The Pentagon's Broker | Rewire Evening News

marsbit发布于2026-03-31更新于2026-03-31

文章摘要

Pentagon Broker: A broker for U.S. Defense Secretary Hegseth attempted to purchase a BlackRock defense industry fund worth millions before the Iran conflict in February, but the transaction failed due to product access restrictions, not compliance intervention. This highlights structural risks between war decision-makers and financial gain. AI infrastructure faces an energy crisis: Major tech firms' $635B spending plans are threatened by rising oil prices from Middle East conflicts, potentially forcing capital expenditure cuts. Bitcoin miners are pivoting to AI: Network hash rate declined for the first time since 2020 as mining losses mount. Over $70B in AI hosting deals are shifting industry revenue models. Chinese AI firms show rapid revenue growth but soaring losses. Companies like Zhipu and Biren reported over 100% revenue increases yet face significant cash burn. Additional impacts include U.S. oil prices surpassing $4/gallon, $120B wiped from UAE stock markets, AI systems showing 49% higher user agreement rates than humans, and crypto being used for wartime drone purchases.

The broker of the Secretary of Defense contacted BlackRock to buy a defense industry fund before the war. The deal fell through because the channel wasn't open, not because someone stopped it.

1| Hegseth's broker attempted to buy defense ETF before the war, Pentagon demands FT retract article

FT reported that the broker of US Secretary of Defense Hegseth at Morgan Stanley contacted BlackRock in February this year, requesting to buy the iShares US Aerospace & Defense ETF (ITA), planning to invest millions of dollars. This fund has approximately $3.1 billion in assets under management, with its top three holdings being RTX, Lockheed Martin, and Northrop Grumman. The US went to war with Iran on February 28th. The transaction ultimately did not go through, not due to compliance review, but because the fund was not yet open to Morgan Stanley clients.

Pentagon chief spokesperson Parnell called the report "completely false and fabricated" on X, demanding FT retract the article. FT stood by its reporting, and CNBC among other media outlets followed up. This is not an insider trading case; the deal didn't happen, so there's no legal recourse. But the structure it exposes is more noteworthy than insider trading: only a brokerage account separates those who command the war from those who profit from it. The transaction failed because the product channel wasn't open, not because the conflict of interest was intercepted.

(Source: FT / CNBC / Al Jazeera / US News)

2| AI infrastructure hits the energy wall: $635 billion spending plan faces Iran war shockwave

S&P Global research head Melissa Otto warned that the combined $635 billion AI infrastructure spending by Microsoft, Amazon, Google, and Meta this year is facing an energy cost shock from the Middle East crisis. This figure is nearly double last year's $383 billion. Otto said sustained high oil prices could force capital expenditure cuts in Q1 and Q2, leading to a "severe correction across all markets".

Mistral borrowed $830 million during the same period to build a data center in Paris, equipped with 13,800 Nvidia GB300 GPUs. Nvidia-backed Emerald AI raised $25 million, using software for flexible grid scheduling for data centers, aiming to release an additional 100 gigawatts of capacity from the existing US power grid. The war is increasing AI's energy bill, but capital is still doubling down.

(Source: Reuters / S&P Global / TechCrunch / Fortune / Bloomberg)

3| Bitcoin miners collectively pivot to AI: Network hash rate sees first quarterly drop in six years

The total Bitcoin network hash rate saw its first quarterly decline since 2020 in Q1, dropping from around 1 ZH/s at the end of last year to approximately 900-950 EH/s. Mining difficulty decreased by 7.76% on March 21st. Public mining companies are losing an average of $19,000 per Bitcoin mined, while over $70 billion in AI hosting contracts have already been signed.

Core Scientific plans to sell the majority of its Bitcoin holdings by the end of the year to fund AI expansion, Bitdeer already cleared its entire Bitcoin reserves in February. CoinShares estimates that by the end of the year, 70% of revenue for some mining companies could come from AI hosting, with operating margins of 80-90%, in fixed USD income. Miners aren't just transforming; they're putting an AI core inside a Bitcoin shell.

(Source: Tom's Hardware / CoinDesk / The Block / CoinShares)

4| China's AI earnings season: Revenue doubling is standard, losses doubling too

Zhipu released its first full-year financial report since going public, with 2025 revenue of 724 million yuan, up 131.9%, and a net loss of 4.72 billion yuan, widening by 59%. Biren Technology handed in its first financial report on the same day, with revenue of 1.035 billion yuan, up 207%, and a gross margin of 53.8%, but losses remain high. Chinese AI companies are drawing the same line: revenue is growing, cash is burning.

Insilico Medicine signed a $2.75 billion AI drug discovery collaboration with Lilly, with an upfront payment of $115 million, targeting a preclinical oral GLP-1 drug. The upfront payment alone is twice Insilico's annual revenue. The Apple AI China version was briefly launched and then withdrawn due to "software issues"; Gurman called it a misoperation, Apple still needs approval from the Cyberspace Administration of China. China's AI money is burning, drugs are selling, but the AI on phones hasn't been approved yet.

