A war of attrition that cannot be won or negotiated 1. From the "quick victory illusion" to being "caught in a dilemma" On February 28, 2026, when the US and Israel jointly launched an attack on Iran, the Trump team was optimistic: relying on overwhelming air superiority, they could replicate the "Venezuela model," quickly paralyzing nuclear facilities, weakening the regime, and even facilitating regime change. The reality, however, was quite different. Five months later, although Iran's nuclear capability was severely damaged, it was not completely destroyed; missile and drone capabilities remain; the Strait of Hormuz repeatedly became a focal point of contention; US interception ammunition was heavily consumed, and Central Command has repeatedly warned of diminishing marginal returns on targets. Trump himself swung repeatedly between the hardline threat that "Iran will cease to exist" and "pausing airstrikes—resuming attacks—pausing again." After 13 consecutive days of airstrikes in late July, he once pressed the pause button, influenced by both military advice (target depletion, interception missile inventory pressure) and political considerations (midterm elections approaching, declining domestic support, oil price fluctuations). Now, again expressing "loss of confidence," he essentially admits that limited strikes alone cannot force Iran into substantive compromise. 2. The dual game of negotiation and strikes US side: Use military pressure to maintain negotiation leverage while avoiding full-scale ground war (Trump explicitly unwilling to deploy ground troops). Iran side: Use limited retaliation (missiles, drones, targeting commercial ships) to test US red lines while buying time to repair capabilities and drain US political and military resources. The result is a typical war of attrition—neither side can defeat the other, nor reach an agreement. Trump says, "They always want to talk but always break their word," while Iran believes US threats outweigh credible promises. Mutual trust has long been broken. 3. Real options and constraints for further strikes Trump’s talk of "preparing for another strike" currently shows roughly three possible paths: Limited escalation: Targeting energy facilities, power grids, bridges, and other infrastructure (previously threatened multiple times). This can increase domestic pressure on Iran but will significantly raise oil prices and may be seen as excessive by the international community. Joint expanded operations with Israel: Media reports indicate both sides are discussing joint strikes on key energy targets. The biggest irony of this war is that Trump, despite having overwhelming military superiority, finds himself trapped in a triangle of "cannot win by fighting, cannot stop, cannot negotiate." Although the Iranian regime is severely damaged, it shows greater resilience than expected; the US is consuming hard power while also depleting Trump’s own political capital. For the Middle East, this means proxy conflicts, strait risks, and energy premiums will persist. For global markets, oil prices and risk premiums will fluctuate in phases, but the probability of systemic collapse still depends on whether either side truly crosses the "full-scale infrastructure strike" red line. When Trump says "loss of confidence," he is actually admitting: he underestimated Iran’s will to survive and overestimated the decisive effect of limited US military actions. The actual scale and target choices of strikes in the coming weeks will speak louder than any verbal threats about whether this war will move toward limited damage control or continue sliding deeper into the quagmire. #特朗普称对伊失去信心，酝酿再打击

On July 30, 2026, the Ethereum mainnet celebrates its 11th anniversary since official launch. Since the genesis block was created on July 30, 2015, this blockchain, known as the "world computer," has achieved a rare feat in distributed systems: 100% uptime for eleven consecutive years, never experiencing a full network shutdown or block production interruption. This is not a marketing slogan but a fact repeatedly confirmed by the community and data. After enduring the DAO hack hard fork, multiple major upgrades, a complete transition from Proof of Work to Proof of Stake (The Merge), and countless client coordinations, Ethereum has maintained uninterrupted block production. This is a very high standard for any centralized system; for a decentralized network maintained by global validators without a single operating entity, it is an almost unbelievable reliability record. Eleven years: From experiment to infrastructure Ethereum's beginnings were full of idealism. When the Frontier mainnet launched in 2015, it was still an "unfinished" experiment: the Gas mechanism was crude, development tools were rudimentary, and user experience was difficult. The early core team witnessed the genesis moment in a small office in Berlin. The key milestones since then are clear: 2016: The DAO event hard fork established community governance and the boundaries of "code is law" discussions. 2020: Beacon Chain launched, officially starting the Ethereum 2.0 path. 2022: The Merge completed, switching from PoW to PoS, drastically reducing energy consumption and fundamentally changing the security model. 2023: Shanghai/Capella upgrade enabled staked ETH withdrawals. 2024: Dencun upgrade introduced Blob transactions (EIP-4844), significantly lowering L2 data costs. 2025: Pectra and Fusaka upgrades continued to advance scalability and data availability. 