May 12, 2026
The AI investment narrative continues to broaden beyond the dominant U.S. semiconductor names. While many investors remain heavily concentrated in domestic AI exposure, the next phase of the trade may increasingly involve the global infrastructure ecosystem supporting artificial intelligence, particularly memory, data centers, and international semiconductor supply chains.
During a recent television appearance, Michael Landsberg, Chief Investment Officer of Landsberg Bennett Private Wealth Management in Punta Gorda, Florida, discussed why diversification within the AI ecosystem matters, especially after the massive run seen across many U.S.-based AI-related equities.
Rather than focusing solely on U.S. names that have already appreciated significantly, Landsberg pointed toward international exposure and memory-related infrastructure as a way to broaden participation in the AI buildout.
One of the key themes raised was that memory remains a critical input for AI workloads. Large language models, cloud infrastructure, inference systems, and data center expansion all require substantial memory capacity and storage throughput.
That trend has helped fuel the strong performance of memory-related businesses and ETFs tied to the semiconductor ecosystem.
Landsberg emphasized that many portfolios remain overly concentrated in the U.S. AI narrative while overlooking the broader global infrastructure chain supporting AI demand. He noted that only about 37% of global data centers are located in the United States, reinforcing the idea that AI growth is increasingly international.
One of the more notable insights from the conversation was not necessarily bullishness on AI itself, but concern around concentration risk.
Landsberg explained that even when holdings perform remarkably well, positions are routinely trimmed and rebalanced to prevent risk from becoming excessive inside portfolios. That discipline becomes particularly important after rapid appreciation.
The broader point is relevant for today’s market environment. Many investors now hold substantial exposure to a narrow group of AI-linked equities without realizing how concentrated their portfolios have become.
The discussion also highlighted an important distinction between “staying with winners” and allowing position sizing to become uncontrolled. Long-term conviction does not eliminate the need for risk management.
Global semiconductor demand continues to be supported by AI infrastructure expansion. According to the Semiconductor Industry Association, worldwide semiconductor sales reached record levels in 2025, with AI-related demand contributing heavily to memory and accelerator growth.
Meanwhile, key cloud providers continue expanding AI data center infrastructure globally, not just within the United States. This reinforces Landsberg’s point that AI exposure increasingly extends beyond domestic equities alone.
The conversation also touched on IPO markets and transparency concerns surrounding private companies. As more large firms remain private longer, retail investors may gain access to private markets without receiving the same level of disclosure traditionally available in public markets.
Landsberg argued that if retail participation in private assets continues to expand, reporting standards and transparency should improve alongside it.
The AI trade is gradually evolving from a narrow “chip winner” story into a broader infrastructure and ecosystem story.
That shift matters because leadership may eventually rotate away from the heavily crowded names toward adjacent beneficiaries such as memory, networking, data center infrastructure, exchanges, cybersecurity, enterprise software, and international suppliers.
The other important takeaway is behavioral. Investors often confuse strong performance with reduced risk. In reality, rapidly appreciating assets can increase portfolio risk through concentration. The discipline of trimming, rebalancing, and diversifying internationally may become increasingly important as AI-related valuations continue climbing.
Investing involves risk, including the possible loss of principal, and no investment process or strategy is guaranteed to be profitable. Past performance does not guarantee future results, and any investments mentioned may not be suitable for all investors. Hypothetical examples are for illustrative purposes only, do not represent actual results, and actual results and individual experiences may differ.
This material is for general informational and educational purposes only. It is not individualized investment, tax, legal or insurance advice, or a recommendation to buy or sell any security or insurance product. Consult qualified professionals regarding your individual circumstances. Views expressed are solely those of the [author(s)/speaker(s)] and do not necessarily reflect the views of Hightower Advisors, LLC or any of its affiliates.
All information referenced is current as of the publication date and may change without notice. Although obtained from sources believed to be reliable, Hightower Advisors, LLC, and their respective affiliates do not guarantee its accuracy or completeness. References or links to third-party content do not constitute an endorsement of any third party or its products, services, or views. Click here for definitions of and disclosures specific to commonly used terms.
Landsberg Bennett Private Wealth Management is a group comprised of investment professionals associated, and/or registered, with Hightower Advisors, LLC, an SEC registered investment advisor. Some investment professionals may also be registered with Hightower Securities, LLC, member FINRA and SIPC. Advisory services are offered through Hightower Advisors, LLC. Commissionable securities are offered through Hightower Securities, LLC. Hightower Advisors and Hightower Securities are affiliated entities.