CPO penetration rate continues to rise, Tiantong shares and other stocks lead gains, Nvidia's layout attracts attention

robot
Abstract generation in progress

In the capital markets, concept stocks related to CPO have recently been active. Multiple stocks such as Tiantong Shares and Guangxun Technology hit their daily limit-ups, while companies like Realsil, Ruijie Networks, Hengtong Optoelectronics, Huagong Technology, and FiberHome Communications also saw their stock prices rise. Behind this phenomenon is the continuous increase in demand for high-speed data transmission in the AI computing power sector, as well as the growing importance of optical technology in data center construction.

According to the latest report from TrendForce, NVIDIA’s next-generation AI computing cabinet architecture plan indicates that future GPU designs will focus on increasing chip interconnection density and data transfer speeds. Specifically, cabinet-level chip interconnection (Scale-Up) and large-scale interconnection across cabinets (Scale-Out) are becoming key topics in data center planning. Traditional copper cable electrical transmission solutions are limited physically and cannot meet the needs of ultra-large-scale data transfer, creating opportunities for optical transmission solutions. The organization predicts that CPO (Co-Packaged Optics) will continue to grow in penetration within AI data center optical communication modules, reaching 35% by 2030.

NVIDIA’s recent investments in optical technology further confirm this trend. The company announced a $2 billion investment in Lumentum and Coherent, along with multi-year procurement agreements to secure priority supply of advanced lasers and optical products. This move is seen as a strategic reserve for NVIDIA’s key components of Scale-Up optical interconnects, marking its deep involvement in the R&D of lasers and optical components. The reliance of future computing infrastructure on optical technology is expected to further increase.

The upcoming NVIDIA GTC conference, a major indicator in the global AI computing power sector, is about to commence. It is reported that GTC 2026 will be held from March 16 to 19 in San Jose, California. NVIDIA founder Jensen Huang revealed that a “world-first” new chip will be showcased. Industry speculation suggests that the conference will focus on the new generation GPU architecture Rubin platform, the next-generation Feynman architecture GPU, as well as key technological advancements such as CPO switches, power architecture upgrades, and liquid cooling. These innovations are expected to further drive the upgrade and iteration of AI computing infrastructure.

Guosheng Securities analysts pointed out that with the continuous growth in AI computing power demand and the upcoming GTC conference, market attention to the optical communication and computing infrastructure industry chain is increasing. Investors are advised to focus on high-speed optical modules, optical communication segments, optical devices, and optical interconnection supporting fields. Breakthroughs in technology and commercialization in these areas will be key drivers supporting the sustained development of AI computing power.

View Original
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
  • Reward
  • Comment
  • Repost
  • Share
Comment
Add a comment
Add a comment
No comments
  • Pin