Wall Street opened August with a broad risk-on move as falling oil prices, lower Treasury yields and a renewed bid for AI and cloud stocks pushed major U.S. indexes higher. The Dow Jones Industrial Average rose 1.32% to a record close, while the Nasdaq Composite gained 2.13% and the S&P 500 climbed 1.48%. Markets repriced energy risk after U.S. President Donald Trump said talks with Iran had begun and suggested the Strait of Hormuz could reopen soon. Iranian officials denied direct talks with Washington, saying discussions were limited to shipping security arrangements with Oman, but crude still sold off sharply. WTI dropped 7.42% and Brent fell about 6%, easing inflation pressure and helping Treasuries rally. The 10-year U.S. yield slipped about 5 basis points to around 4.68%, while gold held above the $4,000 mark. The strongest equity action came from AI and cloud names. Amazon rose 4.58% and crossed a $3 trillion market capitalization for the first time after stronger-than-expected AWS results. Nvidia gained nearly 3% and moved back above $5 trillion, Meta rose more than 6%, and Microsoft and Google each added close to 5%. Crypto-linked stocks also mostly advanced, with SoFi Technologies up more than 10%, IREN up over 8%, and Hut 8 and Robinhood both up more than 4%.

Bitget’s UEX daily market report on Aug. 4 put the focus on three linked themes: easing geopolitical tension around the Strait of Hormuz, a sharp rebound in U.S. technology stocks, and a still-cautious crypto market. Donald Trump said on Aug. 3 Eastern Time that talks with Iran had begun, with reopening the Strait of Hormuz as the first stage and denuclearization as the second, adding that the waterway could reopen as soon as tomorrow. The report said that comment quickly pushed down the risk premium in crude, even as Iran’s military adviser Mohsen Rezaei rejected any U.S.-backed non-Iranian route through the strait. In equities, the Dow, S&P 500 and Nasdaq all rose, while Amazon’s market capitalization moved above $3 trillion for the first time. Meta, Microsoft, Google, Nvidia and Tesla also gained. Palantir stood out after reporting second-quarter revenue of $1.94 billion and adjusted EPS of $0.41, both above expectations, and raising its full-year guidance, sending the stock up more than 14% after hours. Crypto lagged the move in tech. BTC traded around $62,420 and ETH around $1,850, while total crypto market capitalization was about $2.27 trillion. The report also noted $265 million in net outflows from spot BTC ETFs the previous day and 24-hour liquidations totaling about $244 million.

Citi said in a July 31 research note that second-quarter data from Microsoft, Amazon, Google and Oracle is challenging a market narrative that heavy AI infrastructure spending would crush profit margins. The four cloud vendors posted a combined $103.8 billion in quarterly revenue, up 50% year over year, accelerating from 41% in the prior quarter. According to the note, revenue growth is now outpacing the rise in capital expenditures, suggesting that AI investment has moved beyond a pure cash-burn phase. Azure revenue growth accelerated to 43%, AWS to 37%, Google Cloud to 82%, and Oracle Cloud Infrastructure to 92%. Citi also highlighted improving monetization metrics, including AWS AI annualized revenue rising from $15 billion to $25 billion in one quarter and Google reporting token usage growth of 37.5% quarter over quarter to 22 billion per minute. The bank kept Buy ratings on Microsoft, Amazon, Google and Oracle, with price targets of $600, $350, $447 and $330, respectively. Citi argued that operating leverage is helping offset pressure from depreciation and expansion, while growing backlog at AWS, Google Cloud and Microsoft points to sustained AI demand and stronger revenue visibility in coming quarters.

Volatility in U.S. semiconductor stocks has picked up sharply since July, and RockFlow argues the move should not be read as a simple collapse of the AI trade. In its latest market analysis, the research team says AI demand still looks intact: Nvidia GPUs remain tight, HBM orders are still full, cloud companies are still spending on AI infrastructure, and data center buildouts have not stopped. What has changed, in RockFlow’s view, is the market’s pricing framework. Investors are no longer willing to pay up solely for distant AI upside, and the period of buying anything tied to AI without discrimination has largely passed. The note says the current drawdown is forcing investors to sort semiconductor holdings by type rather than treat the entire AI chain as one basket. RockFlow groups relevant names into four categories: “toll-taking” assets such as TSMC, shovel-selling leaders such as Nvidia, cyclical high-beta names including parts of memory, equipment and optical module suppliers, and pure narrative stocks with weak fundamentals. The report also argues that higher cloud capex is no longer an automatic positive. Markets now want clearer evidence that spending by Microsoft, Google, Meta and Amazon can turn into durable, high-quality revenue and justify the pressure on free cash flow. That leaves earnings season as a tougher test than many investors expect, with guidance, monetization and return on investment all under closer scrutiny.

