A commentary published by MarsBit sketches a widening split between China’s primary and secondary capital markets, using hard-tech projects and recent IPOs to show how money is being priced very differently across the two. The piece starts with a world-model startup that, according to the author, went from being a difficult, low-monetization AI bet last year to one of the hottest fundraising stories this year, with capital chasing anything tied to simulations of the physical world. The author says the project raised billions of yuan within half a year and is already discussing a new round at a valuation approaching the high hundreds of billions, a sign of intense fear of missing out in private markets. The article then broadens to listed names and the IPO queue. It cites Qingke Research data showing 5,944 equity investment cases in China in the first half of 2026, up 14.7% year over year, with total investment reaching about 565.4 billion yuan, up 31.9%. Yet, by the author’s account, roughly 90% of first-market money this year has concentrated in AI, robotics, world models, quantum technology, controllable nuclear fusion, integrated circuits and commercial aerospace. On the public-market side, several hard-tech listings are described as peaking at the open and then retreating quickly, including Unitree and Pinzhun Laser, while ChangXin Technology is presented as a contrasting case backed by much stronger earnings and lower valuation multiples. The article frames the gap as a test of how much hard-tech issuance public markets can absorb and whether “patient capital” is being practiced or merely invoked.

Vitalik Buterin has shared insights on recent research into Ethereum transaction formats, clarifying the distinction between 'actions' and 'dependencies' within transactions. According to Buterin, actions such as sending ETH can be optimized separately from dependencies like signatures, UTXO Merkle proofs, and ZK-SNARKs/STARKs. Dependencies can be processed in parallel, with state-dependent ones handled by the mempool and pure dependencies potentially replaced by STARK proofs, reducing execution and data processing overhead. Buterin noted that over 90% of Ethereum's activity does not require dynamic flexibility, suggesting that a clear separation between dynamic and statically analyzable parts is key to scaling. The evolving EIP-8141 transaction type is moving toward a generic, minimal call list, alongside new state types, recursive STARK mempools, and keyed nonces, all aimed at optimizing efficiency.

Vitalik Buterin posted on X that recent research on transaction formats has given Ethereum a clearer understanding of 'actions' and 'dependencies,' which can be optimized separately. Dependencies, including signatures, Merkle proofs, and ZK-SNARKs/STARKs, can be processed in parallel; over 90% of Ethereum activity does not require dynamic flexibility. EIP-8141 aims toward a generic, minimal call list, with new state types, recursive STARK mempools, and keyed nonces also part of this direction.

Crypto markets opened Sept. 4 with a clear macro catalyst: traders sharply cut expectations for a Federal Reserve rate hike this month. CME pricing showed the odds of a 25-basis-point increase at roughly 50%, down from 70% a day earlier, after comments from Fed Governor Christopher Waller and New York Fed President John Williams signaled that a hike is not a settled outcome. Bitcoin briefly moved above $82,000 in early trading, while most major tokens in centralized exchange volume rankings posted gains, including BTC, ETH, XRP, SOL, BNB, DOGE, UNI, and ZEC. The day’s policy picture was mixed. A Galaxy research executive said House Republican leadership had canceled voting weeks in late September, leaving the CLARITY Act with very low odds of passing before the midterm elections. At the same time, SEC Chair Paul Atkins told Fox Business he expects the bill to be voted on in the Senate on Sept. 15 and ultimately sent to the president for signature. Separately, G20 finance ministers and central bank governors backed a clearer regulatory path for digital assets and called for stronger anti-money-laundering enforcement in large virtual-asset jurisdictions. Elsewhere, Coinbase said it is working to launch single-stock perpetual futures in the U.S., Polymarket rolled out perpetual contracts more broadly, Robinhood Chain DEX volume hit a record $1.89 billion in 24 hours, and Nvidia agreed to acquire Hugging Face for $12.9303 billion.

PANews has published its August 2026 Top 10 column content ranking, a list compiled from all column articles on the platform during the month based on a combined assessment of content quality and readership. Tiger Research took two spots on the list, one on regulatory gaps in Asia’s prediction markets and another on capital outflows from South Korea’s crypto market. Other entries covered Binance’s attempt to build price discovery for U.S. stocks over weekends through Perp and bStocks, the fading wealth effect of exchanges, shifts in AI infrastructure bottlenecks, crypto market rotation after gains in Bitcoin and Ethereum, and the rise of RWA perpetual contracts as an exchange growth engine. PANews also said it created posters for the five most-read articles and used the release to invite more contributors to open columns on the platform. According to the announcement, high-quality submissions may receive homepage recommendations, app push exposure, banner placement, and distribution through core communities. PANews added that its September column ranking is expected to be released around October 5, 2026.

Yangtze Memory Technologies Holding Co., Ltd., or YMTC Holding, has moved a step further in its planned listing on Shanghai’s STAR Market. The Shanghai Stock Exchange website showed on Sept. 3 that the company’s IPO review status changed from "accepted" to "under inquiry," just 13 days after its application was accepted on Aug. 21. The company plans to raise 33 billion yuan, a figure that would set a new record for proposed fundraising on the STAR Market, topping ChangXin Memory Technologies’ 29.5 billion yuan and Semiconductor Manufacturing International Corp.’s 20 billion yuan. CITIC Securities and CSC Financial are serving as joint sponsors. According to Counterpoint Research, YMTC Holding reached a 14% share of global NAND bit shipments in the second quarter of 2026, surpassing Kioxia and Micron for the first time and ranking third worldwide behind Samsung Electronics at 25% and SK hynix at 22%. Shipment volume rose 22% year over year. The prospectus also shows that the company’s second-phase production line is already running at full capacity, while construction of a third-phase project is advancing. Of the funds to be raised, 20.8 billion yuan is earmarked for production line technology upgrades and 12.2 billion yuan for research and development-related projects.

Yangtze Memory Technologies Holding Co., Ltd. has seen the review status of its initial public offering and proposed listing on Shanghai’s STAR Market change to “inquiry received,” according to ChainCatcher. The status indicates the company has formally entered the Shanghai Stock Exchange’s listing review process, where it will face one or multiple rounds of written questions and answers focused on matters including information disclosure and its science-and-technology credentials. YMTC is described as China’s largest 3D NAND flash memory manufacturer and the country’s only original 3D NAND producer. The company has independently developed the Xtacking architecture and operates as an integrated device manufacturer, covering chip design, manufacturing, and packaging and testing.

Alibaba said it will place 710 million new shares at HK$112.70 each, raising HK$80 billion, with the net proceeds going entirely into full-stack AI capabilities and infrastructure. Over the past two months, Zhipu, MiniMax, Deepexi and QuantGroup have also tapped the market through IPOs, follow-on placements or new financings. The article argues that surging AI demand, tighter supply in memory and compute, and rising capital spending are pushing the sector into a far more capital-intensive phase. Alibaba’s move is notable because it comes despite cash of RMB 474.5 billion on its balance sheet at the end of June. Zhipu and MiniMax, both newly listed in Hong Kong, have already started to pursue A-share paths, showing how short the funding window may be for AI companies that need to keep buying chips, building data centers and funding model development. Upstream suppliers such as memory makers are capturing more of the economics, while application-layer AI firms continue to bear heavy losses.