Orbit Post Sitemap
Big Tech's earnings delivered the reality check the AI trade has been dodging. Alphabet and Tesla both reported, and both stocks sank, not on weak results (Google Cloud grew 82%) but on rising AI capex guidance. The market has flipped: spending on AI used to be rewarded as vision, now it's scrutinized as cost.
This is the same story that hit semis all week, viewed from the demand side. Investors aren't questioning whether AI is real; they're questioning the return on hundreds of billions in capex before the revenue catches up. For crypto it's a useful mirror: narratives get repriced the moment the market demands proof over promise. Risk-off today (BTC $64K) echoes that same "show me the ROI" mood bleeding across tech.
Just my read, not advice.
#USIranStrikePause #OpenWeightSupport OKXOrbit🚨 Two mega-cap earnings. One clear message: the market is looking beyond headline beats.
Alphabet posted a strong quarter with $119.8B in Q2 revenue, while Google Cloud continued to deliver impressive growth. Yet $GOOGL fell more than 4% after hours.
Why? Investors cared more about the outlook than the quarter itself.
Alphabet raised its 2026 capex guidance to $195B–$205B, up from $180B–$190B, while free cash flow turned negative. AI remains a compelling long-term growth story, but Wall Street is becoming increasingly focused on the cost of financing that growth.
Meanwhile, Google, Microsoft, Meta, and Amazon are expected to spend a combined $725B on capex in 2026—roughly 77% higher than last year.
The takeaway: markets are rewarding more than earnings beats. Forward guidance, cash flow, and AI spending discipline now matter just as much.
Tesla told a different story.
The company still holds 11,509 $BTC , unchanged since 2022. Despite recording a $112M quarterly Bitcoin-related loss, Tesla neither added to nor reduced its position.
No panic. No accumulation. Just HODL.
📊 Why this matters for crypto:
• $BTC continues to benefit from steady ETF inflows.
• Crypto remains closely tied to the Nasdaq 100, making Big Tech earnings an increasingly important macro driver.
• Upcoming reports from Microsoft, Meta, and Amazon could influence both equity and crypto sentiment through their guidance.
One key advantage for crypto traders: while U.S. stock markets close after hours, crypto trades 24/7.
With OKX tokenized U.S. stocks settled in $USDT, assets like $XGOOGL and $XTSLA remain tradable through earnings releases and weekends.
👀 Will the next wave of Big Tech earnings strengthen—or weaken—crypto sentiment?
#EarningsRealityCheck #CLARITYActStalled 价格新高下的流动性背离：市场并非普涨，而是资金在高度集中
市场表面看似强势，但真实定价与资金分布之间是否存在明显裂痕？
事实层面，原文指出价格虽创下更高点位，但流动性并未同步扩散。未平仓合约在近期出现重置，交易量仍处高位，说明交易行为正从追逐每一波突破转向极端选择性。资金并未均匀流向整个加密市场，而是集中于少数资产，如JELLYJELLY、OPG、SLX、LAB、BSB、ALLO、CHIP、MEME、EDEN、HUMA、ZKP、METIS；同时，BEAT、EDGE、COAI、TRUMP、RAVE、SPACE、SOPH、IP、AVNT、ZAMA、OFC、PIEVERSE、VIRTUAL、ACU、H、MEGA等代币流动性持续流失。
市场结构变化：BTC仍是流动性中枢，ETH代表机构资本风向，SOL作为高贝塔领头羊，DATA映射AI基础设施敞口，WLD承载AI身份叙事，HYPE衡量投机欲望，ZEC与DOGE则反映散户参与度。这种分层定价表明，资金正从广泛的山寨市场回撤，集中押注于少数有明确叙事或流动性支撑的标的。
传导逻辑：当BTC和ETH维持高位但未拉动多数山寨时，意味着增量资金并非来自散户涌入或被动配置，而是存量投机资本在主动筛选。这会导致流动性进一步向头部资产倾斜，山寨的弱势参与反过来抑制风险偏好，形成负反馈。
偏多路径：如果BTC/ETH能持续站稳并放量突破，可能吸引被动配置资金回流，带动流动性从集中区向边缘扩散，重新激活山寨市场。条件是宏观环境稳定，且BTC未出现持仓集中度风险。
偏空风险：流动性持续收缩而价格虚高，一旦BTC或ETH出现回调，集中持仓的资产可能面临踩踏，导致山寨市场加速失血。失效条件包括未平仓合约快速膨胀或交易所存量BTC大幅流入。
结论：当前价格新高更多反映资金集中而非真实需求扩散，弱势参与比弱势价格更值得警惕。风险在于流动性假象下的结构脆弱性。
$BTC $ETH $SOL$BTC
I did my best to explain my thesis on all of these important questions that almost no one is talking about.
No one is talking about Bitcoin’s trend angles.
Throughout its entire history, Bitcoin has consistently respected its long-term trend angles. With every market cycle, those trend angles have continued to decrease, and they have now reached a point where making new all-time highs is becoming increasingly difficult.
Everyone tells you that every Bitcoin bear market lasts exactly one year. But history tells a different story.
Out of Bitcoin’s four major market cycles, only the last two bear markets lasted around one year.
First bear market: 154 days
Second bear market (2013–2015): approximately 630 days
Third and fourth bear markets: around one year
Yet almost everyone continues to copy the post-2018 pattern and claims there are only 1–5–2 months left in the bear market, while completely ignoring Bitcoin’s earlier history.
Before 2018, Bitcoin had two bear markets that did not last one year—they lasted 154 days and approximately 630 days.
At the same time, most people tell you not to look for the bottom. Instead, they recommend buying randomly from now through the next 1–5–2 months, assuming history must repeat exactly as it did after 2018, while ignoring the rest of Bitcoin’s market history.🚨 NVIDIA CEO Jensen Huang says:
«"No chip bust for a while."
"This time is different."»
He believes the semiconductor industry still has room to grow 5–10x.
But here's what stands out:
- Meta: 21% of NVIDIA's revenue
- OpenAI / Oracle: 17%
- xAI: 16%
Just three customers account for 54% of NVIDIA's total revenue.
Meanwhile:
- 2026 hyperscaler capex: $785B
- 2027 forecast: Nearly $1T
- TSMC capex: $60–64B
- Intel capex: $20B
- U.S. chip factory utilization: 72.2%
The AI boom is being fueled by massive spending from a relatively small group of companies.
If even one of those major buyers meaningfully slows its AI infrastructure investment, the industry's growth outlook could change much faster than many expect.
The key risk isn't that spending stops—it's that capex growth slows.
That's what the market will be watching.
#EarningsRealityCheck #CLARITYActStalled $SHIB SHIB suddenly pulled up—is it about to take off, or is it just another scam?
SHIB's volatility today was quite insignificant, with its price rising over 20% at one point and a noticeable increase in 24-hour trading volume. On the surface, it seems like a sudden start, but after looking around, I didn't find any super positive news that could completely change the project's fundamentals.
This surge seems more like several factors coming together.
Recently, some whales have repurchased SHIB, while SHIB balances on exchanges continue to decline. Simply put, the coins are moved into on-chain wallets, and in the short term, the amount of chips willing to dump may decrease. Moreover, SHIB's liquidity is not as deep as Bitcoin's, so whenever funds suddenly flow in, the price can be quickly pushed up.
But I think people shouldn't start fantasizing about "removing two zeros immediately" just because SHIB goes up.
SHIB's current circulating supply is still close to 589 trillion, which is simply too large. Although the project has been burning for a long time, the amount burned in the past 30 days is still very limited compared to the overall supply, making it difficult to drive long-term price increases through burning alone.
Shibarium is still running, and its ecosystem hasn't completely gone flat. However, on-chain activity has been unstable recently, with daily trading volume even dropping noticeably not long ago. This shows that SHIB's strongest driving force is still not the app explosion, but community sentiment, whale movements, and MEME rallies.
So my view on this wave is very straightforward:
You can see this as a signal of SHIB's reactivation, but you can't immediately assume a major rally has arrived.
What really matters is not how much it rose in a single day, but whether trading volume can be maintained, whether funds will continue to flow out of exchanges, and whether there is heavy selling after prices rise.
These coins do surge rapidly, but when they turn around, they never say goodbye.
Do you think SHIB is truly starting this time, or is it just a rally to keep tying people in?🚀 $HYPE — LONG Setup 📈
Trade Bias: LONG ✅
📍 Entry Zone: 59.1 – 59.6
🎯 Take Profit Targets:
• TP1: 60.5
• TP2: 62.0
• TP3: 64.0
🛑 Stop Loss: 58.2
⚠️ Risk Level: Medium
📊 Technical Outlook:
$HYPE continues to show strong bullish momentum after a $6.177K short liquidation around 59.418, adding fuel to the recent upside move.
