As the artificial intelligence (AI) industry advances, investors' attention is rapidly shifting beyond GPU (graphics processing unit) manufacturers to so-called 'AI Bottleneck Stocks' that are essential in the process of building AI infrastructure.

As AI models grow in scale and data centers become ultra-large, overall infrastructure—including data transmission, memory, power supply, and cooling facilities—is emerging as a critical factor that determines AI system performance, alongside computing power. The market defines AI bottleneck stocks as companies likely to sustain high demand due to supply shortages or technical constraints during the expansion of AI services.

With GPUs alone making it increasingly difficult to maximize data center performance, the value of core technology companies that support them is being highlighted as well. Below are the key sectors and representative companies drawing attention as AI bottleneck stocks.

■ Optical Communication — Key to AI Cluster Expansion

Optical transceivers, optical modules, and laser technologies are essential to connect tens of thousands of AI servers at ultra-high speeds. Representative AI optical communication beneficiary stocks include Coherent, Lumentum Holdings, Applied Optoelectronics, POET Technologies, Credo Technology Group, and Astera Labs. The industry expects the importance of ultra-high-speed optical interconnect technology to grow as AI clusters expand.

■ Network Equipment — Arteries of Data Transmission

How fast and reliably data is transmitted between GPUs determines the performance of AI systems. Consequently, interest is growing in the network solutions of Arista Networks, Broadcom, and NVIDIA.

■ Memory (HBM·DRAM) — Essentials for AI Accelerators

AI accelerators require large-capacity, high-speed memory as an essential component. Micron Technology, Samsung Electronics, and SK Hynix are evaluated as key beneficiary companies. The industry has also raised the possibility that a shortage in HBM supply may persist for the time being if the proliferation of AI servers continues.

■ Power Infrastructure — Key to Data Center Competitiveness

As power consumption in AI data centers surges, growth potential for power generation and electrical equipment companies is gaining attention. GE Vernova, Bloom Energy, Eaton, and Schneider Electric are among the leading companies in this space. Analysts note that a stable power supply capacity has become a determining factor in AI data center competitiveness.

■ Cooling Technology — Solution for High Heat Generation

The latest AI servers are seeing a rapid increase in demand for liquid cooling over traditional air cooling. Vertiv Holdings, Modine Manufacturing, and Carrier Global are cited as leading companies in this area.

■ Storage — Surging Demand for AI Data Storage

The storage sector is also drawing attention to store the exploding volume of data generated during AI training and inference processes. SSD and storage device companies such as SanDisk, Seagate Technology, and Western Digital are classified as stocks expected to benefit from the expansion of AI data centers.

■ Outlook

The market expects that the future competitiveness of the AI industry will likely be determined not simply by the ability to secure GPUs, but by the supply capabilities across the entire infrastructure, including optical communication, memory, power, and cooling. Accordingly, there are ongoing projections that AI bottleneck stocks will establish themselves as a central pillar of next-generation AI investment strategies.

[This article was generated with the assistance of AI. This article is not intended to solicit investment, and the final judgment and responsibility for investments lie with the investor.]