2026: Glamsterdam upgrade (including ePBS, block-level access lists, etc.) is in the testing phase, aiming to further increase Gas limits and efficiency. Each hard fork is a "network-wide coordinated surgery" but has never caused mainnet downtime. This continuous evolution capability is Ethereum's core competitive advantage over many successors. Current ecosystem achievements: Consolidated settlement layer status Eleven years later, Ethereum has evolved from a "programmable blockchain" into the core settlement layer of the global crypto economy. Key data as of late July 2026 roughly includes: Stablecoins: The mainnet hosts about $148–156 billion, accounting for more than half of the global stablecoin supply. Tokenized Real-World Assets (RWA): Approximately $15.5–17.2 billion, leading all networks. DeFi: Still the largest ecosystem, with TVL significantly higher than any single chain (some statistics show the mainnet and ecosystem combined advantage is several times that of the second place). Layer 2: Daily transaction volume has reached 10–14 times that of the mainnet, with Base, Arbitrum, Optimism, and others carrying the vast majority of user activity, while the mainnet focuses on final settlement and high-value transactions. Developers and institutions: Hosting the industry's largest developer community and gradually becoming the default entry point for institutions into the on-chain world. Behind these numbers are real-world implementations of stablecoin cross-border payments, DeFi lending and trading, NFT and digital ownership, and tokenization of traditional financial assets. Ethereum is no longer just an "experimental field" but infrastructure carrying real funds and real use cases. Challenges and the next chapter The 11th anniversary is not without flaws. ETH price has significantly retreated from the 2025 peak, mainnet fee revenue has declined due to L2 diversion, and the Foundation organization is undergoing adjustments. These are realities mature networks must face. But the roadmap remains clear: the Glamsterdam upgrade will continue to improve throughput and user experience, with subsequent upgrades like Hegotá focusing on censorship resistance and long-term security. The community repeatedly emphasizes "Credibly Neutral" — no company owns it, no one can unilaterally shut it down, and everything on-chain is verifiable — which becomes even more precious as institutional capital accelerates entry. #以太坊主网十一周年：十一年不间断运行与生态成就

Why is Google willing to "backstop"? Data center developer Nexus Data Centers is negotiating with a bank syndicate led by Morgan Stanley to raise about $15 billion in debt financing for a large AI data center project located in Hubbard, Texas. The key driver behind this is Google—the search and cloud giant under Alphabet. Google has agreed to provide multi-billion-dollar credit guarantees (backstop) for Anthropic (developer of Claude) covering lease and power payment obligations in this project, which includes four data center lease contracts and related power purchase agreements. In return, Google is expected to receive approximately 20% equity in the data center and its supporting power plant. This is not a simple "friendly sponsorship" but a typical move by tech giants in the current AI infrastructure boom, leveraging financial engineering to drive large-scale computing power expansion. Core structure of the deal The project is led by Nexus Data Centers, located in Hubbard, Texas, planning to build a 1.6 GW natural gas power plant (behind-the-meter, i.e., on-site direct supply) to solve the most challenging power bottleneck for AI data centers. The financing plan roughly consists of: About $14 billion in bridge loans + revolving credit facilities, totaling approximately $15 billion. The bank syndicate is led by Morgan Stanley, with Evercore serving as Nexus's financial advisor. Google's guarantee only covers the "minimum necessary amount" required by the banks, not a full backstop. If Anthropic defaults, Google must bear the corresponding payment obligations but retains the rights to take over the underlying leases or sublease them. Anthropic plans to deploy TPUs (Tensor Processing Units) jointly designed by Google and Broadcom in this campus. The chips themselves are financed through independent supplier financing arrangements between Anthropic and Broadcom, further dispersing risk. For Google, this is a dual gain of "credit for equity + chip orders": it does not directly record all construction costs on its balance sheet, while locking in computing power customers and sharing long-term project returns. Why is Google willing to "backstop"? This is not an isolated event. Google has sharply increased the scale of its credit guarantees in recent years. By mid-2026, the nominal amount of potential backstops for third-party data center lease payments has surged from about $6.5 billion nine months ago to approximately $44 billion. These guarantees are mostly accounted for as credit derivatives, with the fair value recorded as liabilities far below the nominal amount. The core logic has three points: Accelerate TPU ecosystem expansion Facing Nvidia's absolute dominance, Google needs more third-party customers to actually use TPUs. By providing financing support for data center projects, Google enables developers to build campuses at lower costs and faster speeds, thereby absorbing its chip capacity. A delicate balance of risk and capital efficiency Building data centers directly would significantly increase capital expenditures and debt. Through the "guarantee + minority equity" model, Google transfers some construction and operational risks to banks, developers, and tenants (Anthropic), while retaining upside returns. If the project succeeds, the 20% equity can bring considerable returns; if problems arise, Google has the right to take over the assets. Following the industry trend of "off-balance-sheet" financing Peers like Meta and Microsoft also use similar structures (SPVs, leases, guarantees) to isolate debt. Google's move is both a competitive necessity and a proactive use of Wall Street financing tools.#谷歌为AI数据中心债务兜底，换取两成股权

