How "Heavy Assets, Low Obsolescence" became the hottest investment thesis of 2026
Executive Summary
- Goldman Sachs reports capital-intensive stocks have outperformed capital-light peers by 35% since 2025, as the market reprices scarcity in the AI era — the most dramatic sector rotation since the dot-com bust.
- The HALO trade — Heavy Assets, Low Obsolescence — has emerged as Wall Street's dominant defensive strategy, funneling capital into energy, industrials, utilities, and infrastructure while software and IT services face existential valuation compression.
- The convergence of AI disruption fears, geopolitical fragmentation, and a Middle East war is accelerating the rotation into physical assets, creating a potential multi-year secular shift in market leadership that challenges a decade of asset-light supremacy.
Chapter 1: The Birth of a New Acronym
In early February 2026, financial commentator Josh Brown of Ritholtz Wealth Management coined an acronym that would come to define the year's most consequential investment shift: HALO — Heavy Assets, Low Obsolescence. Within weeks, Goldman Sachs had formalized the concept in a landmark research note, Morgan Stanley's trading desk had built entire hedging frameworks around it, and the term had migrated from financial Twitter into the lexicon of every institutional allocator on the planet.
The thesis is deceptively simple: in an era when artificial intelligence is systematically destroying the profit margins of software companies, IT service providers, and knowledge-work intermediaries, the safest place to park capital is in businesses whose value derives from physical things that cannot be replicated by an algorithm — pipelines, power grids, oil refineries, heavy machinery, restaurant chains, farmland.
But HALO is not merely a defensive crouch. It represents the market's collective judgment that the economic structure of the AI era will look radically different from what Silicon Valley promised. Rather than making the economy "lighter," AI is paradoxically making it heavier — driving the largest capital expenditure cycle in corporate history as tech giants build out the physical infrastructure required to run their models.
The numbers are staggering. Goldman Sachs calculates that the five major U.S. tech companies will have spent $1.5 trillion on capital expenditures between 2023 and 2026 — more than 2.5 times their cumulative investment in the entire pre-ChatGPT era. In 2026 alone, their combined capex is expected to exceed $650 billion, roughly equivalent to the GDP of Switzerland. This is not the behavior of an industry that believes the future is asset-light.
Chapter 2: The Great Rotation in Numbers
The scale of the sector rotation is without modern precedent. While the S&P 500 index has remained broadly range-bound since late October 2025, the internal dispersion tells a completely different story.
Morgan Stanley trader Kunal Sodha documented what he called the most extreme internal rotation in recent market history:
| Metric | Change (Q4 2025 – Feb 2026) |
|---|---|
| Growth vs. Value return spread (MSZZGRVL) | -24% drawdown |
| Beta long/short factor (MSZZBETA) | -21% pullback |
| S&P 500 Info Tech + Consumer Discretionary | -11% combined |
| Industrials | +13% |
| Consumer Staples | +16% |
| Materials | +22% |
| Energy | +25% |
The return spread between the best and worst-performing S&P 500 sectors has reached its third-highest level in five years. Goldman Sachs' proprietary basket of capital-intensive stocks (GSSTCAPI) has outperformed its capital-light basket (GSSTCAPL) by approximately 35% since the start of 2025.
To put this in perspective: during the dot-com bust of 2000–2002, the rotation from tech into value stocks took roughly 30 months to achieve a similar magnitude of divergence. The HALO rotation has accomplished this in approximately 14 months.
The geographic dimension is equally striking. The FTSE 100, heavily weighted toward "old economy" companies in energy, mining, and consumer staples, posted its strongest February since November 2022 — its eighth consecutive monthly gain. The pan-European Stoxx 600 hit record highs. Meanwhile, the Nasdaq Composite has been the worst-performing major index globally in 2026.