Buyers remain in control, with price holding above key support and momentum favoring further gains as long as the breakout zone remains intact.
🔹 A sustained hold above the entry zone keeps the bullish outlook valid.
🔹 A break above 60.5 could accelerate momentum toward the higher targets.
Stay disciplined, manage your risk, and wait for confirmation before adding exposure.
Not financial advice. Always do your own research.
Let's go $HYPE! 🔥
#EarningsRealityCheck #CLARITYActStalled #USIranStrikePause
$BTC $ETH $DOGE 🚨 Two mega-cap earnings reports. One takeaway: markets are looking beyond headline beats.
Alphabet delivered a strong quarter, reporting $119.8B in Q2 revenue, while Google Cloud continued to post robust growth. Yet $GOOGL fell more than 4% after hours.
Why? Investors focused on the outlook rather than the results.
Alphabet raised its 2026 capex guidance to $195B–$205B, up from $180B–$190B, while free cash flow slipped into negative territory. AI remains a powerful growth story, but Wall Street is becoming increasingly concerned about the cost of funding it.
Google, Microsoft, Meta, and Amazon are now expected to spend a combined $725B on capex in 2026 ,around 77% more than last year.
The market is rewarding more than earnings beats. Forward guidance, cash flow, and AI spending are becoming just as important.
Tesla told a different story.
The company still holds 11,509 BTC, unchanged since 2022. Despite reporting a $112M quarterly loss tied to Bitcoin's earlier decline, Tesla neither sold nor added to its position.
No panic. No accumulation. Just HODL.
📊 Why this matters for crypto:
• $BTC continues to benefit from steady ETF inflows.
• Crypto remains closely correlated with the Nasdaq 100, making Big Tech earnings increasingly influential.
• Upcoming results from Microsoft, Meta, and Amazon could shape both equity and crypto sentiment through their guidance.
One advantage for crypto traders: while US stock markets close after hours, crypto never sleeps.
With OKX tokenised US stocks trading 24/7 in $USDT , assets like $XGOOGL and $XTSLA remain tradable through earnings releases and weekends.
👀 Will the next wave of Big Tech earnings strengthen or weaken crypto sentiment?
#EarningsRealityCheck #KoreaAIChipPush BSB on-chain data shows that 90% of the tokens are distributed across 8 whale addresses, showing a high concentration! The funding rate has been falling all along, yet the funding rate has always been bought by the bulls, rewarding the bears. This is suspected of whale manipulation and short-selling of the market, feeding retail investors on funding rates. When bulls encounter market-poor market players, they just accept their bad luck!一边全线回落一边疯狂扩产，AI芯片赛道出现完全割裂的两极行情
当下市场同时上演两种反差极强的景象，一边板块个股集体走出回落行情，另一边全球头部科技企业仍在不计成本加码算力、存储产能，两种完全相悖的市场行为同时存在，藏着整条AI硬件产业链的深层变化。
先梳理近期盘面客观波动数据，本轮AI硬件板块分化十分明显。英伟达收盘定格206美元，持续在前期高点下方承压运行；博通单日录得2.7%的跌幅，存储赛道的美光波动幅度更大，单日下跌幅度达到7个百分点，硬件存储标的集体走弱。
宏观层面的变量正在持续影响整个科技板块，下周即将迎来FOMC议息会议，当前市场对于利率变动的敏感度已经拉满。资金形成统一共识，一旦利率环境收紧，高估值科技企业会率先承受资金流出压力，这也是近期芯片、存储板块承压的核心宏观背景。
但跳出短期盘面涨跌，产业链实体端的布局节奏完全没有放缓。头部AI企业的算力采购需求依旧旺盛，A社、OpenAI持续大批量购入算力硬件；海外产能建设同步提速，英伟达落地韩国AI超级工厂的扩建计划持续推进，各大存储厂商持续落地新增产能项目。
韩国本土产业链竞争格局也迎来新变化，以往海力士独占HBM高端存储赛道的局面被打破，三星加速入局抢占HBM4供应链份额。当前高端AI服务器硬件架构持续升级，单台设备搭载GPU数量不断提升，算力规格升级直接带动高端HBM存储的需求持续走高，各大存储厂商都在争抢这条核心增量赛道的供货话语权。#韩国存储双雄获AI双巨头大单
产业端扩张与二级市场回落形成鲜明反差，短期盘面价格和实体产业布局走向完全背离，这种割裂状态也是当下AI硬件板块最值得细细拆解的特征。下周议息会议的表态，会成为判断本轮板块调整属性的关键参考，能够分清本次下跌属于短期回调，还是高估值赛道估值重塑周期开启的分界节点。
不少人只盯着单日个股涨跌做判断，很少结合实体产业落地进度、宏观利率周期综合梳理板块节奏。单纯看K线很容易被短期波动误导，把产业长期增量和二级市场短期震荡混为一谈。
你们平时分析科技赛道，会优先参考二级市场价格波动，还是重点跟踪全球厂商产能、订单落地的实体产业消息？TRUMP TEAM JUST MOVED MORE $TRUMP
Trump's team moved 10.84M $TRUMP worth ~$16.91M to BitGo. The next stop is likely exchanges.
This is the third big transfer in the last 5 months.
So far, they have moved 48.25M tokens worth $172.4M.
After every previous transfer, the price went lower.
The token is now down 66%+ over the same period.
$TRUMP is also down more than 90% from its all-time high since President Trump took office, and it continues to trend lower.
This doesn't look like normal treasury management.
It looks like they are selling into the market.In 2026, semiconductors will be dominated by AI computing power, with significant market growth. HBM and advanced packaging are in the shortest supply, mature processes remain in demand, and advanced process iteration is slowing down. The US, Japan, South Korea, and Taiwan control global high-end supply chains, normalizing geopolitical controls. China is focusing on filling gaps in equipment and materials, automotive-grade chips are performing steadily, while traditional consumer electronics chips are weak. In the long term, caution is needed regarding risks of overcapacity and insufficient computing power usage next year. ##韩国存储双雄获AI双巨头大单 ##$SNDK Players familiar with the $SOL community know an unwritten rule: to gauge Solana's overall atmosphere, you can catch a glimpse of BONK by watching closely. Recently, a large number of short-lived new meme users on the chain have quickly faded, with funds shifting back to established meme brands. BONK has taken advantage of this trend to start a rally. Today, let's analyze the underlying logic in detail. First, the underlying background: BONK is a milestone meme born during Solana's downturn. Back when Solana experienced a sharp decline and a large number of users fleeing, the market was pessimistic, and $BONK emerged. It lacks grand technical narratives, relies on community consensus to ignite the entire chain's popularity, and is also Solana's first top-tier meme to break out. Since then, countless meme imitators have emerged one after another. For a long time, BONK has become more than just an ordinary meme—it has become a recognized ecological mood thermometer within the community. As soon as Solana retail trading enthusiasm returns, funds often immediately think of this established token. Considering the current market situation: SOL has recently steadily recovered, driving a revival of trading atmosphere across the entire public chain. Recently, the market has been frantically chasing the endless stream of new meme products, but most new projects have very short lifecycles, quickly stalling after just a few days of gains, causing many players to fall into traps and incur losses. Funds learned from this lesson, began to avoid illiquid new coins, and shifted to established stocks with ample trading depth. BONK saw massive buying and simultaneously surged in rallies. In-depth analysis of the core logic of this round of gains: Currently, it is an internal capital rotation within the sector. The Solana ecosystem is not short of speculative funds, only capitalSIMD-0096 isn't a technical update, but it directly rewrites Solana's economic logic. Previously, half of the priority fee was burned, and half went to validators. What about now? Validators eat half of the burned parts, and losing them sounds like giving miners a chicken leg, right? But looking further down, validators receiving more $SOL liquidity incentives means their willingness to lock up is stronger. Circulating pressure is reduced, and selling pressure naturally decreases. This isn't speculation; it's a slow explosive. Look at the recent $SOL price—it's stuck at a high level but not going down, and trading volume hasn't shrunk. Meme coins like $BONK and $WIF in the ecosystem are even showing signs of a second restart. I think the market is already voting with its feet, waiting for retail investors to react. Most likely, it's another chance to take over. Solana is very smart this time. They turn validators into community of interests. The more congested the network, the higher the fees, and the more validators earn. So what will they do? Buy more $SOL to stake, forming a closed loop—that's the real moat. Sisters, stay steady—don't be scared off by market volatility. On-chain data doesn't lie. Solana's current daily active address count and DEX trading volume have long left other L1s far behind. For those still waiting for $SOL to pull back to 80 or 90 to bottom-fish, I just want to ask: in bear markets, you've dropped to that level before—did you buy? Now that the rules have changed and the economic model has been upgraded, inflationary pressures have even decreased, making you less afraid to get on board? Honestly, I think in this bull market, more than a hundredfold coin $SOL will emerge in the Solana ecosystem$MSFT $GOOG $AMZN $TSLA $META All have fallen below the 50-day and 200-day moving averages.
