Li Xiaolu timeline review: 16 years old Golden Horse Award → 40 years old with no roles to play. As long as a person keeps self-destructing, no matter how high the starting point, they can ruin themselves!

#财报观察员：亚马逊指引不及预期，股价却反涨9% After the market closed on July 30, Amazon delivered a strong earnings report with a conservative outlook for Q2. Revenue, profit, and AWS growth all exceeded expectations, but the Q3 revenue guidance midpoint was significantly below Wall Street consensus. The result? The stock surged nearly 10% after hours and continued to strengthen the next day. This is an unusual move during the recent tech giant earnings season! Alphabet and Meta were hit due to high spending or growth details, yet Amazon increased capital expenditure to $220 billion and was rewarded by the market with real money votes. What actually happened? Q2 revenue was $200.6 billion, up 20% year-over-year, exceeding the market expectation of about $196.5 billion. Operating profit was $27.5 billion, a 43% year-over-year increase. The real highlight was AWS: revenue of $42.2 billion, growing 37%, marking the fastest pace in the past 18 quarters, with an annualized run rate reaching $169 billion. Backlog orders surged to $496 billion. AI-related business and self-developed chip business each surpassed $25 billion in annualized revenue. Advertising business continued to grow at 26%. On the retail side, same-day and next-day delivery product volume increased by more than 40% in the first half of the year. The profit figures look exaggerated (diluted EPS of $5.75) mainly because the fair value change of the investment in Anthropic contributed $53.4 billion in non-operating income. Excluding this one-time factor, core operating performance remains solid. Why was the guidance below expectations? The company’s Q3 revenue guidance is $197 billion to $202 billion, corresponding to 9% to 12% year-over-year growth. Wall Street previously expected about $204 billion, so the midpoint is clearly low. Operating profit guidance of $22.5 billion to $26.5 billion is also somewhat conservative. However, management specifically explained two points: first, this year’s Prime Day was moved up to Q2, resulting in a higher comparison base for Q3; second, exchange rates had an adverse impact of about 80 basis points. Excluding the Prime Day timing difference, actual growth could be nearly 400 basis points higher. In other words, the guidance itself is clearly "technically conservative." At the same time, full-year capital expenditure was raised from the previous $200 billion to $220 billion. The main reasons are rising memory prices and AI demand continuing to exceed supply capacity. CEO Andy Jassy clearly stated that capacity may still lag demand in 2026 and 2027. Why doesn’t the market buy the pessimistic case? This is the most interesting part of this earnings report. Recently, the market has developed a conditioned reflex to AI spending: whoever dares to significantly raise Capex is suspected of longer return cycles and worsening free cash flow. Alphabet was hit after raising Capex, Meta was also hit while maintaining high spending. Amazon not only raised Capex but also turned free cash flow negative by $7.6 billion, which should have put pressure on the stock. The market instead chose to believe another logic: AWS growth is truly accelerating, driven by AI pulling the core cloud business rather than just spending money. The backlog nearing $500 billion means high revenue visibility for the coming years. When growth accelerates again, the market is willing to pay for "investment turning into demand" rather than just focusing on short-term cash flow. The contrast is clear. Microsoft reported Azure acceleration the day before while controlling capital expenditure, and its stock soared; Amazon proved with higher growth and clearer backlog orders that its spending is supported by demand. The paths differ, but the market’s conclusion is consistent—high spending is acceptable as long as growth materializes. A deeper signal This is not just a reaction to an earnings report but a market re-pricing of the AI infrastructure investment narrative. In recent months, investors have increasingly worried about "too much spending and too slow monetization." Amazon pushed back on this concern with real data: AWS growth returned to a high level, AI business has scaled, and chip business is starting to contribute. Spending is not a bottomless pit but a response to real enterprise demand. Of course, risks remain. Negative free cash flow, rising depreciation pressure, and high memory costs will continue to test the income statement in coming quarters. If Q3 guidance ultimately falls short, market sentiment could quickly shift. But at least on the evening of July 30, the market’s vote was very clear! It cares more about whether growth is accelerating again than whether guidance is perfect. For Amazon holders, the core message of this earnings report is actually simple: the AI story is shifting from "investment narrative" to "growth validation." When validation appears, the stock’s tolerance for guidance will significantly increase. In the coming quarters, the real focus should not be whether capital expenditure will be raised again, but whether AWS growth can remain high and backlog orders can continue converting into revenue. As long as these two points hold, the market will likely continue to give Amazon greater margin for error.

Today in the elevator, I overheard two women talking: "He earns 30,000 a month, but he's too honest." Brothers, the word "honest" nowadays is basically an insult in this context.

Code/Terminal AI Super Assistant! Still Googling error messages and writing scripts line by line? What a waste of time! Install this free CLI tool: Chat directly with AI in the terminal to modify code, fix bugs, and generate features with powerful context awareness. Coding speed increases by 10 times. Once you use it, you can't go back!