Top HALO performers in the Stoxx 600 (YTD 2026):
- Frontline (oil tanker shipping, Cyprus): +57%
- Kongsberg Gruppen (defense/energy systems, Norway): +46%
- Rheinmetall (defense, Germany): +38%
- Shell (energy, UK): +29%
- Caterpillar (heavy equipment, US): +24%
Chapter 3: Why AI Is Making the Economy Heavier, Not Lighter
Goldman Sachs' February 24 report, titled "The HALO Effect: Heavy Assets, Low Obsolescence in the AI Era," articulated a profound irony at the heart of the current moment: the very technology that was supposed to make physical assets obsolete is instead creating the largest demand surge for physical infrastructure in decades.
The report identified four converging forces driving the repricing:
1. AI Capital Expenditure as Physical Asset Demand
Every dollar spent on AI ultimately materializes as physical demand — for semiconductor fabrication plants, data center buildings, copper wiring, power generation equipment, cooling systems, and the energy to run it all. The International Energy Agency estimates that data centers will consume 9–17% of U.S. electricity by 2030, up from roughly 4% in 2023. This is not a virtual phenomenon. It requires turbines, transformers, transmission lines, and fuel.
2. Geopolitical Fragmentation and Supply Chain Reshoring
The parallel fragmentation of global supply chains — driven by U.S.-China decoupling, the SCOTUS IEEPA ruling, and now the Iran conflict — is forcing companies to build redundant physical capacity in multiple jurisdictions. Goldman noted that "after more than a decade of under-investment, particularly in Europe, corporates are shifting decisively back toward physical assets."
3. The SaaSpocalypse Effect
The AI-driven destruction of software company margins — what the market has dubbed the "SaaSpocalypse" — has not merely rotated capital away from tech. It has fundamentally challenged the market's belief that asset-light models deserve premium valuations. When Anthropic's Claude Cowork plugin can replicate functionality that took enterprise software companies decades to build, the terminal value of those businesses approaches zero. Goldman stated bluntly: "The AI revolution is calling into question the profit margins and terminal values of software and IT services."
4. The Hormuz Premium
The outbreak of military conflict between the U.S.-Israel coalition and Iran in late February 2026, and the subsequent IRGC declaration of Hormuz Strait as a "war zone," has injected an acute scarcity premium into every physical commodity on Earth. Energy infrastructure, shipping capacity, and strategic commodity reserves have become not merely profitable assets but matters of national survival. The HALO thesis, originally conceived as an AI hedge, has been turbocharged by geopolitical reality.
Chapter 4: The Three Pillars of HALO
Morgan Stanley's framework categorizes HALO assets into three distinct clusters, each with its own investment logic:
Pillar 1: Industrials and Manufacturing
Companies like Deere & Co., Caterpillar, and Siemens Energy derive their competitive advantage from physical production lines, proprietary equipment, and mature supply chain networks. These businesses have technology refresh cycles measured in decades, not quarters. Their production processes involve tacit knowledge embedded in physical systems that cannot be digitized or replicated through AI.
Key characteristic: High barriers to entry based on engineering complexity, regulatory approval, and time-to-build. A competitor cannot "prompt" a new semiconductor fabrication plant into existence.
Pillar 2: Energy and Commodities
ExxonMobil, Shell, BHP, and Glencore sit atop physical resource reserves whose value is determined by geological scarcity, not software innovation. Their refining facilities, logistics networks, and extraction expertise represent decades of accumulated physical capital. The Iran conflict has reminded markets that energy infrastructure is among the most strategically critical — and vulnerable — asset class on Earth.
Key characteristic: Inelastic demand combined with inherent physical scarcity. AI cannot synthesize crude oil or mine copper.
Pillar 3: Consumer Chains and Offline Networks
McDonald's, Walmart, and Coca-Cola possess global networks of physical locations, distribution systems, and brand infrastructure that AI cannot displace. McDonald's owns or controls approximately 40,000 restaurant locations globally — a physical footprint that generates stable rental income and brand premium regardless of AI developments.
Key characteristic: "You still need this on Monday morning" businesses, as Saxo's Ruben Dalfovo puts it. Offline consumption scenarios have low sensitivity to technological disruption.