$NVDA Falls below the 50-day moving average but remains above the 200-day moving average.
$AAPL Strongly crushing two moving averages.
This is not weakness. This is the classic Mag7 rotation, paving the way for the next wave of gains.
Apple is currently a quality leader—its highest valuation is justified (ecosystem, cash flow fortress, AI potential). When the other Mag7 members lag significantly while AAPL remains strong, history shows that once the rotation ends, the laggards will fiercely catch up.
We are witnessing a pattern exactly like the previous two major Mag7 surges. Oversold stocks + a clear leader = rocket fuel for the entire group.
The only way I know is 🚀Two days ago, $CORE was hovering around 0.023, looking toward a new all-time low. Today, it suddenly rebounded above 0.028, with a 24-hour gain of +8.9%, and its market cap returned to around $32 million. The market looks lively, but if you compare the candlestick + chain chain, the flavor doesn't quite match. K-line: Rebound Gained, Confidence Still Lacking. In the past two days, CORE has pulled a decent bullish candlestick from its stage low, temporarily regaining the short-term moving average. However, the daily 20/50/200MA remains a standard short range, MACD bars haven't fully closed, and RSI climbed up from 32.5 without a golden cross. Veteran players are familiar with this pattern—the technical corrections in the decline outweigh trend reversals. Two key points to watch: • Volume: During the rebound, trading volume has not significantly expanded, indicating a weak recovery pattern of "some buy at low prices, but no one really buys"; • Resistance: 0.030–0.032 is the previous breakdown level; if it doesn't rise, it means double top preserve. On-chain: The main force is topping up exchanges—this is the real signal. Even more straightforward than candlesticks is on-chain. In the past 48 hours, an address that has been watched for a long time 0x611f... d09d (the market calls it the "core main force") repeatedly deposited millions of CORE tokens on OKEx—this old "withdrawal→pumping → recharging back to exchanges" scenario is obvious to you. Hedging data is another side: Santiment shows whales in the million–tens of millions have net increased holdings by about 420 million in recent weeks, with small retail investors dumping and large players taking over. But be careful—those receiving goods and those depositing exchanges may not be the same group; internal differentiation is a comparison tableWhile others are watching the K-line chips, smart money is "laying pipelines" on the chain
Date: July 27, 2026
Today's market sentiment survey shows that over 70% of retail traders are still chasing AI concept coins and zoo-like meme projects, with long positions accounting for as much as 83% of the total 24-hour liquidation amount in the entire cryptocurrency market. But if you only focus on these, you fall into the trap of the "illusion of flow."
I want to make an unconventional judgment: the last thing you should watch right now is the gainer rankings; What should be watched most now are those infrastructure layers that are so deserted that almost no one pays attention to them.
Let's first look at a set of facts that just refreshed today: the average daily active addresses on Ethereum's Layer 2 network (L2) officially surpassed 6.8 million today, setting a new all-time high. In contrast, the median gas fee on Ethereum mainnet fell to 0.8 Gwei—the lowest level since the end of the 2022 bear market. What does this indicate? This indicates that on-chain activity is truly exploding, but speculative enthusiasm is rapidly fading. Large funds are quietly and patiently completing "turnover": withdrawing from highly volatile assets and moving into underlying protocols that can generate real returns.
Another overlooked data point is that today, the total deposits of global stablecoins (USDC+USDT) in decentralized lending protocols actually increased by $420 million against the trend compared to the same period last week. This is a textbook departure from the bleak secondary market. Retail investors are selling, institutions are holding.
Why am I bringing this up? Because I've seen too many people make the same mistake: scoff at the 5% annualized yield of liquidity mining on perpetual contract DEXs (decentralized exchanges), only to break their thigh after it surged tenfold.
Today, the daily average trading volume of leading decentralized perpetual contract protocols has quietly climbed to 8.7% of the total perpetual contract volume on centralized exchanges (CEX), compared to only 2.1% in the same period in 2025. This is no longer a "geek toy"—it's a real cash migration. South Korean exchange Upbit just updated its asset reserve proof early this morning, showing that its holdings in DeFi blue-chip tokens increased by 217% over the past 30 days. This logic follows its previous listings on Morpho and Euler—the listing departments of mainstream exchanges understand the value of the "underlying asset" better than retail investors.
The current market situation is:
· Where Others Go Crazy: Narrative-driven tokens, with daily turnover rates over 80% and jaw-dropping volatility.
· Places others fear (or completely ignore): interest rate derivatives protocols, decentralized credit scoring protocols, and governance tokens that "have no sexy stories, only stable cash flow."
The favorite tactic of manipulators and smart money is to open positions when liquidity runs dry, and then distribute when liquidity floods. At this point in time, USDT's OTC premium has returned to positive levels, and Korea's premium has nearly dropped to zero—this is precisely the standard characteristic of being "ignored."
You don't need to go all in right now. What you need is: spend three hours today without looking at any market software, just look at the liquidation data, historical fluctuations of funding rates, and the protocol's revenue-sharing mechanisms for three decentralized perpetual contract protocols.
What I regret most isn't missing out on a hundredfold coin, but that when on-chain options protocols first emerged in 2024, I found them complicated and illiquid, and didn't allocate 5% of my position for 'trial and error.' As a result, two years later, the leading agreement in that sector had already paid over 200 weeks of stable dividends.
Remember: what makes you money is never "knowing," but "doing" and "arriving early."
While everyone is crowding the main roads to grab gold, the real winners are already selling shovels and repairing highways. Today, the "concrete" of on-chain finance has only just been fully dried. Are you sure you want to wait until the trucks are fully loaded before chasing after it?
(This article does not constitute any investment advice. The market carries risks, and decisions must be made independently.) ）This is going to be a very interesting week for $BTC .
Over the past 12 months, eight of the last nine FOMC meetings have been followed by a relatively large sell-off.
Across those eight flushes, BTC declined roughly 10% on average over the following week.
During last month’s meeting, price was trading in almost exactly the same region as it is today.
BTC traded around $66K, then dropped roughly 12% to $58K, setting new cycle lows.
The one exception was the previous meeting in May, when BTC produced the opposite reaction and rallied roughly 5%.
So another bearish reaction is not necessarily guaranteed. We have already seen this pattern fail once during the current bear market.
But 8 out of 9 is still not a statistic I am interested in betting against.
If the same reaction plays out again, we’re likely to see a key test of the range lows.
I’m personally watching whether $61K can hold as support.
That level is the gatekeeper between another pullback inside the current range and a potential flush to new lows.
Either way, the reaction we see after this meeting is going to be a good indicator as to whether we see fresh cycle lows again soon.The crypto ETF race is heating up. Several digital assets have already secured spot ETF approval in the US, while a growing list is still waiting for the green light. ✅ Already live in the US (9): $BTC — Jan 2024 $ETH — Jul 2024 $XRP + $DOGE — Sep 2025 $SOL — Oct 2025 $LTC — Nov 2025 $DOT + $AVAX — Mar 2026 $HYPE — May 2026 ⏳ Filed & awaiting approval (13): 🔹 $ADA — VanEck, 21Shares, Grayscale 🔹 $LINK — Bitwise, 21Shares, Grayscale 🔹 $XLM — 21Shares, Bitwise, Grayscale 🔹 $BCH — 21Shares, BitI just finished what I was doing this morning, and took the opportunity to check the market during a break in coffee. When I saw Jensen Huang's open letter yesterday, I wondered if NVIDIA might be using news to strengthen today. But when I checked the market, NVDA was still oscillating around 207, even slightly lowering at one point. The market reaction was much calmer than I expected.
This open letter itself is quite noteworthy, with a total of 25 tech companies jointly supporting open-source AI, including Microsoft, Meta, and IBM, and even Elon Musk publicly expressing support.
Many people's first reaction upon seeing this news was: Will models becoming more open-source affect AI companies' profitability?
But if you look at it from NVIDIA's perspective, I think the logic is quite the opposite.
The more open the model is, the more developers participate, the lower the barrier for enterprises to deploy AI, and the faster AI application implementation may accelerate. What truly determines NVIDIA's long-term value is not necessarily the leading model company, but whether the entire AI industry continues to expand.
After all, models can be open source, but the GPU, servers, and computing resources behind training and inference are not free. The fiercer the competition among AI vendors, the more iterative models and expanded deployments become, the demand for high-performance computing power may actually rise.
So in my view, Jensen Huang has always been betting not on a single model, but on the continuous growth of the entire AI ecosystem. As long as the industry keeps expanding, the demand for underlying computing power will rarely disappear overnight.
However, the pace of capital market watching clearly isn't that long.
