Analyst: Bitcoin’s daily chart shows hidden bearish divergence, forming a key resistance near $80,000

Citrini analyst Jukan stated in a recent report that, per his sources, NVIDIA’s upcoming Rubin Ultra may have its HBM specification downgraded from 12 layers of HBM4E to 8 layers. Clients including OpenAI and Anthropic had even requested a 4-layer HBM product, but memory manufacturers rejected the request, with the current downgrade likely stopping at 8 layers. The analyst attributes the adjustment primarily to yield and cost pressures: if Rubin Ultra uses 12-layer HBM4E, combined with price hikes, memory costs could account for roughly 70% of the accelerator’s total bill of materials (BOM). Software optimizations such as model quantization, multi-head latent attention (MLA), and compute task offloading are shifting low-frequency KV caches and model states to LPDDR, CXL, and NAND, leaving HBM to only hold the working set required for real-time computation. As a result, after meeting minimum capacity requirements, customers now prioritize HBM bandwidth over capacity. Jukan added that reducing stack layers would improve packaging yield, lift HBM and AI accelerator shipments, and may even expand total HBM demand. Meanwhile, higher bandwidth requirements reduce the share of chips passing speed binning on wafers, further straining DRAM wafer capacity. Long-term, HBM will eventually be replaced by new architectures, with the ultimate direction being the integration of memory and logic chips. The next two years will be a critical period for memory manufacturers to expand into the logic sector.

Solana Foundation Technical Vice President Jacob Creech has published a post announcing that Solana is set to undergo several key changes. Next week, the network will take the first step toward rent reductions. Transaction V1 will officially launch on September 9, while slot time (the interval between block productions) will be further shortened. Additionally, Solana plans to roll out the Alpenglow upgrade in October and host the "Scale or Die" event in November. Creech noted that with these changes implemented, "Solana’s development will be fundamentally different moving forward."

Robinhood's blockchain token issuance platform Pons stated in a post that it has cumulatively paid $20.93 million to token creators over the past 47 days.

BitMine Chairman Tom Lee told CNBC in an interview that Nvidia’s strong earnings report pushed its stock price higher, breaking the recent pattern where positive earnings results failed to lift share prices, signaling investors still value corporate fundamentals. Software stocks including Salesforce, CrowdStrike, and Okta also rallied, reflecting the market’s positive response to downstream AI deals, with overall market reactions remaining healthy. Regarding the underperformance of AI-related stocks such as Meta, Amazon, Alphabet, AMD, and Micron, Tom Lee noted that some investors who had underweighted Nvidia may be selling other tech stocks to raise funds to increase their Nvidia holdings. He pointed out that Nvidia’s profit expectations have been significantly revised upward, yet its stock price has not fully caught up, with its price-to-earnings ratio still contracting, meaning its current valuation remains low. Tom Lee also stated that data center construction is gradually becoming a political issue in the U.S. midterm elections, and even some Republican-led states and states that support data center development are starting to consider pausing related projects. This could be one reason for the recent poor performance of AI infrastructure-related stocks, as funds are shifting more toward downstream AI assets.

According to on-chain analyst Ai Yi (handle @ai_9684xtpa), address 0x90c…888ac spent roughly $1 million to purchase 12.25 million "Niu Lai" tokens over the past hour, re-establishing a position in the related token. During the buy, "Niu Lai"’s market capitalization briefly topped $90 million. This is not the address’s first trade of "Niu Lai": it opened a position 13 days ago, then sold out entirely at $0.03162 five days ago, marking a re-entry after liquidating its holdings.

Beating AI Express News: Whistleblower Leo states OpenAI’s next major model Astra has moved from internal dogfooding to early partner testing. A small number of partners are currently seeing it under the name “ultima-alpha”. If feedback this weekend is positive, OpenAI plans to expand early access next week and further widen the availability scope by the week ending September 3.