Xin Lang Cai Jing·2025-11-30 11:43Group 1 - The core viewpoint is that the computing power industry is in a high prosperity cycle, driven by the rapid growth of demand due to continuous innovation in AI models [1] - Major companies like Google and NVIDIA have reported earnings that exceeded expectations, indicating a positive feedback loop between model innovation and computing power construction [1] - North American giants are increasing capital expenditures significantly, with Meta raising its forecast for capital spending from $66-72 billion to $70-72 billion, Google from $85 billion to $91-93 billion, and Amazon's expected capital expenditure around $125 billion for the year [1] Group 2 - Leading optical module manufacturers are expanding production capacity both domestically and internationally to meet explosive demand for computing power [2] - Companies like NewEase and Zhongji Xuchuang are establishing production bases in Thailand to optimize cost structures and mitigate geopolitical risks, with significant expansions planned for 2025-2026 [2] - Financial data supports the intensity of capacity expansion, with Zhongji Xuchuang reporting a 1765.7% increase in construction projects compared to the beginning of the year [2] Group 3 - A new round of concentrated capacity release is expected in Q1 2026, which will open up growth opportunities for the industry [3] - The optical module industry is facing supply bottlenecks due to rapidly growing demand, prompting leading manufacturers to adopt a dual-line expansion strategy in mainland China and Thailand [3] - The industry is transitioning from "capacity release" to "performance climb," with expectations of steady performance growth as production scales up and processes mature [3] Group 4 - Companies to watch in the computing power sector include leading optical module firms like Zhongji Xuchuang and NewEase, as well as other key players in the optical device market [4] - The focus is also on domestic computing power supply chains, including liquid cooling segments and edge computing platforms [5]