The U.S. officially steps in, supporting the yen jointly for the first time in nearly 30 years From July 30 to August 1, 2026, a rare scene unfolded in the foreign exchange market: after Japan launched a large-scale yen-buying intervention, the U.S. Treasury made a rare direct move by selling euros and buying yen through the New York Fed. This marks the first time in nearly 30 years that the U.S. and Japan have jointly supported the yen through "actual currency purchases," and it is the first directional intervention targeting the yen by the U.S. since 2011. The yen rebounded strongly by over 6% within just two days from near a 40-year low (around 164), briefly surging to the 157 range. This intervention has escalated from a unilateral Japanese action to a coordinated U.S.-Japan effort, signaling a substantial shift in the market landscape. Why is the U.S. personally stepping in now? The U.S. has long held a stance of "understanding but not directly participating" in Japanese interventions; this direct action is highly significant. The main motivations can be understood on three levels: Preventing chain risks triggered by further uncontrolled yen depreciation The yen approaching the 164 level has already increased Japan's import costs and inflationary pressures, while also intensifying volatility in global risk assets. If the yen continues to fall disorderly, Japan might be forced to sell U.S. Treasuries on a larger scale to obtain dollars, further pushing up U.S. Treasury yields (the 30-year yield has risen to its highest since 2007). U.S. intervention can mitigate the impact of Japan's solo actions on the U.S. Treasury market. Policy coordination and alliance signaling Biden and Kazuo Ueda maintain close communication, emphasizing the "strong relationship and close coordination" between the U.S. and Japan. The timing of this intervention aligns with the Bank of Japan maintaining a 1% interest rate while signaling a hawkish stance (increasing the probability of a rate hike in September). It also sends a clear message to the market that "the U.S. and Japan will not allow the yen to depreciate disorderly." South Korea's simultaneous intervention further strengthens regional coordination. Strategic and political considerations Under the Trump administration framework, the U.S. tends to actively manage exchange rates to protect its interests and allies' stability. Biden himself has a strong background in forex trading and is familiar with the Japanese market. Direct involvement can suppress speculative short positions and enhance U.S.-Japan policy coordination influence in forums like the G20. Historically, the last large-scale coordinated U.S. intervention in the yen was after the 2011 Great East Japan Earthquake (then selling yen to prevent excessive appreciation). This time, the operation is the exact opposite—"supporting the yen"—and is conducted through actual transactions rather than verbal warnings, making it extremely rare. Single or short-term interventions often only bring temporary rebounds and rarely reverse trends completely. The true medium-term direction of the yen is still determined by: U.S.-Japan interest rate differential: The Bank of Japan maintains 1%, but Kazuo Ueda's press conference was hawkish, and the market is beginning to price in a possible rate hike in September. If Japan accelerates rate hikes, the intervention's effect will be supported by fundamentals. U.S. policy stance: Biden has clearly stated the yen is undervalued; whether actual currency purchases continue and their scale will be key points to watch. Speculative forces and capital flows: The yen's long-term depreciation is driven by structural factors such as the U.S.-Japan interest rate gap, Japanese corporate overseas investment, and arbitrage trading. Intervention can only raise short-selling costs but cannot change fundamentals. Risk appetite and geopolitics: Energy prices, Federal Reserve policy path, and geopolitical situations will continue to influence the dollar's direction. Historical experience shows that after multiple large-scale Japanese interventions, if there is no follow-up policy (rate hikes or structural reforms), the exchange rate often gradually gives back gains. However, this time with direct U.S. involvement, the deterrent effect far exceeds previous instances, and short sellers betting on yen depreciation will face higher risks. Trader takeaways #日元干预战升级，美方准备介入

⚡️ $BTC Ethereum 4H key level update, weekend intraday long and short entry points reference (if no events affect the weekend, there may be no entry opportunities) ⚡️ $ETH Ethereum 4H key level update, weekend intraday long and short entry points reference (if no events affect the weekend, there may be no entry opportunities) Develop the habit of scaling in and out for entries and take profits.

