The race to build artificial intelligence (AI) infrastructure is fundamentally reshaping the global industrial landscape. The early market focused on computing semiconductors epitomized by Nvidia, and attention later shifted to the massive power infrastructure needed to operate data centers.
Recently, the "data bottleneck" has emerged as a key issue. No matter how fast the brain (GPU) and how abundant the power supply, AI performance inevitably deteriorates significantly if the infrastructure to connect them and transmit data is insufficient. This is the backdrop against which optical communication networks are rapidly emerging as a new investment theme, centered on the United States.
In the past, internal data center communications relied primarily on copper wires based on electrical signals, but AI-dedicated data centers are now rapidly transitioning to optical communications using light. As AI models grow larger, "clustering" technology that binds tens of thousands of GPUs into a single unit becomes increasingly important. To process the vast amounts of data generated in this process without delay, optical fibers and optical transceivers that overcome the physical limits of copper wire are essential. Recent capital expenditure trends among North American hyperscalers show that the proportion of optical communication components within network equipment is rising sharply. An "optical revolution" is underway, in which the pathway connecting chips and servers is shifting from copper to light.
This shift is also confirmed by the movements of global companies. Optical communications firm Lumentum has already filled most of its supply volume through 2028, and Meta has signed a contract with Corning to supply optical cables worth up to $6 billion through 2030. This is because expanding optical communication networks is essential to handle data traffic that is growing to encompass video AI.
From an investment perspective, the optical communications value chain exhibits high growth potential similar to semiconductors, while also offering the stability of an order-based industry tied to large-scale network construction. That is why it can be viewed as an attractive theme that secures both growth potential and downside stability.
However, it is not easy for individual investors to directly invest in the U.S. optical communications value chain. Not only is the technology itself difficult to understand, but analyzing and purchasing individual companies one by one amid a rapidly changing market environment entails considerable inconvenience and risk.
In such cases, it is worth paying attention to exchange-traded funds (ETFs) that allow investors to invest in core U.S. AI optical communications companies all at once. In particular, ETF investment has the advantage of diversifying individual company risks while tracking the growth of the entire sector, making it a suitable method for investing in the optical communications value chain, which has high technological barriers and rapid growth.
The flow of AI investment is expanding from "points" to "lines." Beyond the performance of individual chips (points), we have entered an era in which the efficiency of the networks (lines) that organically connect those chips determines the value of the entire system. It is time to take note of the fact that a new investment map is being drawn on the "highway of light" through which data flows.