CertiK researchers said they found two security flaws in Google’s EdgeTPU, tracked as CVE-2026-0150 and CVE-2026-0153. Google acknowledged the findings and listed them in its June 2026 Security Bulletin, rating the bugs High and Critical. Beyond the disclosure itself, the research points to a broader shift in how companies need to think about AI security. The report argues that as AI systems move from generating content to carrying out tasks, the security focus can no longer stay limited to model behavior alone. Enterprises are now deploying AI across identity verification, facial recognition, edge inference and other business-critical functions, while AI agents are increasingly connecting to databases, APIs and third-party tools. In that setting, risk can emerge not only inside a model, but across the interfaces and trust relationships that link infrastructure, applications and external systems. CertiK also ties this change to a wider industry trend. It cites McKinsey’s “The state of AI in 2025,” which says 88% of companies have deployed AI in at least one business function and more than 60% have begun exploring AI agents. A separate Google Cloud survey found 83% of respondents believe major infrastructure upgrades are needed to support AI agents at scale. The company says the practical implication is clear: securing AI now means validating the full system, from development and deployment to runtime operations.

Wall Street entered the Federal Open Market Committee meeting with a defensive tilt, as investors rotated out of crowded AI hardware positions and into software names, cash-flow-heavy blue chips, and more traditional sectors. Overnight, the Dow Jones Industrial Average rose 1.03%, the S&P 500 added 0.21%, and the Nasdaq Composite slipped 0.22%. The move came as traders cut duration, reduce exposure to crowded trades, and dial back dependence on narrative-driven positioning before both the Fed decision and a fresh wave of major tech earnings. At the same time, geopolitical tension returned to the foreground. U.S. Central Command said Iran launched multiple ballistic missiles at a U.S. base in Jordan, with all of them intercepted. The attack marked Iran’s first renewed strike on a U.S. target since Washington paused attacks last Friday. Oil responded quickly, with WTI and Brent crude both rebounding more than 4% intraday after three straight sessions of declines. Gold failed to rally on the renewed tension. Reuters’ latest survey showed analysts cut their gold forecasts for the first time since late 2023, while CME data showed open interest in fed funds futures tied to the current meeting hit a record 967,000 contracts. In equities, semiconductor, memory, and optical communications shares sold off sharply, while software stocks and companies such as Apple, Coca-Cola, and Boeing attracted inflows.

Google disclosed last week that it has agreed to cover as much as $44 billion in third-party data center lease payments if tenants default, according to a report by The Information cited by BlockBeats on July 29. The figure marks a sharp increase from $6.5 billion as of the end of September. The report said the move reflects Google’s push to offer companies including Anthropic an alternative to Nvidia AI chips through its TPU lineup. Google has taken on what the report described as lease “backstop” commitments to help expand TPU sales. Two sources said the company made those commitments on the view that revenue from TPU sales would exceed the financial obligations tied to guaranteeing the data center leases needed to house the chips. One of the sources said executives believed the financial math worked in Alphabet’s favor. The report added that Alphabet has provided lease guarantees for roughly 10 projects totaling about 2.4 gigawatts of capacity. None of those projects has been completed, meaning the guarantees have not yet taken effect.

Nvidia CEO Jensen Huang used what the source describes as his second-ever X post to announce the Open Secure AI Alliance, a new coalition with 37 founding members spanning chipmakers, cloud companies, security firms, and open-source communities. The alliance argues that closed and open models should coexist, but says open models and testing harnesses are essential for cybersecurity work. The source draws a sharp contrast between the alliance roster and an earlier open-weight letter backed by 32 companies. Nvidia, Microsoft, Meta, Cisco, Dell, Hugging Face, IBM, Palantir, and SpaceXAI are described as backing both efforts. OpenAI and Google later signed the public letter but did not join the alliance. Anthropic, according to the report, did neither. The article also links the alliance’s case to a July security incident disclosed by Hugging Face and later claimed by OpenAI on July 21. In that account, commercial frontier models refused to analyze attack logs because of guardrails, while Hugging Face used the open-source model GLM 5.2 on its own infrastructure to review more than 17,000 actions and cut forensic work from days to hours. The source further argues that Nvidia’s push for open models also aligns with its interest in expanding the market for local GPU computing as major AI customers pursue in-house chips.