A pricing crisis about "long-term trust," why specifically the 30-year term? The 30-year Treasury bond is the market's pricing tool for the "distant future." It reflects not only current inflation and Federal Reserve policies but also investors' comprehensive judgment on the sustainability of U.S. fiscal policy, inflation persistence, and the creditworthiness of the dollar over the coming decades. Historically, the 30-year yield has remained stable at relatively low levels. After the 2020 pandemic, it briefly fell below 1.5%, and the 2023 peak only briefly hovered around 5%. However, in 2026, it not only broke through 5% but also stayed above 5% for more than 20 trading days cumulatively within the year, setting the longest continuous high-level record since 2007. Several major factors driving this round of increases: the second shock from Middle East conflicts and oil prices The ongoing U.S.-Iran conflict continues to push energy prices higher. Between April and July, multiple inflation data points (PPI, CPI) exceeded expectations. Rising oil prices directly increase costs, repeatedly undermining market confidence in "inflation easing." Investors have begun to reprice the "potentially longer duration of high inflation." Shift in Federal Reserve policy signals The July FOMC meeting held rates steady (3.5%–3.75%) but saw three members dissenting in favor of a rate hike, indicating stronger hawkish voices internally. The chair's statements also emphasized "acting without hesitation if necessary." The market's pricing shifted from "waiting for rate cuts" to "possible further hikes," with the long end reacting particularly strongly. Dual pressure from fiscal deficits and debt supply The scale of U.S. federal debt continues to expand, with deficits remaining high. Meanwhile, tech giants are issuing large amounts of corporate bonds for AI infrastructure construction, competing with Treasuries for long-term funds. Increased supply combined with cautious demand naturally pushes long-term yields higher. The return of the "term premium" Over the past decade, compensation for holding long-term bonds (term premium) was suppressed at very low levels. Now, geopolitical uncertainty, policy reversals, and unclear debt trajectories all demand higher risk compensation. The market is relearning that "lending to the government for 30 years is not cheap." What impact will this have? Mortgage and consumer credit: The 30-year yield is a key anchor for mortgage rates. Sustained high long-term yields will directly raise the cost of newly issued mortgages, suppressing real estate transactions and consumer credit expansion. Stock market valuation pressure: Higher risk-free rates will lower the valuation base for growth stocks, especially high-valuation tech stocks. Although the AI narrative is strong, rising financing costs will test actual returns. U.S. government financing costs: The Treasury's borrowing costs have risen significantly, with interest expenses continuing to climb as a share of the budget, creating a "debt-interest-reissuance" cycle pressure. Global asset repricing: U.S. Treasuries are the benchmark for global asset pricing. Rising long-term yields will drive up long-term rates in other countries and affect the pace of capital flows from emerging markets back to dollar assets.#30年期美债收益率创19年新高

On July 31, 2026, Trump publicly stated at the Camp David cabinet meeting that he was "losing faith" in Iran because "they lie, mislead, and frequently break promises." He specifically mentioned that during the negotiations, Iran fired 5 missiles at the U.S. military base in Jordan (all intercepted). He then declared: "We will hit them hard... at some point, they will say 'we really can't hold on anymore.'" Meanwhile, U.S. media revealed that he had approved a new round of strikes, possibly starting as early as the weekend, targeting energy infrastructure with the intent to force Tehran to "surrender" or return to the negotiating table. This is not the first time. Since the large-scale U.S.