The repricing of "old world" capital-intensive assets is spreading from data centers to the broader supply chain of the real economy. A recent Goldman Sachs research report highlighted that in the AI era, capital is flowing toward "Heavy Asset, Low Obsolescence" (HALO) physical assets, such as power grids, pipelines, utilities, transportation infrastructure, and critical industrial capacity. These assets are difficult to replicate, possess high physical barriers, and are less prone to obsolescence. Since late February, a basket of HALO-related stocks has continued to rise.
Concurrently, shifts in U.S. tariff policy are providing additional support for transactions involving physical assets. Goldman Sachs preliminarily estimates that adjustments triggered by a Supreme Court ruling could lower the effective U.S. tariff rate by approximately 100 basis points. Although the direction of tariff policy remains uncertain, any cost relief for infrastructure and industrial firms reliant on physical imports would strengthen the short-term rationale for a market shift toward tangible assets.
Goldman Sachs strategist Chris Hussey emphasized that investors are assigning higher valuations to traditional capital-intensive companies, with this trend extending both upstream and downstream in the supply chain and into the broader economy. The firm's analyst Oppenheimer explained that this marks the first time in roughly a quarter-century since the commercialization of the internet that technological growth prospects are highly dependent on physical assets like data centers and energy supply. The revaluation of physical assets, from data centers to the entire infrastructure chain, is underpinned by the logic that during a gold rush, the most stable beneficiaries are often the sellers of shovels.
Oppenheimer noted that technological growth in the AI era is increasingly reliant on visible, tangible physical assets, rather than the pure software and platform models that dominated markets over the past twenty-five years. This reassessment has led to a strategic repricing of "hard assets." Goldman defines the highest-quality HALO targets as companies possessing two key traits: first, assets that are costly to rebuild, face deep regulatory barriers, and have long construction cycles, making them hard to disrupt or replace; and second, assets that hold long-term economic value. Specific focus areas include power grids, pipelines, utilities, transportation infrastructure, key machinery and equipment, and long-cycle industrial capacity.
While physical assets gain favor, soft assets are facing continued pressure. Sectors such as software, media, consulting, and certain financial subsectors are undergoing market reassessment. In February, software and IT services stocks experienced significant declines again, with stocks like INTU, WDAY, IBM, and ACN each falling more than 20% for the month. There are concerns that AI could "disintermediate" these businesses or open the door for lower-cost competitors, potentially undermining their core business models.
However, Goldman Sachs analyst Gabriela Borges stressed that not all software companies have identical business models. She pointed out that the agentic technology ecosystem is evolving rapidly, making it challenging to assess final value and establish valuation floors. Therefore, it does not imply that all software stocks should be sold off. Goldman suggests that investors can still focus on companies that can demonstrate their historical experience leads to higher-quality AI outcomes and maintain or improve their fundamentals through profitability in the coming AI era.
Comments