I practice asset allocation every month to accumulate retirement pension assets by using Stirling (STRL) and Comfort systems (FIX). Recently, the stock price growth rates of semiconductor sector ETFs , SOXX SMH, and AI data center-related ETFs , AIPO IVEP CTAI, have been rising. Given that the AI data center industry is predicted to grow over the next five years, I plan to invest a portion of my retirement portfolio in ETFs that hold AI data center infrastructure companies. Let’s take a look at how to select AI infrastructure-related ETFs and determine the proportions for constructing the portfolio.
Apple’s stock price surged significantly following the announcement of plans to introduce various AI models alongside news of its own chip development. AMD is also showing a sharp rise of over 6%, buoyed by an earnings report showing a 35% increase in revenue.
Next year, the so-called M7 companies are scheduled to execute massive capital expenditure (CAPEX) plans exceeding $1 trillion. Consequently, as memory demand is expected to skyrocket, the stock prices of related companies are also showing an upward trend.
Following the recent issuance of $30 billion in corporate bonds, Google issued a 37-year Euro-denominated bond today. This is interpreted as a strong determination to seize dominance in the AI market, even if it means overinvesting, rather than losing the initiative due to investment failures.
The U.S. national debt has surpassed its GDP and is recording its highest level since World War II. As the phenomenon of debt generating interest accelerates, the yield on U.S. 3-year Treasury bonds is already approaching the 5% mark. As Big Tech companies increase their corporate bond issuance, a competitive landscape is forming with government bonds, which are considered safe assets.
If demand swells toward Big Tech corporate bonds that offer competitive interest rates, government bonds are forced to raise their rates to attract investors.
From the Treasury’s perspective, they must lower oil prices to reduce inflationary pressure and defend long-term government bond interest rates.
Considering the overall macroeconomic environment, the semiconductor sector still possesses strong momentum for further gains. For investors managing pension accounts, it is crucial to adhere to principles rather than being swayed by market noise.
A strategy of consistently accumulating semiconductor ETFs such as SMH, SOXX, and MAGS according to a fixed ratio is effective.
A year ago, the advent of the AI era led to a significant increase in the need for GPUs. The demand for NVIDIA GPUs surged, causing NVIDIA’s stock market capitalization to become the world’s largest.
With the release of the H100 Blockwell Vera Rubin, NVIDIA has begun to meet the demand for GPUs. As the demand for memory, which was previously used in GPUs, shifts to the Agent era, the speed and computational power of inference have become critical.
With the emergence of stacked memory known as HBM, inference speeds are gradually accelerating. It is projected that by around 2030, technology will evolve to a level where Agents communicate with each other and pay costs via tokens.
Regarding bottlenecks in AI data centers, there is a surge in demand for securing sites and construction, as well as a rapid increase in demand for generators, transformers, substations, and high-voltage cables to supply power to these centers.
As a basic principle for pension assets, it is recommended to allocate 1/4 each to VTI (a portfolio of all US stocks with low volatility and an upward trend), VXUS (which purchases global stocks outside the US), BND (which pools US Treasury bonds and high-quality corporate bonds with maturities of less than 3 years), and Gold IAUM.
Based on this structure, it would be advisable to partially invest in AIPO IVEP TCAI, an AI power infrastructure ETF responsible for the bottleneck areas in AI data center construction, and to include stocks of companies involved in land and site construction, as well as the electrical and mechanical facilities within AI data centers.
The goal is to accumulate 1% of total investment assets in individual stocks. Let’s pay attention to the tremendous growth rates of AI data center construction companies.
The stock price of Sterling Infrastructure (STRL) was $23.18 in May 2021.
By May 2026, it had risen to $806, representing a growth rate of 3,377%. That is a whopping 34-fold increase.
Sterling Infrastructure (STRL) is a specialized construction company that purchases land for AI data centers, excavates the site to create foundations, and installs electrical and mechanical facilities inside the buildings. It is currently working on constructing AI data centers for Amazon and Meta.
Earnings themselves have skyrocketed. Revenue for the first quarter of 2026 was projected at $592 million, but rose to $826 million. Earnings per share were expected to be $2.19 but recorded $3.59. Naturally, the stock price rose 52% in a single day. Not only were earnings strong, but guidance for 2026 was also significantly raised.
Revenue guidance was presented at $3.8 billion, up from $3.7 billion, and the adjusted EPS was previously $13.59. The upward EPS guidance increased from $18.40 to $19.05. It appears that the market has clearly reached a consensus that Sterling Infrastructure (STRL) will see steady growth in both revenue and net income.
Comfort Systems (FIX) stock price was $86.87 in May 2021, but rose to $1,967 on May 5, 2026.
This represents a 2,614% increase. In just five years, the stock has risen a whopping 26-fold. Comfort Systems specializes in the installation and maintenance of cooling and HVAC systems for AI data centers.
Revenue itself grew significantly, with revenue for 2025 reaching $9.1 billion. This marks a 29.5% increase in revenue compared to 2024.
Net earnings (EPS) stood at $28.88, representing a 97.8% growth year-over-year. The order backlog for 2026 stands at $11.94 billion, already more than double that of 2025.
As the demand for cooling equipment in AI data centers continues to increase in the future, Comfort Systems (FIX)’s stock price is expected to trend even further upward.
STRL and FIX are good choices for upgrade my Asset Allocation Portfolio. My basic portfolio are consisted of following ETSs as middle aged man.
VTI : 20%
VXUS: 20%
QQQ: 10%
SGOV: 10%
BND : 20%
IAUM : 20%
My ideas is changed from the stability is the best to enhancing 1%~5% of my portfolio is better than keeping stable portfolio. After 1 year for 5 years later, I will realized which portfolio is good to me. Even if U.S. stock and international stock market is broken like 2000 Dotcom bubble, I want to sleep well in night. I will tell you more ideas about new AI related stable ETSs in next post!
#leofinance #inleo #invest #investing #stock #finance #financing #hive #hive-kr
Originally published on bomspring.com on May 9, 2026
