Tensions in the Strait of Hormuz have ignited the most notable capital movement in the spring of 2026. On March 9, sectors including oil and gas, coal, and electrical equipment and grids led the gains, with the electrical grid sector alone attracting over 9 billion yuan in net main fund inflows. Over the past week, the oil and gas sector surged more than 10%, energy equipment rose over 9%, and shipping, power, and coal were among the top performers. Market consensus is forming: hard assets like power grids, pipelines, and minerals—which cannot be disrupted by AI—are becoming safe havens in an era of uncertainty. This rally, sparked by geopolitical conflict, has made the unfamiliar acronym "HALO" go viral in investment circles, standing for "Heavy Assets, Low Obsolescence." Goldman Sachs refers to it as "scarcity repricing": under the triple pressures of rising interest rates, geopolitical fragmentation, and surging AI capital expenditure, investors are no longer chasing "scalable light-asset narratives" but are instead embracing physical capacity that is "buildable and hard to replace." Behind this shift lies both fear of AI and a craving for tangible assets: on one hand, AI's "excessive success" has made light assets vulnerable; on the other, the "information disruption" caused by conflict has made hard assets scarce. These two forces intertwine, forming the dual logic behind the rise of the "HALO" strategy. But the real question is: Is this merely a short-term避险 driven by geopolitical conflict, or the dawn of a new era? Once the smoke clears, can the long-term logic of HALO assets support their valuation reassessment?
What is "HALO"? The term "HALO" is an acronym for Heavy Assets, Low Obsolescence. The concept was introduced by Josh Brown of Ritholtz Wealth Management and later featured in research reports from Goldman Sachs and Morgan Stanley, quickly gaining popularity in investment banking circles. Goldman Sachs explained in its report that under the triple pressures of rising interest rates, geopolitical fragmentation, and surging AI capital expenditure, the market is undergoing a "scarcity repricing." Investors are no longer blindly pursuing the "scalable light-asset narrative" that was highly regarded over the past few decades, but are instead turning to "buildable, hard-to-replace physical capacity and networks." Morgan Stanley's constructed "HALO" basket covers seven structural pillars: materials (minerals, non-ferrous metals), utilities (power grids, electricity), railways and pipelines, waste management, defense, and infrastructure like signal towers. The common characteristics of these assets are: extremely high replication costs, natural physical moats, and immunity to obsolescence from revolutionary AI iterations. A straightforward explanation is: You can use ChatGPT to write an analysis report on oil pipelines, but you cannot use code to magically build an oil pipeline in the Texas desert. With the escalation of U.S.-Iran conflict, this strategy has demonstrated a remarkable ability to attract capital. Market reaction has been direct and forceful. On March 2, the first trading day after the escalation of U.S. and Israeli military actions against Iran, A-shares in the oil and gas sector saw a wave of limit-up gains, with the "big three" oil companies罕有地集体封板 for several consecutive days. Looking at Wind China industry indices, over the past week, oil and gas surged over 10%, energy equipment rose over 9%, and shipping, power, and coal were among the top performers. As of March 9, the "HALO" strategy remained effective. Oil and gas and coal still led gains for the day, both rising over 3%. In terms of fund flows, the electrical equipment and grid sector attracted significant main fund buying, with net inflows nearing 8 billion yuan.
Why now? To understand this collective capital migration, one must look back at the recent panic moment dubbed the "bloody February" for the software industry. Initially, after Anthropic launched its legal AI tool, U.S. stocks of legal software and data service companies fell sharply. Subsequently, the sell-off spread to the entire software sector, including semiconductors and AI infrastructure. Then, after online insurance platform Insurify launched a new AI tool, the S&P 500 Insurance Index closed down nearly 4% that day. When wealth management firm Altruist introduced an AI tax planning tool, the entire wealth management sector fell collectively. By late February, a research report simulating the potential chain-reaction economic crisis triggered by AI further intensified market panic. Suddenly, investors stopped asking, "How are this company's fundamentals?" and started asking, "Will this company be wiped out by AI?" Behind the simultaneous selling of semiconductors and AI computing power and the rush into oil pipelines, grid towers, copper mines, and railways lies a profound shift in valuation logic. The爆发 of "HALO" investing is not accidental; "避险 demand" under AI disruption panic is a core reason. Guolian Fund analysis stated that with the rapid proliferation of AI technology, markets widely fear that light-asset, software, service, and labor-intensive companies could be replaced by AI, casting doubt on their profit sustainability. HALO assets, as "AI-immune" assets—physical facilities like power grids, pipelines, and railways that cannot be replaced by code—have become "safe harbors" for capital seeking to avoid technological disruption risk. Simultaneously, the escalation of U.S.-Iran conflict, like a stone thrown into a pond, shattered surface-level market expectations for growth industries. The Strait of Hormuz handles about 20% of global seaborne crude oil trade, and its shipping stagnation directly boosts the "geopolitical premium" of energy assets. Wang Yanbo, a digital economy scholar at the Shanghai Academy of Social Sciences, pointed out that geopolitical events like the U.S.-Iran conflict essentially represent a sudden interruption or distortion of the global collaborative "information flow," amplifying the uncertainty of physical supply chains (flow of matter and energy) and短期内推高ing the premium on "hard assets" that secure these flows. The "dimensionality reduction attack" from the digital world, combined with "flow interruption panic" from the physical world, acts simultaneously, forming the dual logic behind the rise of the "HALO" strategy.
