets, or do you simply don't think it will come back? #The thing I don't like most about cutting issuance discource is lack of positive vision for Ethereum validator set. Original design was not prescient enough, but has clear design goals: hundreds of thousands of individual stakers, running Ethereum all across the world.
Status quo proponents also have a version of positive vision - e.g. Lido has built community staking, is decentralizing and geographically distributing validator set. We're successfully making a staking protocol that can bring Ethereum closer to original goals via opinionated, non-neutral decisions that base protocol can't adopt.
Cutting is just "do it or it's going to get worse", no clear vision of good validator set we're solving for. They're given up on original vision, but don't want to clearly admit it and offer a new one.Sandi, you chased after me to kill me, right?
Family, I'm really speechless.
SNDK short position, average opening price 1423.69, average closing price 1470.83, loss -3.93U, return -68.25%.
Short selling is like a single needle pulling the bomb out, then crashing back in. I'm all too familiar with this script.
📌 How did this order come to an end?
After shorting, SNDK dipped slightly to 1402, with a decent unrealized profit. Then the price began to rebound, directly pushing up to 1483.62.
My short position was forced out at 1470.83, which was precisely crushed, and then the price crashed back to around 1402.
The same familiar formula, familiar flavor: first pull up the bears, then smash them back.
🔍 Why did this needle appear?
(1) #闪迪财报前夕, HBF and storage shortages have sparked heated discussion, with intense capital competition
SNDK's earnings report approaches, and HBF and storage shortages have sparked heated discussion. Market expectations for the earnings report are highly divided, with both bulls and bears betting on it.
In such cases, major funds often use information advantages for short-term harvesting. Pull a needle to blow up the bears, then smash it back to capture both the bulls and bears.
(2) After major players reduce their holdings, the bears are targeted for harvesting
Yesterday, four early long positions sold 8,321.8 shares of SNDK, which is about $11.111 million at current prices.
After major players reduce their holdings, there may be more bearish positions in the market. When the main players saw the short sellers gathering, they immediately triggered a targeted liquidation.
💡 Where did he go wrong?
First, 20x leverage is still too high before the financial report.
The opening price was 1423, the forced discount price was 1479, a difference of 56 pips (about 3.9%). Volatility is already high on the eve of earnings reports, and the 3.9% margin for error is simply insufficient.
SNDK rose from 1402 to 1483, up 5.8%. With 20x leverage, a 5.8% reverse fluctuation is enough to blow up a position.
Second, shorting before the earnings report = betting on earnings falling short of expectations.
Shorting before the earnings report is essentially gambling on "poor earnings." But when it comes to financial reports, no one knows the outcome until the very last moment. Big money has an information advantage, while retail investors can only guess.
Third, it overlooks the common tactic of "pushing up the market and blowing up the short market" before earnings reports.
On the eve of the earnings report, both bulls and bears are betting on it. Major funds often pull a needle before earnings reports to blow out the short market, then pour it back. I've been harvested by this trick several times already.
🎯 What happens next?
1. Before SNDK's earnings report, I stopped touching it.
No matter if it's bullish or bearish, I won't touch it before the earnings report. Wait until the financial report is out and the direction is clear.
2. If you really want to do it, it's also a light position + lower leverage.
Use less than 10 times to allow enough room for volatility. Or just use spot trading without touching contracts.
3. Wait until the financial report is released.
After SNDK's earnings report is released, the market will provide a clear direction. Entering the market with the trend will be much more reliable than betting on the current direction.
Final thoughts
Family, I really admire it.
If you short the market, it will explode; after the explosion, you will smash it back.
Sandi, you chased after me to kill me, right?
But I've come up with a rule: before the earnings report, it's best to hold back.
(This is purely a personal trading record and does not constitute investment advice.) ）
$SNDK
#交易之声: Your experience deserves to be heard USDT Market Cap Decline Reaches Historically Extreme Levels
“Historically, the market's deepest USDT contraction phases have also marked points where selling pressure was closer to exhaustion than to further acceleration.” – Read More$BTC $ETH Terra peaked at 43% LUNA stake at something like 15% APR & 10,000 airdrop and liquidity mining schemes layered on top of staking while being the hottest ponzu in crypto...also having in-protocol delegation and in-protocol LST so less risk...also being in a low-interest-rate QE/covid subsidy macro environment that we'll probably never get again in our lifetimes...