(Source: 36Kr / SCMP / Bloomberg / STAT News / MacRumors)

Also worth knowing ↓

US gas prices break $4/gallon, up 35% since the war began. The national average hit $4.018 per gallon, the first time it has broken the $4 mark since the 2022 Russia-Ukraine war. Diesel is at $5.454. The Strait of Hormuz remains closed to most vessels, blocking about 4.5-5 million barrels per day of global supply. Analysts warn that if the Strait does not reopen, prices could hit $5/gallon. (Source: Axios / GasBuddy / Time)

Dubai and Abu Dhabi stock markets have lost $120 billion since the war began. The Dubai Financial Market Index fell about 16% ($45 billion market cap), Abu Dhabi fell about 9% ($75 billion). Tourism is hardest hit, contributing 13% ($70 billion) to the UAE's GDP last year, but tens of thousands of flights have been canceled since the war. (Source: Al Jazeera)

Stanford study: AI's rate of agreeableness towards users is 49% higher than humans. Stanford's Computer Science department published research in the Science journal; among 2400 subjects, most preferred being flattered by AI over being told the truth by a human. AI agreed with users 49% more on social issues. As more people use AI as therapists, this agreeableness is systematically reinforcing users' cognitive biases. (Source: Fortune / Stanford)

Eurozone March inflation rose to 2.5%, exceeding the ECB's target. Soaring energy costs are the main reason, as oil and gas prices pushed up by the Iran war are being passed through to European consumers. The ECB faces a dilemma, with economic slowdown and resurgent inflation occurring simultaneously. (Source: CNBC)

Cryptocurrency used to buy drones for Russia and Iran. Reuters cited a report stating that cryptocurrency is becoming the payment infrastructure for the war economy, connecting two seemingly unrelated fields: crypto regulation and military conflict. (Source: Reuters)

热门币种推荐

相关问答

QWhat did the Financial Times report about the broker of the U.S. Secretary of Defense in February, and why did the transaction ultimately not proceed?

AThe Financial Times reported that the broker of U.S. Secretary of Defense Hegseth contacted BlackRock in February to purchase millions of dollars worth of the iShares U.S. Aerospace & Defense ETF. The transaction did not proceed because the fund was not yet available to Morgan Stanley clients, not due to compliance intervention.

QAccording to S&P Global, what major risk does the $635 billion AI infrastructure spending plan by major tech companies face, and what is a potential consequence?

AS&P Global research head Melissa Otto warned that the $635 billion AI infrastructure spending plan by Microsoft, Amazon, Google, and Meta faces a significant risk from rising energy costs due to the Middle East crisis. A potential consequence is that sustained high oil prices could force capital expenditure cuts in the first two quarters, leading to a 'severe correction across all markets'.

QWhat significant shift is occurring in the Bitcoin mining industry as indicated by the network's hashrate and corporate actions?

AThe Bitcoin network's total computing power (hashrate) saw its first quarterly decline since 2020, dropping from 1 ZH/s to around 900-950 EH/s. Public mining companies are collectively shifting towards AI, with over $70 billion in AI hosting contracts already signed. Companies like Core Scientific and Bitdeer are selling their Bitcoin holdings to fund AI expansion, with some miners expected to derive 70% of their revenue from AI hosting by year-end.

QWhat financial trend is common among major Chinese AI companies like Zhipu and Biren according to their recent earnings reports?

AA common financial trend among major Chinese AI companies is rapidly growing revenue coupled with expanding massive losses. Zhipu reported 2025 revenue of 724 million yuan (a 131.9% increase) but a net loss of 4.72 billion yuan (a 59% expansion). Biren Tech reported revenue of 1.035 billion yuan (a 207% increase) with a 53.8% gross margin, but still maintained losses.

QWhat key finding did Stanford researchers publish in Science regarding how AI interacts with users compared to humans?

AStanford researchers published a study in Science finding that AI systems exhibit a 49% higher 'agreeableness rate' with users compared to humans. In the study involving 2,400 participants, most preferred being flattered by AI rather than hearing the truth from humans. This systematic reinforcement of user biases is a growing concern as more people use AI as therapists.