#PCE环比转负，GDP增速放缓至1.5% This is not a signal of recession; it’s the moment hunters should draw their swords. Yesterday, the BEA dropped two bombs simultaneously: June’s PCE price index month-over-month -0.1%, the first monthly decline since the 2020 pandemic. Core PCE month-over-month +0.1%, year-over-year down from 3.4% to 3.3%. Q2 annualized GDP growth at 1.5%, significantly slowing from Q1’s 2.1%, and below market expectations. Many have already started shouting that the soft landing has failed, recession is coming, risk assets run. I just want to say one thing: Those who look at data die in the details; those who look at structure live in the main upward wave. What really deserves attention is not that 1.5%, but what is the main cause dragging down GDP? Government spending decline, inventory reduction, huge surge in imports (AI-related equipment imports skyrocketed). What about the private domestic final sales that truly reflect domestic demand? +3.9%. More than double Q1’s 1.7%, the strongest since early 2023. Consumer spending annualized +3.2%, taking off from Q1’s almost flat 0.5%. Businesses and households are spending real money, government is shrinking, inventories are being digested, imports are restocking. This is called improved growth quality, not economic collapse. Inflation is even more interesting. PCE turning negative month-over-month is mainly due to oil prices falling after the temporary US-Iran ceasefire. Core PCE remains at 3.3%, still far from the 2% target, but the direction has shifted from stubbornly rising to slowly cooling down. The Fed just held rates steady at 3.50%-3.75% this week. After the data release, the market’s pricing for further rate hikes is almost zero, and expectations for rate cuts are starting to rise again. What does this mean for crypto? In the past two years, the biggest macro pressure was high interest rates + high inflation expectations. Now these two stones are starting to loosen. When real demand still exists, inflation begins to cool, and signals of peak rates become clearer, the pricing logic for risk assets will switch from defense to offense. Bitcoin, Ethereum, quality RWA, tokenized stocks, AI-related assets... These are far more sensitive to liquidity expectations than to the absolute GDP value. 1.5% GDP is not a disaster. The real disaster is stagflation with growth collapse + runaway inflation. What we see now is a combination of improved growth quality and easing inflation pressure. This is a classic mid-term feature of a soft landing. // To put it passionately: the market loves to panic when things look bad but actually aren’t. Because most people only read the headlines: “GDP slows! PCE turns negative! It’s over, it’s over!” Those who can really make money look at: • Is domestic demand broken? • Are businesses still investing? • Is inflation out of control? • Is the central bank continuing to tighten? The answer to all is “No.” So now is not the time to run for your life. It’s time to shift positions from “risk defense” to “trend capture.” Remember one thing: The cruelest part of economic data is never that it’s bad, but that it’s just bad enough to scare most people away first. The rest is left to those who truly understand the structure. To study, to position, to wait for the moment when everyone else finally reacts. By then, the price will have already completed the first leg!

Economic downturn, which sectors are still growing against the trend? Elevator advertisements for on-site services in high-end residential communities. All the technicians are good-looking, and several companies already have market values exceeding 100 million, ready to go public at any time. While ordinary people are still complaining about the poor environment, some have long found new paths.

RWA transforms from a narrative into assets in your account: when stocks, funds, and commodities all start living in the same OKX account OKX's official Chinese channel recently released a set of data: as of July 21, RWA growth mainly came from funds +$15.6 billion, commodities +$5.3 billion, and stocks +$2.2 billion. In short — real-world assets are truly being integrated into the digital asset market. This is not just a slogan. BUIDL (BlackRock tokenized treasury bond fund) continues to expand on the institutional side, and OKX is promoting its use cases as collateral; on the stock side, there are Unified Tokenized Stocks; commodities and funds are also appearing across different product lines. In theory, one OKX account can simultaneously hold crypto assets, tokenized US stocks, tokenized treasury bond yields, and even more real-world asset exposures in the future. What this changes is the boundary of assets. In the past, crypto accounts and traditional financial accounts were two separate worlds. If you wanted to allocate US stocks, you had to open a separate brokerage; if you wanted treasury bond yields, you had to buy money market funds or treasury ETFs. Now, at least in qualified regions, these assets start sharing the same login portal, the same capital system, and the same risk control and trading tools. 24/7 access, USDT settlement, and the ability to combine with crypto assets — once users get used to these features, it's hard to go back. Of course, reality is not that ideal. Regulatory fragmentation still exists (US and European users cannot access some products), liquidity is clearly tiered, and legal rights (voting rights, dividend handling, bankruptcy isolation) vary by issuer. In the overall active market cap of RWA, stocks still account for a small portion; treasury bonds and funds are the main part. But the direction has shifted: from crypto world inventing assets on its own to integrating real-world assets, trading and settling them in crypto ways. The most direct impact on ordinary users is that asset allocation granularity has become finer. You can express your views on AI chip stocks, your demand for US treasury yields, and your exposure to crypto volatility all in the same account using USDT. The possibilities for portfolio combinations increase, and decision complexity also rises. In the next six months, what’s truly worth watching is not how many more US stocks are added, but whether the capital flow between these assets becomes genuinely active. How much capital shifts from pure crypto positions to tokenized stocks and funds, and how many institutions start using BUIDL-type assets as collateral. If these numbers start to rise significantly, then RWA is no longer just a narrative on a PPT but a real change in account balances. OKX’s choice in this round is very clear: first, perfect the entry point and unified experience, then discuss deeper institutional integration. This might be the most pragmatic path currently. #新手必看：这里有你需要的一切