Short-term funds are now more focused on earnings performance, whether next quarter profits exceed expectations, and whether each tech company's CapEx can continue to improve, rather than on what landscape the AI industry will ultimately develop in a few years. So even if the news is positive, I don't think it's surprising that the stock price doesn't immediately respond positively.
Recently, after the Kimi K3 became open-sourced, discussions in overseas AI circles have noticeably increased, with more and more people rethinking that the future path for AI development may not be limited to closed-source models. Competition between different routes may actually further accelerate the industry's development.
My understanding is that whether open-source models ultimately dominate or closed-source models continue to lead, as long as AI continues to become widespread, the computing power demands from training, inference, and enterprise deployment will most likely keep growing.
Therefore, I won't easily assume that the AI main theme has fundamentally changed just because Nvidia has fluctuated around 207 in the short term. More often, the market is digesting expectations, trading sentiment, and waiting for new earnings confirmation.
Of course, this does not mean the stock price will only keep rising. Short-term fluctuations are still affected by earnings reports, policies, and capital sentiment, so I won't blindly chase highs, but will continue to monitor subsequent earnings realization and CapEx data. If these core indicators do not show a clear weakening, I prefer to interpret the recent volatility as a market repricing rather than that the long-term logic has ended. $NVDA $IBM the radar flagged 26 setups this week before they moved, the tape already settled every single one. gap between "saw it" and "played out" is closed, here's the tape.
split: 11 carried, 15 faded, average outcome -9.9%. but of those 15 faders, 11 were already tagged overheated/high risk before they dropped. that's not the radar missing, that's the radar calling the flush before it happened. un1, WISHBONE, POW all got flagged MEME RUNNING [high risk] and then went to zero, exactly the outcome the tag warned about.
the carry side had a clean pattern too: real squeeze mechanics won. $EUL and $RIF got tagged SHORT SQUEEZE / SHORTS IN CONTROL [med risk] and ran 52%+. $BOP was flagged high risk too but caught a genuine meme wave, +66%. so high risk doesn't mean fade, it means volatile in either direction, the tag is telling you the range, not the outcome.
lesson of the week: overheated longs on thin books fade, squeezes with real positioning behind them carry. radar's still watching, next week's setups are already loading.
NFA$11.0M of $AAVE landed on exchanges this week across 12 venues while price just drifted up 5%, flat enough that nobody flagged it on the chart.
traced the two biggest legs: an old wallet (1+ yr) dropped $4.9M onto Coinbase Prime, and that stack came from 21Shares (21.co) right before. separately Wintermute moved $4.1M onto Binance.
could be an ETP issuer rebalancing and a market maker doing market maker things, could be supply lining up to get sold. inflow like this is possible sell pressure until proven otherwise. watching this one, not calling it 👀At the earnings call, Musk directly stomped on the entire robotics industry.
"99% of demo videos are either pre-programmed or remotely controlled by someone in the background."
Everyone in the industry knew about this, but no one exposed it until Elon Musk spoke up.
The line he drew was clear: a true general-purpose humanoid robot relies on natural commands to work on its own, without pre-programming or human intervention.
Currently, no one has done this.
Including Tesla itself.
The reality of Optimus,
All components require brand-new R&D, and there is no mature supply chain
The Fremont plant is ramping up capacity slowly
Chips are the bottleneck
The first batch of equipment is used only for internal data collection and is not sold externally
Musk is stepping on others, but at the same time, he's giving his own schedule a heads-up.
Physical AI is the real direction, but there is still a long way to go between "being able to work" and "performing and working." #财报观察员: Who can truly understand the real answer sheet from Google and Tesla this time? $DOGE 🚨JUST IN: The Trump team has moved $16.91 MILLION in $TRUMP tokens to Fireblocks custody wallets.
These wallets have previously forwarded $TRUMP to BitGo.
Over the past five months, the team has sent out 48.25 MILLION $TRUMP worth $172.4 MILLION across three separate batches. I recovered the chip structure from April, and obviously the gap between 76,000 and 80,000 has been partially filled. However, the stacked chips of 61k and 63k have reached their peak, which is quite interesting
1. The concentration of massive shares may be a historical bottom, with strong support. Selling pressure cannot be broken, and it is caught by heavy turnover
2. If it breaks below and cannot be recovered in a short time, it will become the strongest resistance level in this bear market, with massive volume trapping chips suppressing the market, potentially triggering panic selling of chips above 80,000, and the market will move to the next bottom consensus zone to reconstruct the bottom
So I believe now is the real turning point for the market. #财报观察员: Who can truly understand the real answer sheet from Google and Tesla this time? $BTC 📊 CoinGecko 24h Gainers Deep Review: Which Are Real Opportunities and Which Are Liquidity Traps?
Today's gainers list looks lively, but in essence, it's a mix of "listing-driven + Meme sentiment + new coin/airdrop liquidity + low liquidity anomalies." Blindly chasing gains based on price increases is very likely to hit a trap.
Based on exchange coverage, contract depth, and event catalysts, I have re-divided these 10 coins into three tiers:
🥇 Tier 1: Real Events and High Liquidity (Key Focus)
$EUL (+65.1%): The strongest listing driver! Officially launched on Upbit KRW spot on July 26, combined with mainstream coverage from Binance/Coinbase, the event is the most solid.
$KAITO (+25.1%): InfoFi narrative leader, with very strong spot and contract liquidity, but note the expectation of a large unlock in August.
$ESP (+16.8%): Post-airdrop TGE market, the core focus is whether the airdrop selling pressure can be effectively absorbed by the market.
🥈 Tier 2: Event/Narrative Short-term Play (Participate Cautiously)
$CROSS: Game "Frost Kingdom" global launch ecosystem driver.
$BOME / $TROLL / $ASTEROID: Pure Meme propagation and sentiment indicators, spreading fast but fundamentals are relatively weak.
🥉 Tier 3: Low Liquidity Warning (Recommended to Avoid)
$PONS / $GRX / $UWU: Although gains look good, trading volume is extremely low (e.g., $UWU only 7.3K, $GRX only 584K), with weak mainstream exchange coverage, making it very easy to encounter a "can buy but can't sell" liquidity crisis!
💡 Core Logic: For short-term trading, don't just look at who gained the most; the key is "whether there is real catalyst, whether mainstream CEX support exists, whether contract liquidity is present, and whether there is an upcoming unlock selling pressure."
Which one did you focus on today? Let's chat in the comments 👇It was just 💥 the last struggle
I don't believe you can keep pulling like this
A 50,000 USD position went all-in to short
The dog farm quickly sold the stock
I'm going to sleep
Wake up and clear the groceries right away
——
$SHIB The weekly major trend has not truly reversed
Although prices have rebounded from their lows,
But it still lags below the MA20
MACD is just a weak fix
This wave is more like an oversold rebound
Once the chasing funds can't hold on,
Whatever you pull up, you might just smash down
——
BTC is oscillating near 64,800
64,000 is the short-term dividing line between bulls and bears
Breaking below is easy to keep pulling back
However, ETFs have seen net inflows for several consecutive days
There is still capital holding the market below
So you can be bearish
But the rebound should not be treated as an unresistible bonus rally
——
$ETH overall performance is clearly weaker than BTC
In the preliminary funding data,
ETH's funding rate once turned negative
Option funds are also more inclined to downside protection
This indicates that market confidence in ETH's rebound remains insufficient
BTC just needs to weaken
ETH is very likely to amplify volatility
——
$LAB Now only around $0.15
Seven consecutive days of decline of more than 13%
Compared to a month ago, it has dropped by nearly 99%.
The previous destruction and the project team's calls
There has been no real restoration of market confidence for now
On top of that, there has been pressure to unlock tokens recently
The rebound seems more like giving trapped investors a chance to escape
This market is indeed on the bearish side
But 20 times the price goes to sleep separately
It's best to set stop-loss points
Don't end up with you waking up to pick up the groceries
Instead, the dog farm collects your position in the middle of the night
#财报观察员: Who can truly understand the real answer sheet from Google and Tesla this time?
#多数党领袖称CLARITY休会前难通过 Turning Point for South Korea's Memory Industry? Concerns and Changes Amid the AI Boom
On July 26, 2026, in the early trading session of the Seoul stock market, shares of Samsung Electronics and SK Hynix both opened higher but quickly narrowed their gains. Just the day before, the two companies announced they had signed a chip supply and technology cooperation framework agreement worth 1,375 trillion Korean won (approximately $940 billion) with a major U.S. tech giant. Some market participants interpreted this news as a major victory for South Korea's semiconductor industry, but a sober look at current data and industry logic reveals that beneath the surface prosperity, structural contradictions are rapidly accumulating.
1. Capacity Expansion Outpaces Market Demand Absorption
According to the latest statistics released by South Korea's Ministry of Trade, Industry and Energy on July 24, semiconductor exports in the first half of 2026 reached $68.7 billion, a year-on-year increase of 12.3%, but the growth rate has clearly slowed compared to 28.6% in the same period last year. Among these, memory chip exports still accounted for 62%, but contract prices for DRAM and NAND Flash have remained flat for three consecutive months.