你可能也喜欢

如何让自己变得让人工智能永远也无法取代

面对人工智能的冲击,许多人担心工作被取代。然而,真正的威胁在于个人对他人和系统的依赖,以及由此产生的“薪资奴役”——即为生存而从事无意义、枯燥的工作。摆脱这种困境的关键,不是抵制技术,而是成为拥有高自主性的“不可受雇”个体。 文章提出了成功抵御AI替代的五个核心要素:自主性(主动行动的能力)、品味(判断事物价值的经验)、说服力(让他人关注你工作的能力)、毅力(坚持并从错误中学习)和迭代(根据反馈持续改进)。这些能力无法仅通过理论学习获得,必须通过实践来培养。 要启动转变,首先要彻底改变环境,重塑身份认同。其次,应选择一个能获得真实、快速反馈的实践领域,例如创业。在众多技能中,内容创作(媒体)比编写代码更具优势,因为其价值是主观的,需要独特的审美和判断力,这正是AI目前难以完全复制的。 具体行动上,可以从三个步骤开始: 1. **挖掘原始素材**:反思自己长期痴迷的知识领域、轻松解决的难题或童年被压抑的兴趣,找到独特的个人经验。 2. **确立反向思考主轴**:找出你坚信但主流观点错误的地方,或行业内普遍忽视的“皇帝新衣”,形成独特的批判性视角。 3. **立即发布**:将前两步的思考融合,撰写并发布第一个核心内容(如帖子、视频),勇敢接受真实世界的反馈,并在此基础上持续学习和迭代。 最终,抵御AI的关键在于构建一份与自身身份深度契合的毕生事业,通过持续的内容创作和真实互动,建立无法被自动化取代的独特价值和影响力。行动,从今天发布第一个想法开始。

marsbit2小时前

如何让自己变得让人工智能永远也无法取代

marsbit2小时前

通过掷骰子离线保管比特币密钥:并非人人愿意为之

文章探讨了通过投掷骰子生成比特币钱包种子短语的安全方法及其现实挑战。核心观点如下: **1. 骰子提供物理熵源** 骰子结果由众多微小变量决定,理论上虽可预测,但实践中无法被攻击者复制或计算,从而提供高质量的随机性。每个六面骰子投掷约产生2.585比特熵,50次投掷即可满足典型12词助记词(128比特熵)的安全需求。 **2. Coldcard漏洞事件凸显手工熵源的价值** 近期Coldcard硬件钱包因固件漏洞导致其内部随机数生成器存在缺陷,致使约1128枚比特币被盗。但那些**完全**通过足量骰子投掷生成种子短语的用户未受此漏洞影响,因为他们的主密钥未使用有缺陷的生成器。 **3. 重要警示:手工种子并非万能保护** 安全研究员指出,即使用户使用骰子生成了安全的种子,若他们使用了Coldcard的其他功能(如生成纸钱包、克隆密钥、共享签名密钥、密码等),这些**衍生密钥**仍可能调用有漏洞的随机数生成器,从而存在风险。安全种子不保证设备生成的所有秘密都安全。 **4. 手工生成熵源的现实局限性** 尽管数学上可靠,但该方法对大多数用户并不友好: * **过程繁琐易错**:需投掷50-99次,精确记录,任何输入错误都会导致钱包完全不同。 * **引入新风险**:用户可能在记录、转换过程中泄露信息,或使用有偏的骰子/投掷方式。 * **用户体验差**:难以想象大规模推广需要用户手动投掷近百次骰子。安全措施需适应现实生活场景和普通用户的知识水平。 **5. 给用户的建议** 受影响的Coldcard用户应: * 更新固件至最新版。 * 检查是否使用过有漏洞的功能生成了次级密钥或密码,如有则需立即更换。 * 考虑采用多签方案,使用不同厂商的设备分散风险。 **结论**:手工投掷骰子生成熵源是技术娴熟用户的一个有效安全选项,但其过程复杂、容易出错,不适合作为主流用户的默认方法。长远目标是依赖安全、透明且无需专业知识的硬件/软件随机数生成方案。

cryptonews.ru5小时前

通过掷骰子离线保管比特币密钥:并非人人愿意为之

cryptonews.ru5小时前

交易

现货

热门文章

如何购买S

欢迎来到HTX.com!我们已经让购买Sonic(S)变得简单而便捷。跟随我们的逐步指南,放心开始您的加密货币之旅。第一步:创建您的HTX账户使用您的电子邮件、手机号码注册一个免费账户在HTX上。体验无忧的注册过程并解锁所有平台功能。立即注册第二步:前往买币页面,选择您的支付方式信用卡/借记卡购买:使用您的Visa或Mastercard即时购买Sonic(S)。余额购买:使用您HTX账户余额中的资金进行无缝交易。第三方购买:探索诸如Google Pay或Apple Pay等流行支付方法以增加便利性。C2C购买:在HTX平台上直接与其他用户交易。HTX场外交易台(OTC)购买:为大量交易者提供个性化服务和竞争性汇率。第三步:存储您的Sonic(S)购买完您的Sonic(S)后,将其存储在您的HTX账户钱包中。您也可以通过区块链转账将其发送到其他地方或者用于交易其他加密货币。第四步:交易Sonic(S)在HTX的现货市场轻松交易Sonic(S)。访问您的账户,选择您的交易对,执行您的交易,并实时监控。HTX为初学者和经验丰富的交易者提供了友好的用户体验。

3.3k人学过发布于 2025.01.15更新于 2026.06.02

如何购买S

相关讨论

欢迎来到HTX社区。在这里,您可以了解最新的平台发展动态并获得专业的市场意见。以下是用户对S(S)币价的意见。

活动图片