Recently, while tracking investment opportunities in global internet infrastructure, I found that the market's understanding of 5G concept stocks remains stuck in old thinking. The true driver of 5G demand is not smartphone upgrades, but the data traffic explosion brought by AI.

Senior executives at Qualcomm recently predicted that future traffic demand could be three to seven times current levels, with over 30% driven by AI. This is not nonsense; there are very specific structural changes happening behind the scenes.

Let's look at the most critical shift first. The proportion of AI inference computing power has risen from one-third in 2023 to half in 2025, and this year, it has surpassed two-thirds for the first time. What does this mean? Training is concentrated and cyclical, but inference is continuous and dispersed, accumulating as each new application is added. To reduce latency, inference must be deployed at the edge, no longer stacked in a few large data centers. This directly increases the demand for network infrastructure.

Next, consider the growth of AI agents. It is estimated that by 2026, the global number of AI agents will be about 50 to 100 billion, potentially skyrocketing to 2 to 5 trillion by 2036. Along with this growth, global bandwidth usage will surge from about 100 exabytes per day to 8,100 exabytes. The compound annual growth rate of AI-driven network traffic is as high as 51%, and overall traffic volume is expected to grow five to nine times.

What are the actual investment opportunities behind these numbers? I break down the industry chain of 5G concept stocks into four levels.

The upstream is optical communications and optical modules. Internal data center interconnects have shifted from traditional bundles of fewer than 1,000 fibers to thousands of fibers. North American hyperscale data centers see annual traffic growth exceeding 30%, and high-speed specifications over 800G, such as optical transceivers, have become bottlenecks. Companies like Corning are directly benefiting from this upgrade demand. Silicon photonics and low-power optical modules (LPO) are also core growth drivers for 2026, with Taiwanese manufacturers having a complete ecosystem in this area.

The midstream is network equipment providers. Companies like Ericsson and Nokia are transforming from traditional telecom equipment vendors into enablers of enterprise edge AI and private 5G. Ericsson recently signed multi-year strategic partnerships with NTT DATA to promote private 5G in manufacturing, mining, ports, and other scenarios. These orders' visibility is directly reflected in revenue, with the largest fluctuations.

The downstream is telecom operators. Traditional telecom stocks like AT&T and Verizon have long been considered defensive dividend stocks, but after the surge in AI traffic, fixed broadband, fiber access, and 5G fixed wireless access all have renewed growth potential. Valuations have been discounted long-term, and as the market reassesses growth sustainability, corrections may occur.

Taiwan's supply chain also has clear opportunities. The upgrade to 800G optical modules and the acceleration of private 5G directly boost MediaTek's 5G-Advanced chips, Wistron and Macroblock's power amplifiers, ZTE's switches, and Lianya and Huaxing's optical communication modules.

If you prefer not to bet on a single stock, ETFs like FIVG and NXTG, which cover equipment, chips, infrastructure, and operators, can effectively diversify risk.

But be aware of three risks. First, telecom operators have yet to find effective business models to turn traffic into profits; there is a time lag between investment and returns. Second, in some regions, government approvals, land permits, power supply, and tariff pressures may delay deployment. Tariffs have already increased hardware costs for chips, RF modules, and antennas, leading Asian supply chain vendors to face longer procurement cycles. Third, the 6G narrative has already emerged early, with some funds shifting from 5G equipment to 6G concept stocks, which could pressure the growth-phase 5G equipment stocks.

The current core logic is that network bottlenecks have shifted from download bandwidth to uplink bandwidth, low latency, and reliability. AI inference demand has already surpassed training, with 24/7 online AI agents generating continuous, bursty, and highly elastic mixed traffic. 5G's inherent advantages in ultra-low latency, high reliability, and massive connectivity align naturally with AI agent needs, so equipment vendors and optical communication supply chains directly benefit from the structural growth in AI data center capital expenditure. In the short term, deployment delays, 6G funding, or telecom profits falling short of expectations may cause volatility, so it's wise to leave some flexibility. If you're interested in deepening your understanding of these 5G concept stocks, the market trends and related assets on Gate are worth paying attention to.