One core aspect of the agreement is raising Samsung and SK Hynix's monthly HBM (High Bandwidth Memory) production target from the originally planned 130,000 wafers by the end of 2027 to 190,000 wafers, representing a 46% increase in capacity. However, major global AI chip customers—NVIDIA, AMD, Broadcom—reported in their Q2 earnings in mid-July that inventory turnover days rose to 98, 87, and 92 days respectively, all higher than the 75-80 day range in the same period last year. Downstream customers' willingness to stockpile is marginally weakening.
2. Mismatch Risk Between Non-Binding Agreements and Rigid Capital Expenditures
The signed document is not a long-term purchase contract with penalty clauses but a memorandum of understanding covering technology roadmaps and capacity reservations. Yet, Samsung Electronics announced on July 22 an additional equipment investment of 4.2 trillion Korean won for the P4 production line at its Pyeongtaek plant; SK Hynix confirmed on July 20 that the construction period for the Cheongju M15X plant has been shortened from the originally planned 32 months to 26 months. These are irreversible physical capital investments.
According to the Bank of Korea's "Corporate Investment Intention Survey" released on July 27, the semiconductor industry's equipment investment execution rate in Q2 has reached 78% of the annual budget, compared to 63% in the same period last year. The front-loading of investments is significant, but at the same time, global cloud service providers' capital expenditure growth slowed from 34% in Q1 to 22% in Q2. A Morgan Stanley report on July 23 pointed out that the HBM supply-demand gap is expected to narrow from 18% this year to 4%-6% in 2027, and if the expansion plans are fully implemented, an 8%-10% supply surplus may occur in 2028.
3. Increased Vulnerability from Exchange Rates and Foreign Capital Flows
The Korean won to U.S. dollar exchange rate was 1 USD to 1,378 KRW in early trading on July 27, near the low range since October 2022. Foreign investors have net sold in the Korean stock market for nine consecutive trading days, totaling 2.3 trillion KRW, with over 70% of net outflows from the semiconductor sector. Data from the Financial Supervisory Service of Korea shows that from July to date, foreign ownership of Samsung Electronics shares dropped from 34.1% to 32.7%, the largest single-month decline since 2021.
The depreciation of the won and capital outflows form a negative feedback loop. The 5-year CDS (credit default swap) spread for Korea rose to 47 basis points on July 26, up 12 basis points from a month ago, reflecting a repricing of South Korea's sovereign credit risk in international markets.
4. Real Demand Absorption Capacity Faces Tests
The demand growth for HBM from AI servers is undeniable, but bottlenecks in commercial monetization are emerging. Microsoft, Google, and Amazon reported in mid-July that AI-related business revenue accounted for 5.2%, 4.8%, and 3.9% of total revenue respectively, while corresponding capital expenditures accounted for a high 18.7%, 16.3%, and 14.2% of revenue. The investment return gap remains wide.
If from the second half of 2026 to the first half of 2027, the AI application revenue growth of major North American tech giants continues to lag behind capital expenditure growth, companies will inevitably reassess their procurement budgets. The Korea Development Institute (KDI) warned in its "Supplementary Economic Outlook Report" released on July 25 that if major customers lower their 2027 procurement forecasts, the idle capacity cost for South Korean memory companies could reach as high as 90 trillion KRW annually, equivalent to 32% of South Korea's current account surplus last year.
Conclusion
What the two South Korean memory giants have now is more like an entry ticket requiring a huge upfront stake. The shortened expansion cycle, the conversion of non-binding agreements into rigid expenditures, pressures from exchange rates and capital flows, and uncertainties in downstream commercialization progress together form a complex picture similar in logic but different in path from Japan's semiconductor industry in the 1990s. Physical capacity expansion is easy; sustained realization of industry value is difficult. When the tide recedes, who is swimming naked may become apparent even before the end of 2027. Today's South Korean semiconductor industry stands at a peak, but the mountain winds are biting.$CATI is trying to recover after a sharp sell-off and has already formed a decent rebound structure. The price climbed from $0.03619 to almost $0.03985 before entering a healthy pullback. It is now trading around $0.03845, where buyers are attempting to build support.
📍 Entry Price (EP): $0.03830 - $0.03850
🎯 Take Profit (TP): • TP1: $0.03920 • TP2: $0.03985 • TP3: $0.04050
🛑 Stop Loss (SL): $0.03770
Holding above the current support could open the door for another move toward the recent high. Wait for bullish candles with increasing volume before adding larger positions.
Let's go $CATI 🚀
#EarningsRealityCheck #KoreaAIChipPush #ETHExitQueueZero $SOL Solana Absorbed $1.41 Billion in Stablecoins This Week, 3.7 Times the Net Growth of the Entire Market
The supply of stablecoins on Solana reached $16.48 billion, a 9.34% increase this week, equivalent to $1.41 billion in new capital flowing into the chain.
🔸 Meanwhile, the total market capitalization of stablecoins only increased by $383 million, meaning Solana's liquidity is being drawn from elsewhere, not just through overall expansion.
🔸 The structure is also changing: USDC now accounts for only 47.1% of the stablecoin supply on Solana, while other assets (USD1, USDG) reached a record high of $4.8 billion.
👉 This is a very strong signal for Solana. The influx of stablecoins into the chain is not just speculation but real capital for DeFi and payment applications to function. The diversification of stablecoins also shows that the ecosystem is maturing. This is a different story from previous bull runs; it focuses on real liquidity and utility rather than memecoins.
💬 Do you think stablecoins are the best measure of a blockchain's true health?
News is for reference, not investment advice. Please read carefully before making a decision.Two hours ago, BUB was just an ultra-early-stage project with "shallow liquidity but temporarily scattered chips"; Now, that judgment has failed. Its price dropped from about $0.0001624 to $0.000002384, and main pool liquidity dropped from about $31,600 to about $2,740. In the past hour, there were 1,456 sell and 235 buys. Even if the liquidity certificates still show that all locked, additional issuance, and freezing permissions have been revoked, the funds in the pool that can truly support trading have collapsed, and I will stop observing. BUB contract: 4FaSuBUp15t9Qiar9MdpaspkZJU5RK6A3QLnybNCpump https://dexscreener.com/solana/J1GuZspgz3kxJqgngTGsR5QyJioSLAoZnApFd2yvtVsR HBULL temporarily different. It currently has a market value of about $0.001695, a market cap of about $1.7 million, main pool liquidity of about $126,600, and 24-hour trading volume of about $1.1 million. Within two hours, the price fell by about 10%, but trading volume did not disappear; About 99.97% of the main pool liquidity certificates are locked, and the rights for additional issuance and freezing have been revoked. I still only treat HBULL as a regular observer, since one address holds about 25.70% of the tokens. The project team claims this is a pledged vault, but I haven't seen it yetNvidia and SK Group announced over $500 billion in AI infrastructure plans, but the crypto market reacted lukewarmly, BTC and ETH showed divergence, and altcoins were generally under pressure.
Does this mean that the spillover effects of AI narratives on the crypto market have been fully priced in, or is the market waiting for clearer signals of capital flows?
- Event facts: Nvidia and SK Group jointly announced an AI project. SK Telecom will build a 2 GW AI data center using Nvidia Vera Rubin chips and SK Hynix HBM4 memory. SEC filings show that SK Telecom plans to increase AI data center capacity to 15 GW by 2035. The total project valuation exceeds $500 billion.
- Market Structure Changes: After the announcement, BTC fluctuated narrowly around $105,000, ETH weakened relative to BTC, and altcoins overall declined. This indicates that the long-term benefits of AI infrastructure have not directly translated into demand for crypto risk assets. The market may interpret this event as "traditional tech capital continuing to flood into AI on a large scale," rather than "accelerating the integration of AI and crypto."
- Expectations and Repricing: Previously, some market participants anticipated that the large-scale expansion of AI infrastructure would spill over into the crypto market through computing demand, tokenization, or decentralized computing networks. Current price action shows that this spillover effect is either priced in advance or has yet to form a verifiable transmission mechanism. The market is reassessing the risk premium of "AI concept coins," especially those projects that rely on short-term narratives rather than actual on-chain activity.
- Positioning Behavior and Derivatives Risk: From the derivatives market perspective, BTC perpetual contract funding rates remain in the 0.01%-0.02% range, with no significant increase, indicating that bulls have not increased their positions due to this news. ETH options implied volatility has slightly declined, indicating a reduced market expectation of ETH's short-term volatility. Altcoin futures open interest has declined, suggesting speculative funds are pulling out. If BTC fails to break through the $108,000 resistance, it could trigger long liquidations and intensify the pullback.
