Plans by Iran to restrict passage of "hostile" vessels through the Strait of Hormuz are lifting oil prices and increasing geopolitical uncertainty in the energy sector. Bitcoin once again tested a long?running downward trend line amid minimal spot volumes and large inflows of BTC to exchanges, which may determine the near?term direction of the crypto market. A Counterpoint report showed that more than 92% of sovereign large?language models run on Nvidia GPUs, highlighting a critical concentration of hardware in the AI space. Finally, Google's negotiations over a multibillion?dollar deal with startup Mechanize show how the tech giants are using hiring and licensing to accelerate development of AI tools for programming. Together, these stories reflect a mix of geopolitical risk, technological centralization, and trading opportunities for investors and market participants.

Oil is climbing on Iran's plans to ban US and Israeli ships from the Strait of Hormuz

On Friday, oil prices rose for a second day after Iranian authorities published a draft plan for managing the Strait of Hormuz. The document proposes banning passage through the strait for ships Tehran deems "hostile" (explicitly naming US and Israeli vessels) and imposes heavy fines for violations. That effectively pushes back prospects for a quick return to normal shipping through one of the world's key energy chokepoints.

Brent climbed above $83 per barrel, while US WTI is trading around $78, with WTI's gain on Thursday exceeding $3 per barrel. The upward movement began after the Fars news agency published the "initial draft of a strategic plan for managing the Strait of Hormuz."

The draft's key provisions include a ban on passage for vessels classified as "hostile" and fines of up to 20% of the cargo's value for violators. Transit fees would be 5–7% of cargo value for ships using the strait (Oman is reportedly considering its own fee of roughly 3%).

The United States is already demanding the complete removal of such fees. Iranian lawmaker Behnam Saidi stressed that the agreement reached with Oman "does not mean reopening the Strait of Hormuz." He said the strait will remain closed until the threat to Iran disappears and the war is officially declared over.

Analysts say the market is now pricing not a return to full normalcy but something like a managed or conditional corridor for shipping rather than a full restoration of prior traffic. According to Lin Ye of Rystad Energy, investors and traders are assessing the likelihood of partial, controlled transit rather than free and open passage.

For traders, these developments create volatility?driven trading opportunities: strategies targeting rallies or corrections in oil contracts and oil?related instruments can be considered.

If you want to act, note that the instruments discussed are available for trading on the InstaForex platform. To take advantage of the market situation, you should open a trading account with InstaForex and download the company's mobile app for convenient position management and rapid reaction to news.

Bitcoin tests the key descending trend line of the bear market

Bitcoin has once again approached the descending trend line that has capped every meaningful rally since the record high above $126,000 set in October 2025. Traders believe this could be a decisive moment for the market's next direction. The test is taking place against the backdrop of spot trading volumes falling to the lowest level of 2026.

On Thursday, August 6, the price of Bitcoin briefly approached $65,000, holding support above $64,500 and running into that same bear?trend line that has restrained the price for almost ten months. A similar attempt in early May failed for the bulls and led to a drop to $60,000, so that prior scenario remains fresh in traders' minds.

Technically, the situation is unfolding according to two main scenarios:

Scenario one — a confident break of the resistance line, which would open the way to $65,600 and could complete an inverted head?and?shoulders pattern.
Scenario two — a failure to advance and a pullback that would extend the current bear trend. The decision will most likely determine the near price range.

A worrying factor is that spot trading volumes have fallen to the lowest level of 2026, a state one analyst called a "ghost?town." At the same time, more than 20,000 BTC — roughly $1.28 billion — have flowed onto exchanges over the past seven days. That is a classic signal that large holders are preparing liquidity or selling.

Mika?l van de Poppe of MN Fund sees the low volumes not only as a risk but also as an opportunity. In his view, a calm market gives a chance to accumulate. He also noted that the share of Bitcoin supply in loss has reached a historic high, a metric that has previously preceded major rallies (for example, the roughly 600% rise from the 2022 cycle lows).

What can traders do? First and foremost, watch the key levels and volumes, and be sure to use risk?management strategies (gradual accumulation / scaled?in positions, stop?losses). It is important to use low?liquidity windows for entries while maintaining discipline.

Nvidia powers 92% of sovereign AI models, report says

Countries that aim to build independent AI systems still overwhelmingly rely on a single U.S. chipmaker. According to a Counterpoint Research report published on August 5, Nvidia GPUs power 92.4% of sovereign large?language AI models worldwide.

The "Sovereign AI LLM Index" study covered more than 170 models from 80 regions. Analysts identified over 55 countries across the Asia?Pacific, the Middle East, Europe, Africa, Canada and Latin America (excluding the U.S. and China) that are intentionally developing their own AI infrastructure. But when it comes to chips, everything boils down to the same player.

AMD holds second place with a 4.1% share, and Cerebras 1.7%. The rest of the market is effectively in the "long tail." Counterpoint Research's chief analyst Mark Einstein says plainly that while sovereign hosting and localization are moving forward, chip purchases do not follow that trend: "Most domestic LLM developers are being pulled into the NVIDIA CUDA ecosystem and use NVIDIA chips to train their LLMs."

Why does this matter? Because political and economic initiatives to create independent AI systems may be vulnerable if the core hardware base remains centralized. Even if countries want to move away from the influence of big tech, they must consider the availability and effectiveness of solutions that have already proven workable for training large models.

So, despite efforts toward technological independence, 92.4% of sovereign LLMs are trained on Nvidia GPUs — effectively a near monopoly. AMD and Cerebras occupy only minor shares at 4.1% and 1.7%, respectively.

Localizing AI hosting is not the same as localizing chip supply chains — software and hardware independence still diverge. Political ambitions to build national AI platforms may be constrained by practical limits on access to specialized processors.

For traders, this looks like a market with clear leaders and speculative opportunities: concentrated demand for Nvidia products and limited competition create focal points for watching stock prices, volatility, and news flow in the semiconductor and AI sectors.

Google in talks on $1.5 billion deal with startup Mechanize

Google is discussing buying Mechanize's technology and hiring parts of the startup's team, rather than pursuing a classic acquisition. According to Business Insider, the talks envisage hiring some employees and obtaining a non?exclusive license to Mechanize's technology. The deal size is substantial — reportedly in excess of $1.5 billion. Sources say details are not final and may change.

What is Mechanize and why does it matter? Mechanize is a San Francisco startup founded about a year ago that builds reinforcement?learning environments to test and train AI agents for programming. These environments evaluate models' ability to generate functions, deploy applications and debug code in unfamiliar codebases. The company currently employs fewer than ten people and is based on Third Street.

Why would Google structure the deal this way instead of a traditional takeover? The Mechanize transaction is a combination of talent hires and technology licensing. This model has become common in the tech industry: large players prefer such formats to integrate promising developments faster while reducing regulatory risk. Microsoft used a similar approach with Inflection AI in 2024, hiring much of the staff and obtaining a license to their technology.

...

Forex analysis review

Message History

Delivery option(s)