“Scarce assets” in the AI era? Goldman Sachs: HALO—heavy assets, never outdated

```

Author: Long Yue

Source: Hard AI

As AI products become easier to copy, the market is beginning to revalue “hard-to-replicate physical assets” like power grids, pipelines, infrastructure, and long-term production capacity.

On February 24th, Goldman Sachs Global Investment Research released its latest report, “The HALO effect: Heavy Assets, Low Obsolescence in the AI era,” suggesting: Amid higher real interest rates, geopolitical fragmentation, supply chain restructuring, and a wave of AI capital expenditure, the core pricing logic of the stock market is shifting from “scalable light asset narrative” to “buildable, hard-to-replace physical capacity and networks.”

Goldman Sachs summarizes this change as a “re-pricing of scarcity.”

“Higher real yields, geopolitical fragmentation, and supply chain restructuring are pulling stock leadership back to tangible productive assets. The market is rewarding capacity, networks, infrastructure, and engineering complexity—assets whose replication costs are high and are less easily displaced by technology.”

What is HALO?

Goldman Sachs refers to these companies as HALO, which stands for the combination of “Heavy Assets” and “Low Obsolescence.”

• Heavy Assets: Business models are built on large-scale physical capital, with high barriers to replication—such as cost, regulation, construction time, engineering complexity, or network integration difficulty.
• Low Obsolescence: The economic relevance of these assets persists across technological cycles.

Typical examples include transmission networks, oil and gas pipelines, utilities, transportation infrastructure, key equipment, and various categories of industrial capacity whose replacement cycles are much slower than digital innovation.

These assets are hard to conjure out of thin air. In an era where digital technology changes rapidly, replacement cycles of such physical assets are extremely slow. Technological innovation cannot easily replace a transnational oil pipeline, nor substitute a vast national power grid with mere code.

Goldman Sachs observes enterprises decisively returning to physical assets. Capacity, infrastructure, and long-cycle assets are ushering in an unprecedented value resurgence.

Why is the “light asset” myth ending in the AI era?

For over a decade, post-crisis global zero interest rates and ample liquidity fostered business models centered around scalability rather than physical capital. Tech stocks and light asset sectors enjoyed high valuation premiums.

But this balance has been broken. The rapid rise of AI is applying a powerful “dual pressure” to global markets.

Firstly, AI is overturning the “new economy” model dominant for the past decade, making the “profit margins and terminal values” of some light asset sectors more uncertain. Goldman Sachs bluntly states: “The AI revolution is questioning the profit margins and terminal value of software and IT services.”

The report singles out software, IT services, publishing, gaming, logistics platforms, and even asset management, saying their moats are being reassessed. Goldman Sachs puts it directly: “Software and IT services have recently seen significant devaluation, not due to short-term profit collapse, but because the market is repricing the durability of terminal value and margins—historic high profitability is seen as more vulnerable to competitive erosion.”

In other words, AI lowers information processing costs and squeezes differentiation, so the market is more cautious about valuing future cash flows.

Secondly, AI is reshaping the landscape of capital expenditures. Goldman Sachs notes: “AI is turning some of the most iconic ‘light asset’ winners into the largest capital spenders in history.”

To stay ahead in the foundational AI model and computing power race, the five major US tech giants have begun an unprecedented investment cycle. Data shows that since the release of ChatGPT in 2022, these giants will spend up to $1.5 trillion in capital expenditures (Capex) from 2023 to 2026. By comparison, through their entire developmental history before 2022, their total investment was just about $600 billion.

Even more impressively, in just 2026 alone, their capital expenditure may exceed $650 billion—one year of spending outstripping the entire sum spent before the AI era. This is the largest and fastest capital expense cycle in tech history.

This means two things: First, “computing infrastructure” itself is a typical cycle of physical assets; second, AI does not make the world lighter, but instead enables more industries to benefit from capabilities that are “buildable, supplyable, deliverable.”

As tech giants become infrastructure junkies of “heavy assets,” the belief in “light asset” superiority naturally wavers.

The market is rewarding HALO with real money

Investors are sharp. The performance differences between Goldman Sachs' “heavy asset portfolio” (GSSTCAPI) and “light asset portfolio” (GSSTCAPL) provide the most intuitive market answer.

Data shows asset density has become the core driver of valuation and return. Goldman Sachs reveals: “Since 2025, our new heavy asset portfolio (GSSTCAPI) has outperformed the light asset portfolio (GSSTCAPL) by 35%.”

This outperformance is not just relative stock price fluctuations but a convergence of valuation logic.

In the early 2020s, as many old economy companies were seen as structural value traps, European growth stock valuations were more than double those of value stocks—with premiums as high as 150%. But now, the valuation gap between heavy and light assets has narrowed dramatically.

More noteworthy is the way the convergence has occurred. Goldman Sachs notes that their valuations are nearly at the same level, but this convergence is “mainly driven by heavy asset companies’ valuation re-rating, rather than a broad devaluation of light asset companies.”