- Multiple paths and conditions: If clear "AI+crypto" collaboration cases emerge in the coming weeks, such as decentralized computing networks receiving official support from Nvidia, or AI data centers adopting tokenized computing power, AI narratives may be reactivated. At that point, BTC needs to hold above $110,000, and ETH needs to break through $4,000 for altcoins to see capital flow back.
- Bearish risk and conditions: If AI projects progress smoothly but the crypto market does not directly benefit, the market may further compress the valuations of AI concept coins. If BTC falls below $98,000 (near the current 200-day moving average), it could trigger broader deleveraging, with ETH and altcoins seeing even greater declines.
- Conclusion: Nvidia and SK's $500 billion AI plan is priced in the crypto market as a "tech stock boom" rather than a "crypto catalyst." BTC is currently relatively strong, but ETH's weakness against altcoins suggests that the market is skeptical about the spillover effects of the AI narrative. Under this structure, risk appetite in the derivatives market is declining. In the short term, more attention should be paid to whether BTC can hold key support rather than chasing the AI concept.
Risk warning: The expansion of AI infrastructure may continue to divert attention from the crypto market rather than generating incremental capital.
$BTC $ETH $AISpot $LINK ETFs have recorded capital inflows for three consecutive days...... For the first time since April.
What is brewing within the Chainlink ecosystem.
Spot $LINK ETFs have just experienced three consecutive days of net inflows—the first time since late April 2026.
These products ended the week with a net inflow of +$2.98M and now hold 1.79% of the circulating supply of $LINK.Is $DOGE preparing for a rally at the end of the $BTC bull market?
I've noticed that since the start of the bear market, every time Bitcoin and the market rebound at the end, $DOGE has systematically surged, and when this happens, a sell-off usually follows
$BTC In my view, it hasn't reclaimed its highs yet, so Dogecoin may have good upward momentum$Short-term (a few hours to 1-2 days) is not a good time; the risk outweighs the opportunity. The reasons are as follows:
· Resistance is evident above: the 1-hour chart shows prices are just below 1,922.68, which is both the intraday high area and close to the SuperTrend resistance level (1,904.67). More importantly, the estimated strong discount price is at 1,892.62, meaning if the price falls below this level, bulls will largely passively close their positions, triggering an accelerated decline.
· Open interest divergence: The open interest (OI) shown in your screenshot clearly declines during price rebounds (from 1.511 billion to 1.478 billion). This is a typical case of short positions closing out a rebound, rather than new long entries. Such rallies often have poor sustainability.
· Funding rate is neutral: Recently, the rate has fluctuated slightly around 0, indicating there is no strong bullish sentiment in the market and a lack of fuel for a trending trend.
Specific operational suggestions:
· Want to go long: At least wait until the price breaks through 1,928 (24-hour high) with increased volume and holds steady, or if it fails to break below 1,890-1,900 before reconsidering, with stop-loss set below 1,880.
· Prefer short-term shorting: If the price struggles again near 1,925, you can take a light position and try shorting, with a stop loss at 1,935, targeting 1,900 and 1,890 first.
· Special reminder: The leverage you used does not show. If you use high leverage (above 10x), going long now is very dangerous. Once it breaks below 1,900, the support below is very weak, and it could directly move toward 1,876 or even 1,865.
Simply put: going long now is "going against the small trend," and the profit-loss ratio is not cost-effective. It is recommended to either wait for a breakout confirmation or a deep pullback; temporarily observing or lightly shorting is more reliable. $$AAVE is demonstrating strong bullish momentum on OKX today, pushing up +5.65% to trade around $97.21 with a 24-hour high of $98.12.
After testing support near its recent low of $87.50, the price has broken back above key short-term moving averages (MA5 at 94.95, MA10 at 92.98, and MA20 at 93.53), signaling a healthy reversal on the daily chart as buyers target the psychological $100 mark.
#DailyOrbit @OKX中文 Next week is tough for the US stock market, with Microsoft, $META, Amazon, and $AAPL all reporting earnings, plus the Fed, GDP, and PCE all squeezed into the same week.
Google and Tesla have already set the tone for the market. Even with decent earnings, stocks still fell because investors are growing increasingly impatient with "continuing to pour money into AI." Revenue growth is only part of the story; how long it takes to recoup the money spent is the key focus for pricing after earnings.
On Wednesday, I'll first look at Microsoft's Azure growth. The market expects revenue of $87.67 billion and earnings per share of $4.24. Meeting these numbers isn't enough; guidance on Azure and AI investments will determine the after-hours direction. If Meta continues to significantly increase capital expenditures, advertising revenue must be strong enough, or the stock price will struggle.
On Thursday, I'll watch Amazon AWS. If growth holds up to expectations, Nvidia, Micron, and SK Hynix in the compute and storage chain can catch a break; if AWS slows, the hardware stocks that rose the most early on will be the first to be cut. For Apple, I only care about China sales and next quarter's guidance; no matter how much AI is mentioned at the launch event, it comes second.
My position will be light. The market won't reward "just okay" next week; if earnings, guidance, or cash flow fall short in any area, high-valuation tech stocks could see direct further declines.$BTC
My pattern plays out again...
Friday weakness. ✔️
Weekend strength. ✔️
This time wasn't different.
The weekend should close above Friday's candle close, bringing the pattern to 12 out of the last 13 instances.$BTC #EarningsRealityCheck Back in 2018, hundreds of domestic exchanges were clustered together, charging coin fees, issuing air assets, and selling customer losses—all sorts of tricks. Now, in 2026, the wave of bankruptcies has arrived—aside from those who just fled, the main problem is that matchmaking deals no longer make money, retail investors have evolved, and regulations are getting stricter. Large firms compete fiercely over services, while smaller firms simply can't survive.
If the crypto world truly wants to revive itself, it must abandon all old tricks and focus on one thing: turning good real-world assets—like US stocks and government bonds—into low-cost, high-efficiency Web3 assets on-chain. This is not something a diploma trader can handle.
---
Looking back at the evolution of finance over the past few centuries:
· The bank has → money that can circulate
· The securities market has → corporate equity that can now be moved
· ETFs have emerged→ allowing a basket of assets to be traded at low cost
· Internet brokerages have emerged→ ordinary people can now buy global assets
· The emergence of blockchain → aims to enable global assets to circulate borderless 24×7 hours a day
The true value of Web3 has never been in building more casinos, but in becoming the next generation of financial infrastructure.
Exchanges that survive aren't about who can create more speculative opportunities, but about who first masters TradFi, carves out a trick on it, and makes Wall Street people take a second look at Web3—that's real skill.
#多数党领袖称CLARITY休会前难通过
#财报观察员: Who can truly understand the real answer sheet from Google and Tesla this time? Last night after work, I watched the market for a while. I originally just wanted to see if there was a chance to reduce some positions, but I ended up seeing many people discussing the storage sector. The comment section still had the same saying: “When cyclical stocks have the highest profits, their P/E ratios are often the lowest.” Of course, I agree with this, but I think it only explains the surface and doesn’t answer the question I really care about—how much longer this cycle can last.
Companies like SK Hynix, Micron, Snowflake, and Samsung currently have relatively low valuations. Many people's first reaction is “Don’t touch them, the cycle is at its peak.” But the market actually already knows that the current profit margins can’t be maintained forever, which is why these companies don’t have very high valuations. The real disagreement isn’t whether the cycle will end, but how many more years the supply-demand tightness can last.
If the industry peaks this year and then immediately enters oversupply, price declines, and margin contraction, then the valuations that look cheap now could very well be classic value traps. But if supply tightness can continue for two or three more years, the cash flow these companies accumulate before the cycle truly reverses might be far more than what the current market valuations reflect.
I personally prefer to focus on feedback from companies in the supply chain rather than just watching P/E ratios. At least the information released by several suppliers currently leans toward the latter. SK Hynix mentioned that supply tightness for some memory products might continue until the end of this decade; Samsung, although more conservative, also believes that noticeable supply tightness will last at least until 2027.
Of course, management’s words can’t be trusted 100%, since everyone wants the market to have more confidence in them. But the information they hold is indeed much more complete than that of outside investors, such as customer contracts, equipment purchases, wafer planning, and packaging capacity. These will ultimately be reflected in the pace of capacity expansion, not just slogans.
Another common concern is whether customers will reorder repeatedly.
If the actual installation rate is low, it could mean customers overestimated demand, or it could be due to limited capacity, so everyone locks in supply early. Looking at order quantities alone makes it hard to judge which is the case. I think contract terms are more worth studying.
In this cycle, many customers are willing to sign multi-year agreements, accept price floors and ceilings, pay prepayments, and even share the funding for new capacity construction. From a business logic perspective, if demand were only short-term, few would lock resources years in advance or willingly bear expansion risks for suppliers. I think this is more meaningful than order numbers.
