On Tuesday, April 14, the computing power industry chain experienced a major surge, with the ChiNext artificial intelligence sector continuing its upward breakthrough. Leaders in CPO optical modules maintained strong momentum, with Zhongji Innolight rising 4.74% to set another record high. Lantech and Changxin Bocon gained over 2.5%, while Eoptolink increased by more than 1%. The computing power leasing concept also saw collective strength, with Xiecong Data hitting a 20% daily limit increase. Yhualu and G-Horse Network, among others, rose more than 4%.
Among popular ETFs, the largest and most liquid fund in its category, the ChiNext Artificial Intelligence ETF (159363), surged 2.9% to reach a historic high, marking its fifth consecutive daily gain. Trading volume exceeded 900 million yuan, with net subscriptions reaching 108 million shares for the day, indicating strong investor interest.
In the CPO optical module sector, overseas optical communication leaders are reporting explosive order growth, leading to global capacity constraints. Industry sources indicate that Lumentum, a leading U.S. optical communications company, stated that demand from tech giants for its optical components is accelerating. At the current rate, full-year 2028 production capacity will be completely sold out within the next two quarters.
Analysts note that the market has refocused on the AI industry theme, with Lumentum's capacity being booked through 2028 demonstrating resilient sector demand. Key areas for focus include: 1) Optical modules, where 2027 demand is becoming clearer, and leading companies show attractive valuation metrics; 2) CPO technology, where NVIDIA's GTC conference clarified implementation timelines, with scale-out scenarios entering mass production first and scale-up scenarios launching in 2028, benefiting upstream core component suppliers.
Regarding computing power leasing, prices are rising across the board, with leading companies forecasting substantial profit growth. Data from New York provider Ornn shows spot leasing prices for NVIDIA's full GPU series in cloud data centers have increased significantly in recent months. Additionally, Xiecong Data announced expectations for Q1 2026 net profit between 650 million and 850 million yuan, representing year-on-year growth of 284.14% to 402.33%.
Market observers highlight that major price adjustments in cloud services, combined with supply-demand imbalances, are driving computing cost increases. Major providers including Alibaba Cloud, Tencent Cloud, and Baidu AI Cloud have implemented price adjustments, benefiting domestic AI chip manufacturers and computing power leasing companies. Furthermore, rising AI service costs will likely pass through to downstream application development, potentially accelerating industry consolidation and efficiency improvements in computing resource utilization.
Over a longer horizon, computing power segments including optical modules and CPO have accelerated breakthroughs since April. The underlying index of the ChiNext Artificial Intelligence ETF (159363) has gained over 176% in the past year, significantly outperforming comparable AI indices. Looking ahead, driven by both computing power supply-demand imbalances and exceeding global AI capital expenditure expectations, the ChiNext AI sector with heavy weighting in optical modules and computing infrastructure appears well-positioned for continued benefits.
The ChiNext Artificial Intelligence ETF tracks the ChiNext Artificial Intelligence Index, which has a base date of December 28, 2018, and was launched on July 11, 2024. The index's annual performance from 2021 to 2025 was 17.57%, -34.52%, 47.83%, 38.44%, and 106.35% respectively. Index components are adjusted according to its methodology, and past performance does not indicate future results.
As AI transitions from training to multimodal applications, computing power demand is growing non-linearly. Optical modules, as core global computing infrastructure, maintain high growth prospects supported by both performance and order visibility. The ChiNext Artificial Intelligence ETF (159363) and its feeder funds allocate approximately 70% to computing power and 30% to AI applications, representing both computing core and AI implementation segments.
Market data indicates bullish technical signals are emerging across multiple stocks in the sector.
Comments