-Israel joint military operation against Iran on February 28, the war has entered its fifth month. The initial goals were clear: destroy nuclear facilities, missile capabilities, weaken the navy and air force, and even push for regime change. Trump confidently claimed it could be a quick victory like the "Venezuela model." The reality, however, is that Iran has resisted stubbornly, the Strait of Hormuz has been repeatedly disrupted, U.S. missile intercept stockpiles are running low, domestic support is declining, and oil prices and inflation pressures are feeding back into the U.S. Why has "losing faith" become a turning point? Trump's core logic has always been "military pressure for negotiations." He repeatedly switched between ceasefire and strikes: strike first, then talk, talks break down then strike again, then pause. Each pause came with "goodwill signals" or third-party mediation (Oman, Gulf countries, etc.), but Iran's response was often missiles, drones, or attacks on commercial ships. After 13 consecutive days of airstrikes in late July, Trump suddenly called a halt, citing ammunition depletion, targets nearly exhausted, and leaving room for diplomacy. But Iran attacked during the negotiation window, directly breaking his patience. For Trump, this was not a simple tactical friction but "they simply don't take the agreement seriously." On a deeper level, this exposes the current U.S. strategic dilemma: Military advantage is hard to translate into political results The U.S. military has clear firepower superiority; Iran's navy, air force, and missile production capacity have been severely weakened. But Iran uses asymmetric retaliation—dispersed deployment, proxy harassment, and threats to block the strait. Without ground troops, airstrikes cannot completely destroy the opponent's will or regime. Trump himself admitted "they are very tough," even half-jokingly giving Iran "props." Dual constraints of resources and domestic politics Patriot missile intercept stockpiles continue to decline, and the Pentagon has issued warnings. Continuing high-intensity strikes will further deplete global U.S. air defense capabilities. Domestically, war support is low, midterm elections are approaching, and oil prices and living costs are major issues. Trump needs to "hit Iran hard," but cannot turn it into a real long-term war. Negotiation space is compressed by mutual distrust The U.S. demands complete abandonment of the nuclear program, missile restrictions, and cutting off proxy support. Iran prioritizes survival and regime security, viewing military pressure as a "war to promote talks" bargaining chip. Both sides are testing each other's bottom line with "one more round of fighting," resulting in ceasefire agreements repeatedly failing. Possible directions and risks #特朗普称对伊失去信心，酝酿再打击

On the surface, it's a strong financial report, but in essence, the reserve structure is quietly undergoing a transformation. Net operating profit is $1.5 billion (a significant quarter-over-quarter increase), gold reserves have increased to 146.2 tons (an additional 14 tons in a single quarter). Excess reserves have been halved from 8.23 billion to 4.11 billion. No matter how high the profit is, it cannot hide the fact that the safety cushion has been cut in half. What really deserves attention is not the 1.5 billion. The 1.5 billion profit mainly comes from U.S. Treasury bonds and repo interest, indicating that in the current interest rate environment, Tether remains a "money printing machine." But this money is being rapidly converted into gold. With 146.2 tons of gold, Tether has become the world's largest non-sovereign gold holder, surpassing the total reserves of many countries. The 14-ton increase in Q2 alone accounts for about 5% of the global central bank gold purchases that quarter. It is not "allocating" but systematically absorbing capital. At the same time, excess reserves dropped directly from 8.23 billion to 4.11 billion. The price declines of gold and Bitcoin contributed to some impairment, but even accounting for these, the entire gap cannot be explained. Assets decreased by about 4 billion, profits were 1.5 billion, implying about 5.5 billion in unrealized losses or fund movements. More subtle is the transparency: Q1 disclosed detailed asset amounts, but Q2 almost no longer provides precise classifications. When volatility increases, information disclosure shrinks, which itself is a signal. What does this mean? Tether is "central bank-ifying." It continuously converts stable interest income into hard assets to hedge long-term risks of the dollar system. The gold proportion is close to 10%, and BTC holdings are also increasing simultaneously. In the future, fluctuations in its balance sheet will indirectly affect the entire crypto market through USDT. The safety cushion is thinning. Excess coverage dropped from about 4.5% to about 2.2%. In extreme redemption scenarios, the buffer space is clearly narrowing. Gold liquidation speed is much slower than U.S. Treasuries, and liquidity flexibility is decreasing. USDT's dominant position is more secure. While the overall stablecoin market is contracting, USDT is slightly increasing against the trend, with market share locked above 60%. In short-term trading, the liquidity advantage of USDT trading pairs will only strengthen. #Tether季度盈利15亿，黄金增至146吨