HBM is also an area I’ve been paying close attention to. Its biggest difference from traditional DRAM is that new supply is no longer as easy to release. HBM consumes more wafer capacity, requires higher yields, and advanced packaging further limits expansion speed. From HBM3E to HBM4 and HBM4E, manufacturing complexity continues to increase, so the newly added capacity is likely absorbed by the higher manufacturing intensity per product rather than simply turning into more shipments.
The same logic applies to TSMC and ASML.
The more advanced AI chips are, the more they rely on leading processes, EUV equipment, and advanced packaging working together. Whether it’s TSMC building new fabs, ASML delivering equipment, or customers completing capacity expansions, the whole process can’t be done in just a few quarters. Supply will definitely increase, but the speed of increase may not be as fast as the market imagines.
On the demand side, some worry whether Nvidia, AMD, and Broadcom will face pressure in the future because big companies like Meta, Google, Amazon, and Microsoft are currently very aggressive in purchasing AI accelerators and custom chips. If data center construction cools down, inventory, prices, and margins could all be affected.
I think this risk does exist, so I won’t keep my positions fully loaded just because I’m optimistic about the cycle. But on the other hand, AI computing demand itself is changing. Training still requires massive computing power, inference demand is expanding, and agents as well as more custom chip projects continue to add new loads. Even if the growth rate of a certain chip category slows, new demand sources might continue to push the entire construction cycle further out.
In the past two years, I think the biggest characteristic of the supply chain is that bottlenecks keep moving. At first, everyone fought for GPUs, then HBM and advanced packaging became the constraints, and later it was optical modules, power, cooling, and data center capacity. The constantly changing bottlenecks themselves indicate that the entire industry is still expanding on multiple physical layers, not yet reaching a very clear endpoint.
My understanding is that semiconductor cycles certainly won’t disappear. Supply will eventually catch up with demand, prices will return to normal, and margins will decline. What’s really worth comparing is whether the market’s current valuations, which reflect expectations about the cycle’s end time, align with actual contract durations, expansion speeds, and supplier feedback.
If AI demand slows earlier than expected, then these low valuations might indeed be warning of risk; but if physical bottlenecks in the supply chain remain unresolved and new capacity release can’t keep up, then what the market is underestimating might not be these companies’ profitability but how long the entire boom cycle can last.
So I’m not blindly overweight just because valuations are low, nor will I avoid the sector outright just because of the saying “low P/E in cyclical stocks means the top.” I prefer to track industry data while adjusting my positions. After all, the cycle will end, but it might not reverse quickly next year as the market pricing suggests. Leaving some room in trading is much more comfortable than betting on a single direction.
#韩国存储双雄获AI双巨头大单
$SKHYNIX $MU #Ethereum validator exit queue has dropped to zero
I discovered a very magical phenomenon.
The Ethereum validator exit queue has been directly cleared, but ETH wanting to be staked has to wait in line for more than 40 days.
On one side, no one wants to leave; on the other, new money is scrambling to get in. Isn't this signal obvious enough?
My view is simple: this wave is not retail investors playing, but institutions bottom-fishing and locking up. Big holders like BitMine have staked 70% of their ETH in one go, clearly not planning to sell in the short term. Plus, with continuous inflows into ETFs, the circulating supply in the entire market is quietly shrinking. Many people anxiously watch the candlestick charts daily, thinking ETH can't rise, but look at this data—no one wants to sell, and new money is still queuing to enter. This itself is a pretty strong signal.
Speaking of Bitcoin, this staking wave has actually brought indirect benefits to it. Previously, people worried about "ETH crashing dragging down BTC," but now that the exit queue is zeroed out, ETH's selling pressure has basically disappeared, and Bitcoin has lost one of its biggest "ball and chains." More importantly, ETH staking locks up a large amount of liquidity, effectively reducing the total market supply. Bitcoin's supply is already decreasing after the halving, and with ETH also exiting circulation in large amounts, both sides are shrinking supply, which is a double support for the price.
Of course, risks are not absent. Validators are too concentrated, and large nodes have too much influence, which is not good for decentralization. But given the current situation, I don't think there's a need to be too pessimistic. After this 40-plus-day queue is digested, market supply will be tighter.
On a side note, meme coins are crazy today. Could a wild bull market be coming?!
$BTC $SHIB $DOGE #韩国存储双雄获AI双巨头大单
Quarterly profits spilled 150 trillion KRW! SK Hynix's financial report explosion—whose 'AI bubble theory' was shattered?
In a couple of days (the 29th), SK Hynix will officially release its Q2 financial report. According to the latest forecasts from 14 Yonhap Infomax institutions, SK Hynix's Q2 operating profit is expected to surpass 64.09 trillion KRW—this quarter's profit alone exceeds last year's full-year 47.2 trillion KRW by a full 17 trillion KRW!
Including 37.61 trillion won in the first quarter, SK Hynix's operating profit in the first half of the year alone surpassed the 100 trillion won mark. If Samsung Electronics' DS division also includes the Q2 forecast of 89.4 trillion won, the combined operating profit of these two major Korean semiconductor giants in the second quarter alone will exceed 150 trillion won.
Seeing this set of exaggerated financial data, honestly, those who previously claimed "AI investment can't break even" and "AI capital spending is a bubble" were all left speechless.
Behind this set of data lies the core truth behind global tech capital flows:
The huge capex (capital expenditure) invested by tech giants on AI infrastructure is not a bottomless pit, but precisely converted into trillions of yuan in fiat cash flow on the books of sellers in storage and computing power. HBM high-bandwidth memory is not telling a story, but the world's most competitive and profitable physical commodity today.
For the crypto market, this explosive financial report is of immense significance.
Recently, US tech stocks pulled back, causing many retail investors in the crypto market to panic. But the quarterly profits of 150 trillion won from South Korea's two major chip giants directly prove that the cash flow and self-generating ability of the AI computing power industry chain are unbreakable.
When the computing power of the physical world becomes the most profitable asset, those junk knockoffs in crypto that rely on unlimited token inflation to print money will only be rapidly eliminated; On the contrary, infrastructure that can truly connect to physical computing networks, promote computing power tokenization (such as Gensyn and Virtuals protocols), and provide on-chain computing power revenue distribution is undergoing a value reevaluation of traditional capital.
My conclusion: With SK Hynix's earnings report officially released on the 29th, it is highly likely to dispel macro capital's last wait-and-see sentiment toward tech stocks and computing power Capex.
Do you think SK Hynix's earnings report on the 29th exceeded expectations and will trigger a new round of counterattacks in tech stocks and the crypto AI sector? Let's talk in the comments.Bitcoin is not safe here.
The whole time it is under $66,000, there is a clear pathway to the Realised Cap at $54,000,
The consolidation under $66,000 only becomes a deviation once Bitcoin has reclaimed that level again.
If it does not reclaim, then it becomes a potential bearish consolidation that leads to deeper lows.
With the current corrective price action, this cannot be ruled out as impossible.
There are a few key things to note however.
This bottoming structure is almost identical to 2022.
And we were correctively moving back then also, with a very similar weekly candle to what we are getting right now.
That candle and the weeks that followed sent the timeline into a massive "$12k is coming" frenzy...
But it did not come, and Bitcoin began impulsing out of thin air.
We also had a bullish divergence, and the same percentage of coins sitting in a loss.
In 2022 we spent 10 weeks below $18,000.
Right now, we are 7 weeks since we tagged below $60,000.
The similarities are uncanny$BTC In the days leading up to the announcement of its closure, crypto exchange BitMart saw its on-chain public asset reserves plummet, dropping from about $12 million on the 12th of this month to about $2.31 million on the 26th. Currently, there are only $1.89 million in assets on-chain: Ethereum about $815,000, Solana about $660,000, BSC about $362,000, Starknet about $37,000, and Bitcoin just about $17,000.$COIN's core contradiction is that its valuation logic is shifting from spot exchanges that rely solely on crypto trading cycles to infrastructure covering asset issuance and settlement, but the proportion of short-term fee income still determines cash flow stability.
Currently, the market views $COIN as an elastic amplifier of the crypto cycle, with its early underlying profit model relying entirely on commission commissions from buying and selling spot currencies like BTC and ETH in US dollars.
The driving factors are ranked as follows: the depth of financial infrastructure expansion beyond trading business, the speed of overall crypto market trading volume recovery, and the hedging effect of non-trading revenue against cyclical downturns.
The trigger for an upward scenario is that asset service boundaries successfully cross single spot trading. If its income growth in issuance and settlement exceeds traditional spot fees, the market will anchor the valuation system as a comprehensive financial infrastructure across the chain, thereby raising the valuation center.
The failure signal of this scenario is a devastating contraction in overall spot trading volume of crypto assets, making it impossible for infrastructure revenue to fill the fee gap.