#财报观察员：微软云收入破千亿，Meta却指引拉胯——AI故事分化了？ Microsoft and Meta released their earnings on the same day, but the market reacted as if they were completely different companies. Microsoft’s Q4 revenue reached $90 billion, up 18% year-over-year, with Azure and other cloud services revenue surging 43%. For the first time, Azure’s annual revenue surpassed $100 billion. Microsoft Cloud contributed $59.3 billion that quarter, and commercial remaining performance obligations (RPO) soared to $678 billion, an 84% year-over-year increase. Microsoft 365 Copilot paid seats exceeded 30 million. Profitability was also strong, with net income of $35.8 billion, up 31%. Meta’s Q2 revenue was $60.8 billion, up 28%, with advertising still driving growth. However, free cash flow collapsed to $784 million, a roughly 91% year-over-year plunge, hitting a recent low. Quarterly capital expenditures were $31.08 billion, nearly consuming all operating cash flow. The full-year capital expenditure guidance was raised to $130–145 billion, and Q3 revenue guidance was $6.1–6.4 billion, with a weak midpoint. The stock price plunged in after-hours trading. Both are investing heavily in building AI infrastructure, but one has started turning Tokens into quantifiable revenue and profit, while the other is still in a burn-then-see phase. Where exactly is the difference? Microsoft follows a combined path of selling shovels + seats. Azure itself is cloud infrastructure, and AI workloads directly become billable compute power and services. Enterprise customers sign multi-year contracts, and the surge in RPO indicates demand is locked in early. Copilot integrates AI into the Office ecosystem, becoming a fixed monthly paid seat. The business model has a clear monetization loop from the start: the more customers use, the higher the bill; the longer the contract, the more certain the future revenue. Nadella’s repeated emphasis on the “cost-to-outcome curve” essentially tells the market: we are turning expensive Tokens into business outcomes customers are willing to pay for continuously. Meta takes the path of ad optimization + supercomputing bets. AI currently mainly improves recommendation algorithms and ad delivery efficiency, directly reflected in ad revenue growth. But the bigger bet is on superintelligence and future new businesses. Zuckerberg continues to ramp up data centers and compute power, maintaining a capital expenditure ceiling of $145 billion, with the floor even rising. The problem is that these investments show almost no corresponding new revenue sources in the short term. Free cash flow is squeezed to extremely low levels, effectively using current cash to buy options for the future. The market is willing to be patient, but patience has a price—when guidance is weak and cash flow approaches zero, that price immediately shows in the stock price. This is not about who tells a better AI story, but that business models inherently determine different monetization rhythms. Most of Microsoft’s AI investment goes directly into existing enterprise contracts and software subscription systems, with short and measurable return paths. Meta’s AI investment largely lays the groundwork for yet-to-be-formed new businesses, with longer and more uncertain return paths. One is delivering, the other is betting. Amazon and Apple follow tonight. Can Amazon’s AWS continue to prove the cloud + AI business loop? Can Apple’s device and service revenue prove that edge AI can also bring real money? The market is becoming increasingly discerning: it’s not just about hearing how important AI is, but about seeing whether revenue, profit, and cash flow keep pace after you spend the money. The AI story has moved from a universal narrative to a differentiated realization phase. Companies that can turn compute power into sustainable revenue will continue to receive premiums; those still burning large amounts of cash waiting for answers will be required to provide clearer timelines. This is not about bearishness but about capital starting to measure ROI with a stricter ruler.

Morning Report 🗓️ 2026-07-29 Wednesday Mainstream coins collectively rose moderately, with Bitcoin up 1.20% to $63,885, Ethereum up 1.90% to $1,907, showing overall low volatility and a mild trend. The most notable is the sharp divergence among altcoins. Mantis surged 66.7% to $1.29, Bitway also jumped 40% to $0.1057, but these small coins are prone to pump-and-dump traps, so beware of the risk of buying at the top. On the downside, Peanut plummeted 14.9%, eCash dropped 10.6%; such one-sided declines often indicate heavy sell pressure or worsening project news. From trending searches, Hyperliquid ranks 10th in market cap attention, indicating that enthusiasm for derivatives trading remains. Other important news: Zcash launched the Ironwood zero-knowledge proof mechanism, claiming to technically avoid privacy vulnerabilities, marking a substantial advancement in the privacy coin sector. Tether signed a tokenization agreement with the Nairobi Securities Exchange, further expanding the stablecoin business landscape. Visa detailed its stablecoin strategy during the Q3 earnings call, showing that the traditional financial giant’s stance continues to warm up.