Why did the stock price plunge despite better-than-expected earnings? Guidance below expectations was the direct trigger. Apple clearly pointed out that constraints on advanced process chips and memory supply will significantly worsen in Q4, affecting the supply of iPhone, Mac, and iPad. At the same time, foreign exchange also dragged down by about 2.5 percentage points. The market is very sensitive to the reaction of "growth slowdown caused by supply constraints." Service business and high growth expectations fell short The service business has always been an important support for Apple's high valuation. This time, both growth rate and absolute value were slightly below expectations, coupled with average performance in China, weakening the narrative of "all-around strength." The market had already priced in a strong hardware cycle iPhone and Mac's strong performance was not entirely unexpected. The previous stock price had already reflected the product cycle dividend; what truly determines the short-term direction is the forward guidance. Cook's transition period amplifies sentiment This is Cook's last earnings call as CEO. The market is less tolerant of any uncertainty during this sensitive transition period. Apple's earnings this time exposed a structural contradiction: There is still strong demand on the hardware side (especially iPhone), but supply chain flexibility is severely squeezed under the AI boom. Memory and advanced process capacity are being diverted by large-scale AI infrastructure construction, making Apple, as a consumer electronics giant, a "victim of capacity." Unlike companies like Amazon and Microsoft that can compete for capacity through capital expenditure, Apple relies more on external supply chains. When supply becomes a bottleneck, even strong demand cannot fully convert into revenue. Apple won the current quarter's numbers but lost in expectation management. The market now focuses more on "whether growth can be sustained" rather than "how good this quarter was." If supply constraints continue into the new iPhone cycle, it will put substantial pressure on revenue and gross margin in the second half of the year. #苹果第三财季业绩超预期，盘后股价大幅下跌

Last night, Amazon delivered an earnings report with "slightly soft guidance, but a big jump in stock price." After-hours, the stock surged nearly 10%, sharply contrasting with some major tech companies whose "decent earnings but stock price got hit." Key data breakdown Current period performance fully exceeded expectations: revenue of $200.6 billion, up 20% year-over-year, beating market expectations Operating profit of $27.5 billion, up 43% year-over-year AWS revenue of $42.2 billion, up 37% year-over-year, marking the fastest growth in the past 18 quarters and clearly surpassing expectations Guidance is the only "flaw" Q3 revenue guidance of $197–202 billion (midpoint about $199.5 billion), below analysts' expectation of about $204 billion Operating profit guidance of $22.5–26.5 billion, midpoint also slightly below market expectations On the surface, it looks like "strong current quarter, weak forward guidance." But the market chose to vote with its feet, focusing on more core aspects. Why did the stock price surge despite guidance missing expectations? AWS is accelerating again, the true core. A 37% growth rate means Amazon's cloud business is re-entering a high-growth trajectory. In the current AI arms race, whoever can prove "investment is turning into revenue" will get a premium. Amazon answered with the fastest growth in 18 quarters. AI and chip businesses begin scaling CEO Jassy clearly stated that the annualized revenue scale of AI business and self-developed chip business has both exceeded $25 billion. This is no longer a "story" but revenue being realized. Capital expenditure increase accepted by the market The company raised its full-year capital expenditure forecast from $20 billion to $22 billion, mainly due to rising memory costs and sustained strong AI demand. Unlike some peers whose "increased spending was punished," Amazon's clear growth support led the market to tolerate it. One-time gains inflated profit numbers, but the market sees more clearly The large net profit increase includes fair value changes on the Anthropic investment. The market focuses more on operating profit and AWS's real momentum, not one-time items. #财报观察员：亚马逊指引不及预期，股价却反涨9%