交易量突破150亿美元，Solana HIP-3市场成第10个达标且首个未启用Growth Mode的市场
Related
Goldman Sachs: World Model May Become a New Engine for Future AI Infrastructure, Adding New Computing Demands
According to Mars Finance, Goldman Sachs stated in its latest report that "world models" may become the second engine of future AI infrastructure demand. Unlike large language models, which primarily process text and images, world models attempt to understand causal relationships in physical and social systems, such as simulating friction, material behavior, supply chain responses, policy shocks, or corporate competitive strategies. Physical world models will support robotics, logistics, autonomous driving, and industrial design; social world models may be used for strategic extrapolation, investment decisions, governance stress testing, and policy scenario analysis. Goldman Sachs believes that world models will not replace large language models, but rather superimpose new computing demands. If their development speed exceeds expectations, current investment forecasts surrounding computing power and electricity may still be underestimated. (Cailian Press)
Serenity: XFAB upgraded to "Buy"; AI and photonic chips may unlock long-term growth potential.
BlockBeats reported on July 6th that Serenity reposted Kelper's latest research report, upgrading semiconductor foundry XFAB's rating to "Buy." The report argues that Melexis demand is continuously improving, and the demand for silicon carbide (SiC), gallium nitride (GaN), and photonic chips driven by AI is expected to offset cyclical fluctuations in the automotive industry. Furthermore, the report notes that specialized foundries, scarce in the West, are receiving higher valuation premiums. Data shows that XFAB's current 2027 EV/EBITDA is only 4.8x, significantly lower than the industry median of 14.5x. Serenity stated that this indicates XFAB's core fundamentals are improving, but they are more optimistic about the growth potential of photonic chips, which the market has not yet fully priced in. They pointed out that XFAB leads the EU's PhotonixFAB project, which aims to build a European silicon photonics industry chain and promote the large-scale manufacturing of photonic chips. In the future, XFAB is expected to benefit from the growth in data center optical communication demand driven by AI giants such as Nvidia. However, since related commercial orders have not yet been implemented on a large scale, traditional valuation models still cannot reflect this long-term growth logic.
Large-scale model stocks in Hong Kong extended their gains, with MINIMAX and Zhipu rising over 16%.
Mars Finance reported on July 8th that MINIMAX-W (00100.HK) rose 17%, and Zhipu (02513.HK) rose 16%. In terms of news, Zhipu, a leading Hong Kong-listed large-scale model manufacturer, saw its first share lock-up period expire today, with several core institutional investors clearly choosing to continue their investment. (Science and Technology Treasure Broadcast)
Goldman Sachs: Buy on dips in chip stocks, but avoid "buying a basket of chips" again.
According to Mars Finance, Goldman Sachs stated in its latest report on July 7th that semiconductor stocks still present investment opportunities after the recent pullback, but AI chip trading has entered a more selective phase, and investors should no longer simply buy the entire sector. The bank pointed out that the PHLX Semiconductor Index has risen by over 80% this year, significantly outperforming the S&P 500 and Nasdaq indices. This strong performance has raised the bar for subsequent earnings realization and made the risk-reward ratio more differentiated ahead of the Q2 earnings season. Goldman Sachs remains optimistic about certain sub-sectors, including CPUs, ASICs, memory, and semiconductor equipment. Goldman Sachs believes these areas will benefit more directly from the expansion of AI infrastructure and have relatively higher demand visibility. In terms of individual stocks, Goldman Sachs specifically mentioned AMD and Applied Materials. AMD benefits from server CPU and AI-related demand, while Applied Materials benefits from advanced process technology and memory capital expenditure. However, Goldman Sachs is more cautious about the mobile phone supply chain and some semiconductor companies with high valuations or weak demand.
OpenAI Codex will integrate with the next-generation flagship model GPT-5.6 Sol Ultra.
According to Beating, Thibault Sottiaux, head of core products at OpenAI, confirmed on social media that the Ultra version of the next-generation flagship model, GPT-5.6 Sol, will be integrated into Codex. Previously, some users complained that OpenAI's decision not to include GPT-5.5 Pro in Codex was a major mistake; if GPT-5.6 Ultra were included, developers wouldn't even need to pay for Claude anymore. Sottiaux subsequently confirmed publicly that the Ultra version is indeed in Codex's plans.
Goldman Sachs released a report on China's AI computing power, predicting that domestically produced chips will account for over 50% of the market share by 2026.
According to a Goldman Sachs report, "China AI Computing Power," published by P Equity Research, China is accelerating the construction of its nationwide computing network. Related infrastructure projects are expected to attract 7 trillion yuan in investment by 2026, with data center investment reaching approximately 2 trillion yuan over the next five years. Currently, capital and technology are shifting massively to computing hubs in western China, while data centers in first-tier cities are transforming to focus on ultra-low latency computing, edge nodes, and AI inference. Although gigawatt (GW) clusters containing more than 100,000 chips are still scarce in China, in typical GW-level computing parks, workloads already consist primarily of inference (accounting for over half), along with training and full-stack R&D. The report predicts that by 2026, the market share of domestically produced AI accelerator chips is expected to exceed 50%. Huawei and Alibaba's Pingtouge lead the domestic camp with 20% and 7% market share respectively, but Nvidia currently maintains its overall market dominance with a 55% share. In terms of cost and performance, domestically produced chips have 40% to 50% lower capital expenditure per unit of IT power consumption compared to imported chips. However, due to the performance gap, their capital expenditure per unit of computing power is 2 to 4 times that of imported chips, and the computing power generated per unit of power consumption is only 10% to 30% of that of imported chips. In addition, the daily token output of Huawei's 910B/910C servers is about one-sixth to one-third of that of NVIDIA's H800, resulting in API profit margins based on this hardware lagging significantly behind those of competitors using NVIDIA hardware.