The trigger for the downward scenario is that business expansion fails to change the dependence on revenue structure. When a bear market cycle causes trading activity to remain sluggish, the profit base relying solely on fees will once again drag down overall financial performance.
The downward scenario is signaling the failure of the downward script: the proportion of non-trading business revenue has broken through a critical threshold, causing the decline in transaction volume to no longer drag down the company's total revenue in tandem.
The most important variable to watch over the next seven days is the marginal trend of crypto spot trading volume in total revenue contribution and the pace of business advancement for non-trading products.
#财报观察员: Who can truly understand the real answer from Google and Tesla this time? #SPCX因星舰发射与解禁引发多空分歧#Korean Storage Giants Secure Major AI Orders from Dual Titans
AI computing power enters the order fulfillment phase
The AI industry chain welcomes another major positive development
South Korea's storage chip leaders Samsung Electronics and SK Hynix simultaneously secured long-term cooperation orders from AI giant Anthropic. Meanwhile, NVIDIA announced a $1 billion investment in South Korean internet giant Naver to build an AI data center and further expand cooperation with SK Group. This series of moves indicates that the global AI competition has shifted from model competition to infrastructure competition.
What truly deserves attention is not just one or two orders, but the fact that global tech giants continue to increase capital expenditures.
Whether it's OpenAI, Anthropic, Meta, or Microsoft, they are all continuously increasing AI computing power investments. HBM high-bandwidth memory, GPUs, servers, and data centers have become the scarcest resources in the AI industry chain.
Previously, the market once worried about a slowdown in AI investments, but recent news breaks those concerns. Intel raised its earnings guidance, Qualcomm announced price hikes, and now Samsung and SK Hynix have secured long-term orders again, all indicating that AI demand remains strong and has gradually expanded from GPUs to storage, networking equipment, and data centers across the entire industry chain.
I believe this means the AI market is entering its second phase.
The first phase was driven by expectations—whoever told the AI story rose; the second phase competes on orders, performance, and capital expenditures. Whoever can continuously secure AI orders has a better chance of market revaluation.
For the capital market, $NVDA, $AMD, $AVGO, $TSM, and other computing power and semiconductor leaders remain the core beneficiaries, while Samsung Electronics and SK Hynix will continue to benefit from the supply shortage of HBM and growing AI server demand.
For the crypto market, this also sends a positive signal.
As AI infrastructure continues to expand, AI sector tokens are expected to keep attracting capital attention. Projects like $TAO, $FET, and $RENDER remain important representatives of the AI sector. Meanwhile, the improved AI industry outlook also helps enhance overall market risk appetite, indirectly supporting mainstream crypto assets like $BTC, $ETH, and $SOL.
What the market really needs to focus on is not whether the AI concept can still be promoted, but whether global tech giants continue to invest.
As long as data centers are still being built, GPUs are still being continuously procured, and HBM remains in short supply, it means this AI industry cycle is far from over. The true beneficiaries in the future will not only be model companies but the entire computing power industry chain and related assets developed around the AI ecosystem. Big Tech's earnings delivered the reality check the AI trade has been dodging. Alphabet and Tesla both reported, and both stocks sank, not on weak results (Google Cloud grew 82%) but on rising AI capex guidance. The market has flipped: spending on AI used to be rewarded as vision, now it's scrutinized as cost.
This is the same story that hit semis all week, viewed from the demand side. Investors aren't questioning whether AI is real; they're questioning the return on hundreds of billions in capex before the revenue catches up. For crypto it's a useful mirror: narratives get repriced the moment the market demands proof over promise. Risk-off today (BTC $64K) echoes that same "show me the ROI" mood bleeding across tech.
Just my read, not advice. The grand finale of the month, just on July 30th
The last few days of July, don’t be fooled by the lukewarm market. On Thursday (7/30) Beijing time, two major events packed:
2 AM, Federal Reserve decision. This time no one is betting on a rate cut—the probability is basically 0, the only disagreement is between "hold steady vs. raise 25 basis points." Two weeks ago, the chance of a rate hike was just over 10%, but with oil prices breaking $100 and the US-Iran conflict starting, now it’s surged to over 35%. The real highlight is the 2:30 AM press conference by Powell; he speaks little, gives no guidance, but one sentence can reprice September.
8:30 PM, Q2 GDP. The current tracking estimate is only 1.7%, even cooler than Q1. Growth is cooling down, inflation is still burning—stagflation vibes.
BTC is stuck at 65,000, sentiment is at freezing point (fear index 29), ETFs are still bleeding. On such a thin market, bad data will only cause bigger ripples.
🔴 Resistance 66,000-67,000
🟢 Support 62,000, 58,000 is the lifeline
My approach: no heavy bets on direction before data release, light positions waiting for the shoe to drop. #财报观察员：谁能看懂谷歌和特斯拉这次的真实答卷？ $BTC $ETH $SNDK ⚠️ Let's start with an unusual phenomenon: the S&P 500 barely fell, the Dow even rose, but the Nasdaq and several tech leaders clearly weakened. This is not a broad rally or a broad decline, but rather capital rearranging its seats. 📊 Latest closing data: As of 00:10 Beijing time on July 27, US stocks are closed for the weekend. The latest effective data is the closing at 04:00 on July 25: SPY: $738.93, +0.10%; QQQ: $684.23, -1.12%; DIA: $518.76, +0.48%; AAPL: $333.02, +3.53%; GOOGL: $319.74, +0.65%. MSFT: $381.70, +0.03% NVDA: $206.84, -0.92% AMZN: $232.11, -0.66% META: $595.19, -1.80% TSLA: $313.03, -2.08% 🍎 Apple is strong, but can't save the entire tech sector Apple rose 3.53% in a single day, with its stock price very close to its 52-week high of $334.99. However, at the same time, QQQ fell 1.12%, with Nvidia, Meta, and Tesla collectively pulling back. This shows that funds are not indiscriminately buying tech stocks, but are instead flowing into a handful of strong companies. Nowadays, the US stock market isn't just about "buying tech"—it's about choosing the wrong stock, and the index rise has nothing to do with you. 🔍 Watching next weekI originally just wanted to try Babylon, but ended up turning the test vault into a life-or-death situation
At first, I really didn’t intend to play this big. When I first entered the Babylon TBV testnet, I only created a small Vault, borrowed a bit of test assets, and the health factor was very safe. At that time, I even pretended to be rational, telling myself I was just experiencing the process and would never leverage.
But within ten minutes, I started to feel the position was too light. The test coins weren’t real money, so what was there to fear? So I gradually increased the borrowing limit bit by bit. Every time I clicked confirm, I felt like I wasn’t taking a risk but rather “improving capital efficiency.”
What gamblers are best at is finding a reason that sounds professional to justify their actions.
The health factor slowly dropped, but I actually liked it more and more. The closer the number got to 1, the more I felt this position was being fully utilized. Until BTC suddenly plunged, I realized that the so-called “high capital efficiency” translated into human language meant I was close to liquidation.
When the page showed a risk warning, I could have repaid some first. But I didn’t. I stared at the candlestick chart thinking, just wait a bit longer, a rebound will make it safe.
But the rebound never came, the health factor fell below 1 first, and the Vaults ahead directly entered the liquidation process.
At that moment, I finally stopped pretending. I started frantically recalling what I did when building the position: which Vault was ahead, which was behind, and whether liquidation would handle all BTC together?
It was only then that I truly realized the Vaults in TBV aren’t just random names on the page. Behind each Vault is an independent UTXO on the Bitcoin network. They don’t mix into a common fund pool but are separated one by one.
Liquidation isn’t about the platform taking as much as it wants but is executed in the preset Vault order, by complete Vault units. In other words, when I found splitting Vaults troublesome before, now that trouble has come, I realize that was actually prearranging my own “liquidation order.”
What impressed me even more was that throughout the process, BTC was never moved to another chain. The borrowing status, health factor, and liquidation conditions changed in external DeFi applications, but the native BTC remained locked on Bitcoin, not wrapped into another asset, nor handed over to any custodian.
Before, when I played DeFi, what I feared most wasn’t the market dropping but that when the market dropped, cross-chain bridges, custodians, and protocols would all fail together. TBV doesn’t solve gamblers’ love for leverage, nor will it let liquidation slide just because you insist otherwise.
But at least it separates two things:
I can be liquidated because my position is too heavy—that’s my judgment error; but I don’t need to give up control of BTC to others just to use BTCFi.
This time it was just a testnet, and the losses weren’t real money, but I was still thoroughly educated. Before, when building a position, the first thing I did was calculate how much I could borrow; now I first think about which Vault would be liquidated first if BTC suddenly crashes, and which Vaults can remain.
Babylon TBV didn’t make me quit leverage.
It just made me understand one thing:
You can keep gambling, but you can’t put the keys up as collateral together.
#Babylon #TBV #Bitcoin #BTCFi #DeFi
$BABY