Except for some software and other light asset sectors directly exposed to AI disruption risk turning weak, overall market evolution shows heavy asset companies have actively pulled up their valuations to match those of light asset peers. This shows market funds are paying premiums for the resilience and strategic value of physical economy assets.

How to define “heavy assets”? Six core criteria

To penetrate traditional industry classifications and precisely identify those truly reliant on physical capital, Goldman Sachs abandons single metrics and instead builds a “capital density score” system including six indicators. This system reflects a new perspective on asset quality in the market.

1. Tangible asset density (Net physical operating assets / Revenue): The higher the number, the heavier the physical base needed for every dollar of income.
2. Fixed asset density (Plant & equipment / Revenue): Reflects dependency on physical bricks and mortar.
3. Fixed asset share (Plant & equipment / Total assets): Shows how much money on the company’s balance sheet is “locked” in long-term physical assets.
4. Capital-to-labor ratio (Tangible assets / Employee count): Distinguishes whether business is machine-driven or people-driven.
5. Capex intensity (Capex / Revenue): Measures the annual “bloodlets” needed to maintain or expand business.
6. Capex burden (Capex / EBITDA): Shows how much operating cash profit is consumed by asset maintenance.

Using these six criteria, Goldman Sachs divided companies into distinctly different camps.

Utilities, basic resources, energy, and telecom are unambiguously in the heavy asset camp. These sectors are strictly regulated, have very high fixed capital requirements, and extremely long asset lifespans.

Conversely, software, IT services, internet, and media platform companies are solidly in the light asset, labor-intensive category.

Interestingly, there’s a “middle zone” in the market. Goldman Sachs finds that autos and aviation are obviously heavy assets; but due to brand assets, production know-how, and long-term investment in processes, luxury and beverage also fall into the “low obsolescence” high-quality asset category. By comparison, consumer services, gambling, and most retailers are structural light assets, their economic lifeblood relies on labor and marketing rather than physical capital.

Macro tailwinds and earnings momentum resonance

Why are heavy assets booming at this moment? The answer lies in a dual resonance of macroeconomic indicators and corporate fundamentals.

On the interest rate side, heavy asset stocks tend to do well during high interest rate periods. High yields relentlessly squeeze the valuations of long-duration, light asset growth companies. Conversely, heavy asset sectors tied to tangible capacity can benefit from stronger nominal economic activity and government fiscal spending. Goldman Sachs notes that today’s policy mix is steering capital toward physical assets, “creating structural tailwinds for capital-intensive companies.”

On the macro cycle side, the contest between manufacturing and services is key. Heavy asset sectors’ fate is closely linked to industrial production and the capex cycle. Goldman Sachs observes that as manufacturing PMI—especially the business expectations component—rebounds and overtakes services PMI, the macro backdrop is tilting back toward heavy asset sectors.

As for earnings, which determine long-term market returns, the fundamentals have shifted.

Over the past cycle, light asset companies enjoyed sustained high-growth profits and long-standing valuation premiums. But from 2025 onward, although heavy asset companies’ short-term earnings have been buffeted by tariffs and trade friction (since they're commodity producers and exporters more affected than service firms), the trend is already clear once you strip out short-term noise.

Goldman Sachs stresses: “Earnings momentum for heavy asset companies has recently turned positive, consensus expectations are rising; meanwhile, light asset companies' earnings forecasts are being downgraded.”

Looking ahead, analysts expect the heavy asset portfolio’s consensus EPS CAGR to reach 14% over the next few years, vs. just 10% for the light asset portfolio. Even more crucially, the core indicator supporting light assets’ high valuation—Return on Equity (ROE)—is showing weakness. The market now expects light asset companies’ ROE to stay steady, while heavy assets’ ROE is likely to improve.

Fund crowding: The rotation to heavy assets is just beginning

Given the logic is so clear and valuations have converged, is this heavy asset rally over?

From the perspective of fund flow, it’s far from over.

The recent heavy asset leadership is closely tied to the market’s urgent desire to escape crowded, expensive “US tech stock” positions. In the past 12 months, European value funds have seen net inflows of 3%, while growth funds have suffered 9% net outflows.

But Goldman Sachs points out bluntly, despite sharp short-term rotation, long-term fund positions are still very thin: “Net cumulative outflows of European value funds relative to growth funds still linger at around -40% of asset management scale.”

This means global investors remain severely underallocated to value stocks (the locus of heavy assets). With this huge position gap, the structural rationale for continued outperformance of heavy asset stocks over light asset stocks remains rock solid.

In this AI-accelerated era of reconstruction, the virtual world’s rapid charge ironically makes the steel, pipelines, and power grids of the physical world uniquely precious. Whether this is a lasting market leadership replacement or cyclical rebalancing, for investors, the “bulletproof vest” quality of physical capital is shining brightly.

This article is from WeChat public account “Hard AI”. For more AI frontier news, click here.

Risk Disclosure and DisclaimerThe market has risks; investment requires caution. This article does not constitute personal investment advice, nor does it take into consideration individual users’ special investment goals, financial circumstances, or needs. Users should consider whether any opinions, views, or conclusions in this article fit their specific situation. Investments made accordingly are at your own risk. ```