there is basically no reason to think $ETH staking will ever get to 50% no less 100%, just a bunch of hysterical bullshit....$ETH #EarningsRealityCheck 特朗普政府正在考虑依据“232条款”，对进口多晶硅及相关产品设置最低价格并加征关税，重点针对中国长期补贴形成的低价产能。政策尚未最终落地，但已进入实质性讨论阶段。
美国这次开始往更上游动手了，准备给进口多晶硅设价格底线、加关税。
对美国本土材料厂当然是保护，但光伏开发商和下游企业就未必笑得出来了，原材料一涨，最后还是项目成本买单。
现在半导体和AI供应链越来越不像纯技术生意，更像政策、关税、补贴一起定价。
对MU和MRVL暂时没什么直接利好，我也不想硬蹭概念。真正值得关注的是，美国正在把“供应链安全”从芯片一路往原材料推，后面成本和利润可能都要重新算。SpaceX首份财报：营收翻倍、亏损收窄，盘后却暴跌8%——老默告诉你解禁才是真正的胜负手
兄弟们，昨晚SpaceX发了上市以来首份财报。
数据相当能打。二季度营收78.14亿美元，同比增长92%，远超华尔街预期的69亿美元。净亏损5.41亿美元，较去年同期的10.08亿收窄46%。调整后EBITDA 35.38亿美元，同比增长191%。每股亏损0.09美元，市场预期亏0.26美元。
三大业务板块，全部超预期。
星链所在的连接业务收入42.91亿美元，同比增长66%，运营利润16.6亿。星链用户达1200万，同比翻倍。AI业务收入25.61亿美元，同比暴增247%，首次实现调整后EBITDA盈利。航天业务收入9.62亿，同比增长29%。
业绩全面超预期，盘后股价一度跌超8% 。
为什么？资本开支炸了。二季度资本支出183.7亿美元，其中AI相关158亿。去年同期才28亿，一季度77亿。电话会上CFO说Q3、Q4资本支出跟Q2差不多。AI业务Q2运营亏损12.6亿美元，收入25.6亿，亏了将近一半。
市场在算一笔账：营收78亿、资本开支184亿——烧钱速度是赚钱速度的两倍多。马斯克说投资回收期不到一年，但市场还没看到回报，只看到钱在往外烧。
但更大的变量在后天——8月6日，9.115亿股限售股解禁。
当前可交易流通股只有总股本的5%左右。8月6日这一波直接翻倍到约12%。年底前流通盘要扩到53.3亿股，占总股本约40%。
空头已经提前布好局了。截至7月29日，SPCX空头持仓2.193亿股，占可公开交易股票的34%。6月16日首次披露时只有2330万股。S3研究主管说得直白："目前的大赌注是围绕解禁展开的，市场押注财报不足以抵消大量解禁股票进入市场的影响"。
27家机构给出"买入"评级，平均目标价223美元。大摩目标价300美元。瑞穗200美元、麦格理250美元。但派杰投资刚把目标价从156下调到140美元。
卖方一致看多、空头疯狂加仓——这种极度撕裂的局面不常见。
财报触发的解禁是自动的，不管业绩好坏。8月6日一开盘，最多9.115亿股可以卖。这批股份成本极低，卖就是纯利润。他们会不会卖？什么时候卖？卖多少？没人能回答。但卖盘压力是实实在在的。
说回盘面。SPCX从6月高点225.64跌到目前125附近，跌幅超44%。盘前一度跌超10%到112.69。上方阻力135-140（IPO发行价区域），下方支撑108-110。
老默说几句实在话。SPCX这份财报数字不差——营收翻倍、亏损收窄、星链用户破1200万、AI收入暴增247%。但市场不买账，因为资本开支太高了，而且还要继续高。
更关键的是后天9亿多股解禁。不管财报多好看，供给冲击是实打实的。空头浮盈巨大还在加仓，赌的就是解禁后有人扛不住要卖。
操作上：有仓位的，今天找机会减仓或设好保护止损。想抄底的，等解禁落地、抛压消化完再说。108-110是第一个观察位，缩量企稳可轻仓试多，止损放100下方。
SPCX这波解禁，你觉得能扛住吗？评论区聊聊。
觉得老默拆得清楚的，点个赞关注，解禁落地我第一时间喊你。$BTC $SNDK $SPCX #财报观察员：业绩喜忧参半，解禁将至！SpaceX后续怎么看？