PURR rose more than 20%, MSTR gained 12%, and US-listed crypto concept stocks extended their rally.

The hacker behind the GTA 6 leak launched the Solana meme coin CYBERLEEK using unreleased Grand Theft Auto 6 gameplay footage. The token reached a peak market cap of roughly $25 million on August 23. Hours before Rockstar Games officially unveiled GTA 6 gameplay for the first time, on-chain data showed the token’s contract owner withdrew approximately $146,000 in Wrapped SOL and 15.4 million tokens in fees, then converted all tokens to about $125,000 worth of SOL. Data indicates the funds were later split into four new wallets, with at least 91 SOL transferred to KuCoin. Independent on-chain analysts estimate the total withdrawal amount may exceed $250,000, though the figure has not been confirmed by the leaker or the exchange. CYBERLEEK previously hit a high of $0.0344, and is now trading at around $0.0097, down roughly 46% in 24 hours, with its market cap falling to approximately $7 million. Rockstar Games has responded to the GTA 6 leak, saying the incident has left the development team "heartbroken". Separately, Take-Two is reported to have issued subpoenas to X, Microsoft, and Discord to identify the leaker.

SK Hynix’s CEO stated that no significant signs of a downturn in the memory market have been observed, and expects memory shortages to persist until the end of 2030.

Beating AI Express, as reported by The Information, the Trump administration’s recent executive order aimed at establishing a new AI self-regulatory body has stalled. Sources familiar with the matter said the White House has circulated drafts of the relevant executive order internally in recent weeks, but the plan has not advanced further. The Trump administration had previously pushed for a unified AI regulatory framework and sought to prevent U.S. states from enacting conflicting AI rules. If this new regulatory body plan remains stalled, it could further delay the development of a federal-level AI industry regulatory system in the U.S.

Bitcoin has rallied around 23% in a single week recently, driving a sharp rebound in the stocks of Bitcoin mining firms that had underperformed earlier. Some mining stocks have even outperformed AI infrastructure-related equities, indicating a renewed rise in investor attention to direct Bitcoin exposure. Data shows that Canaan, American Bitcoin, and Cango have gained between 41% and 67% recently. In comparison, CoreWeave rose around 21%, Nebius climbed 17%, and IREN increased 15%, while some mining stocks that had shifted to AI and high-performance computing (HPC) businesses earlier were largely flat or declined. BlocksBridge Consulting attributes Bitcoin’s recent rally to three main factors: the U.S. Treasury expanding its long-term U.S. Treasury bond repurchase scale, the Trump administration pushing Congress to pass the crypto market structure bill CLARITY Act, and a short squeeze triggered by Bitcoin’s breakout, which led to over $1.6 billion in crypto market liquidations within 24 hours. Notably, despite mining firms shifting to AI and HPC infrastructure in recent years, Bitcoin prices still have a significant impact on the performance of mining stocks. A previous analysis by BlocksBridge shows that since 2026, the AI and HPC business revenues of nine listed mining firms have reached around $341 million, while their related capital expenditures have hit $5.11 billion—meaning that for every $1 of AI-related revenue, mining firms have spent an average of about $15.

Crude oil shipments through the Strait of Hormuz are slowly recovering, but a notable gap persists between the accelerated loading of Gulf oil producers and actual imports to Asia, leaving the market skeptical about the authenticity of the supply recovery. Traders currently estimate that around 6 million to 8 million barrels of crude oil pass through the strait daily, roughly half the pre-conflict level. Saudi Arabia, Iraq, Qatar, and Kuwait have all seen a rebound in loading activities recently, with Iraq’s daily loading capacity once exceeding pre-conflict levels. However, some crude oil still requires transshipment via locations such as Oman’s Sohar Port and the UAE’s Fujairah, so increased loading does not mean the oil has actually reached Asian buyers. Data from Kpler shows that Asia’s crude oil imports in August are projected to hit 23.12 million barrels per day (bpd), about 14% lower than the average of 26.91 million bpd in the three months before the conflict. Among these, India’s August crude oil imports are expected to be only 4.51 million bpd, with around 1.45 million bpd coming from the Middle East—just half the pre-conflict average. Additionally, Middle Eastern crude oil arriving in Asia in August stands at around 11.11 million bpd, higher than July’s 10.76 million bpd but still nearly 30% below the pre-conflict level of 15.82 million bpd. Analysts note that the current gap between loading volumes and arrival figures may partly result from shipping delays, and data in the coming weeks will be critical to determining whether supplies through the Strait of Hormuz have truly recovered.

According to CNBC, prominent financial host Jim Cramer said the core reason for the accelerating demand for Nvidia’s AI chips is that downstream clients can now generate profits quickly via AI infrastructure, with returns on investment no longer a distant expectation but an ongoing reality. Nvidia CEO Jensen Huang previously stated on an earnings call that some $500 billion worth of data center projects have seen their capital return periods shortened to less than a year. Cramer believes this means the market’s debate over when AI infrastructure will turn profitable is gradually being resolved. Additionally, Amazon Web Services (AWS) announced it will purchase an extra 2 million Nvidia GPUs between 2027 and 2028, deploy Vera CPUs, and apply related technologies to the robotics sector. Notably, even as Amazon actively develops its own in-house AI chips, it continues to expand its procurement of Nvidia products. Nvidia CFO Colette Kress noted that the company’s quarterly growth will be driven primarily by non-hyperscale cloud provider clients, including "new cloud" firms like CoreWeave and Nebius, as well as enterprise customers. Cramer pointed out that roughly 50% of Nvidia’s clients are not hyperscale cloud providers, indicating AI computing power demand is spreading further into the broader enterprise market.