Cap (CAP) Surge

SuperPure Applied Materials (Stock Code: 301717) debuted on the Shenzhen Stock Exchange's ChiNext board. The company issued 25.4615 million shares at an offering price of 65.99 yuan per share, raising approximately 1.68 billion yuan. The stock opened at 450 yuan, soaring 582% from the IPO price, and closed at 503 yuan, up 662%, giving the company a market capitalization of 51.2 billion yuan. The company specializes in special coating processes and related technologies and materials, providing precision components and services for the semiconductor manufacturing and precision optics sectors. Its products cover core equipment parts for wafer fabrication, packaging, and silicon wafer manufacturing, with notable technical advantages in etching, lithography, measurement/inspection, annealing, and thin-film deposition equipment. Financially, SuperPure reported revenues of 169 million, 260 million, and 496 million yuan for 2023, 2024, and 2025 respectively, with corresponding net profits of 64.8 million, 82.26 million, and 185 million yuan. For Q1 2026, revenue reached 147 million yuan, a 62% year-over-year increase. The company forecasts H1 2026 revenue between 270 to 290 million yuan and net profit between 103 to 110 million yuan. The company's controlling shareholder and actual controller is Chai Jie, who directly and indirectly controls 48.23% of the voting rights. Together with his brother Chai Lin, a concerted action person, they control a total of 68.84% of the voting rights. Major pre-IPO shareholders included GSD Venture Capital, BYD, and AMEC.

"Chips Giant with 700 Billion Market Cap Sees Inventory Soar" In its 2026 first-half financial report, Cambricon, a leading Chinese AI chip design company, reported significant growth in both revenue and profit. Revenue reached 59.96 billion yuan, up 108.13% year-on-year, while net profit grew 122.61% to 23.11 billion yuan. However, a major point of investor focus was the sharp rise in the company's inventory, which surged 66.83% from the end of the previous year to 82.48 billion yuan. This inventory now represents 45.32% of its total assets. The increase is primarily attributed to a substantial growth in raw materials (up 77.95% to 52.78 billion yuan) and work-in-process with contract manufacturers (up 61.85% to 24.18 billion yuan). Cambricon management explained that the inventory build-up was due to increased purchases of raw materials and outsourced processing. The company acknowledged the risk of inventory impairment should market conditions change. Industry analysts noted that as a fabless chip designer, such inventory accumulation is a strategic move to secure wafer capacity. However, it carries significant risk due to the industry's rapid technological iteration. If downstream demand slows, the company could face substantial inventory write-downs. Additionally, Cambricon's prepayments soared nearly threefold to 29.14 billion yuan. Brokerage analyses interpret the concurrent rise in prepayments and inventory as a positive indicator, suggesting strong future order fulfillment and pointing towards significant revenue growth in the third quarter. Despite strong operational metrics, the company's cash and cash equivalents have shown a declining trend over the past three years, dropping to 6.26 billion yuan by mid-2026. Cambricon's stock price hit a historic high of 1,620 yuan per share in June, briefly pushing its market capitalization above one trillion yuan. As of August 13th, its share price was 1,105.5 yuan, with a total market cap of approximately 694.6 billion yuan.

Yushu Technology is going public with a market valuation of 60.9 billion yuan and a P/E ratio of 219x, raising over 6 billion yuan. At least 12 securities firms had indirectly invested before the IPO via limited partner (LP) models, primarily through top-tier PE funds. During the subscription phase, 34 securities firms participated in the offline share placement. As the lead underwriter, CITIC Securities played four roles: an early-stage investor (achieving ~1300% return), a mandatory co-investor in the科创板 listing, the underwriter earning fees (~145 million yuan), and the manager of the employee strategic placement plan. This represents deep, multi-layered involvement. For retail investors, the article highlights three key points: 1) "Indirect investment" through LPs is standard practice for many securities firms. 2) The total exposure of the securities sector to Yushu is significant, combining both pre-IPO stakes and IPO subscriptions. 3) With a 219x P/E ratio, underwriters, their investment arms, and clients are all heavily exposed, creating shared risk. The performance in 2026, with projected H1 revenue of 1.05-1.13 billion yuan, will be crucial to validate the high valuation and the humanoid robotics narrative.

Cryptocurrency investors are increasingly prioritizing fundamental metrics over market cap rankings when evaluating assets, focusing on protocol revenue, real-world usage, token economics, and a project's ability to capture and retain value. Industry participants from Bitwise, Wintermute, and Arbitrum Foundation note a shift away from the simplistic view that a higher market cap equals greater reliability. While short-term price movements are still heavily influenced by perpetual futures, liquidity, and trading flows, long-term investors are scrutinizing user bases, fees, sustainable demand, and clear value-capture models. Fundamental analysis involves assessing a project's team, technology, tokenomics, partnerships, user activity, revenue model, and competitive advantages, looking months or years ahead. Technical analysis remains relevant for short-term trading. Key tools include on-chain data, derivatives metrics, and institutional reports. Market cap is no longer the primary benchmark. Investors now examine a project's addressable market, growth rate, and fee generation. Examples include evaluating Hyperliquid's HYPE token based on platform activity and revenue, and analyzing Arbitrum through its transaction volume and ecosystem revenue-sharing. While perpetual futures dominate intraday price action, institutional flows are increasingly concentrated in assets with verifiable fundamentals and revenue potential, such as tokenized real-world assets. Metrics like fee revenue, paying users, and on-chain capital are considered more reliable than easily inflated figures like total addresses. Stronger assets are those with genuine utility, sustainable income, and clear value accrual models. Bitcoin remains a distinct macro asset. Ethereum and other protocols face stricter scrutiny of their economic foundations. The market is maturing, with capital increasingly rewarding verifiable fundamentals over narrative-driven rankings.

Chinese AI giant SenseTime, once a star in the AI "Four Dragons," recently projected its first-ever profit for H1 2026, sparking a temporary stock surge. However, its current market cap of ~HK$64.6 billion remains over 80% below its peak of ~HK$300 billion in early 2022. The article analyzes SenseTime's dramatic fall from grace. It excelled in the "AI 1.0" era, dominating computer vision for applications like facial recognition and smart cities. However, it was disrupted by the "AI 2.0" generative AI revolution led by ChatGPT, which shifted focus from recognition to creation. Compounded by US sanctions, the death of its founder, a short-seller report, and a decline in its core smart city business, the company faced a perfect storm. Its workforce was cut by nearly 60%. To survive, SenseTime pivoted, actively "killing" its old self. It transitioned from project-based solutions to a generative AI and visual AI dual-engine model. Now, over 70% of its revenue comes from large language models, a more scalable, subscription-like business. Cost-control measures on expensive model training have also contributed to its path to profitability. Yet, significant challenges remain. The current AI race is an ecosystem battle dominated by giants like Microsoft/OpenAI, Google, and Chinese tech firms with integrated clouds, apps, and vast user bases. SenseTime, as an independent AI company, lacks such a super app or traffic gateway. The key question is whether it can build a sustainable moat based solely on model capability and industry deployment in this new competitive landscape.