Is "HALO" overpricing AI's "destructive power"? How long will "HALO" remain hot? The market inevitably shows divergence. Optimists believe this is not short-term避险 but a long-term "new continent." The notion that "the end point of AI is energy" reveals the rigid dependence of computing power explosion on physical energy. The development of new quality productive forces does not weaken but rather strengthens the strategic position of traditional energy. The exponential energy demand from data centers and AI computing centers makes power and oil & gas—'old energy'—the physical foundation of the digital economy, transforming their value logic from cyclical commodities to growth infrastructure. This integration means new quality productive forces must be built upon solid 'old assets.'" Jiang Han, senior researcher at Pangoal Institution, noted that heavy assets, due to their long construction cycles and high approval barriers, form natural supply constraints, giving them a monopolistic premium akin to "digital era oil" within the AI wave, significantly enhancing their strategic value. This reaffirms a fact: the ultimate destination of AI is electricity; regardless of its final form, it cannot escape the "weight" of the physical world. However, a pointed质疑 arises: Is the current capital rush into HALO sectors overpricing AI's "destructive power" while underestimating humanity's ability to overcome bottlenecks through innovation? Wang Yanbo believes the "HALO" strategy is more like a "stage-specific answer" within a particular technological cycle, not a long-term "new continent." The current surge in AI computing power demand has indeed hit expansion cycle bottlenecks in physical capacity like power grids, copper mines, and chip manufacturing (the physical foundation required for information processing is temporarily scarce). However, the history of technological progress shows that humans always overcome physical bottlenecks through information innovation. He cited examples like more efficient chip architectures (e.g., in-memory computing), advanced materials science (e.g., conductive materials replacing copper), or better energy utilization methods (e.g., nuclear fusion) to reduce the material and energy consumption per unit of computing power. He further indicated that the focus of development is gradually shifting toward large-scale, implementable application layers. Only by transforming advanced computing power and algorithms into actual products and services that meet real-world scenario demands can sustainable commercial value and profits be created, completing the cycle from technology to industry. "This is a prominent advantage of the Chinese market and enterprises: possessing a vast user base, rich application ecosystem, agile iteration capabilities, and deep understanding of localized needs." Wang Yanbo believes that once the application端 achieves scaled profitability, it will inject strong momentum into the entire industry chain. Historical experience shows that every technological revolution undergoes a process from panic to excessive flight, followed by value reassessment. The internet did not "kill" Walmart; instead, Walmart learned e-commerce. Ultimately, what may truly be淘汰ed are enterprises that lack inherent barriers and survive solely on information asymmetry. If AI's evolution stops within the digital world, continuing to serve as a "cost-cutting and efficiency-boosting" tool like Software-as-a-Service, then the reassessment of "HALO" assets might just be a周期轮动 driven by inflation expectations. But if AI, as the industry expects, progresses from language models to world models, from digital intelligence to embodied intelligence, from "thinkers" to "actors," then the value reassessment of the physical world may have only just begun. In this sense, HALO and AI are not opposites. Those heavy assets possessing power grids, pipeline networks, transportation networks, and mineral resources are becoming the "toll gates" of the AI era. "This year's barbell strategy involves tech stocks on one end and HALO assets on the other," Yang Delong of Qianhai开源 Fund pointed out. This bears some similarity to last year's barbell strategy of tech plus high-dividend sectors, but the logic has slightly changed. One end embraces the certainty of technological change through computing infrastructure and semiconductor equipment; the other end hedges against the dual risks of technological iteration and geopolitical turmoil by deploying heavy assets like energy, non-ferrous metals, and grid equipment. The market is never black and white. Truly smart capital never limits itself to simple choices.
Comments