Chapter 5: Historical Precedents — When Markets Rediscovered the Physical
The HALO trade is not without historical parallels. Each major technological disruption has eventually produced a counter-rotation into physical assets:
The Dot-Com Bust (2000–2002)
The collapse of internet stocks redirected capital into energy, materials, and industrials — sectors that had been neglected during the late-1990s tech mania. This rotation persisted for nearly seven years, with the Bloomberg Commodity Index rising 160% between 2001 and 2008. Energy stocks outperformed the Nasdaq by over 200 percentage points during this period.
Similarity to today: Extreme valuation compression in technology, rotation into "real" assets.
Difference: The current AI disruption is arguably more fundamental than the internet boom because AI threatens the profit margins of incumbents, not just startup valuations.
The 1970s Stagflation (1973–1982)
The oil shock and inflationary spiral of the 1970s created a decade-long regime in which physical commodities dramatically outperformed financial assets. Gold rose 2,300%, oil rose 1,100%, and farmland values doubled. The S&P 500, adjusted for inflation, lost approximately 60% of its value over the decade.
Similarity to today: Energy supply shock (Hormuz), inflation persistence (PCE 3.0%), central bank credibility crisis.
Difference: The 1970s rotation was purely macro-driven. Today's HALO trade has an additional structural driver — AI-induced obsolescence risk.
The Post-GFC Commodity Supercycle (2009–2011)
China's massive stimulus program after the 2008 financial crisis drove a supercycle in physical commodities, benefiting mining, energy, and infrastructure companies. BHP's share price tripled between 2009 and 2011.
Similarity to today: Massive fiscal spending programs (EU SAFE bonds, U.S. defense buildup, global rearmament).
Difference: Today's physical asset demand is driven by both AI infrastructure AND geopolitical rearmament — a dual-engine catalyst.
Chapter 6: Scenario Analysis — How Long Does HALO Last?
Scenario A: Secular Regime Change (40%)
Thesis: The HALO trade marks the beginning of a multi-year, possibly decade-long shift in market leadership from asset-light to asset-heavy companies.
Supporting evidence:
- The 2000–2007 commodity supercycle lasted seven years after the dot-com bust
- Structural underinvestment in physical infrastructure since 2010 suggests years of catch-up capex ahead
- Geopolitical fragmentation (supply chain reshoring, rearmament) is a multi-year force
- AI capex cycle is projected to persist through at least 2030
Trigger conditions: Continued AI disruption of white-collar employment; sustained Hormuz premium; no rapid resolution of U.S.-China decoupling.
Investment implications: Energy, defense, mining, and utilities become the new "growth" stocks with premium valuations. FTSE 100 and European industrials structurally outperform Nasdaq.
Scenario B: Extended Rotation with Reversion (35%)
Thesis: The HALO trade persists for 18–36 months before AI companies find sustainable business models and tech valuations stabilize.
Supporting evidence:
- Citadel Securities' rebuttal of the Citrini "doom memo" argues AI displacement follows an S-curve, not exponential acceleration
- Historical tech disruptions (electrification, internet) eventually created more jobs than they destroyed
- AI capex spending will eventually translate into measurable productivity gains
Trigger conditions: AI productivity data shows measurable economic benefits; Iran conflict resolved; software companies successfully integrate AI into defensible platforms.
Investment implications: Tactical overweight in HALO assets for 12–24 months, with gradual rebalancing toward AI platform winners. Barbell strategy: physical infrastructure + AI platform oligarchs.
Scenario C: HALO Overshot (25%)
Thesis: The rotation has already priced in most of the structural shift, and HALO valuations become stretched.
Supporting evidence:
- Goldman notes capital-intensive firms in Europe are now MORE highly rated than capital-light firms on P/E basis — a historic first
- Energy sector up 25% YTD despite global recession risks
- War premium in energy prices may prove temporary if Iran conflict de-escalates
Trigger conditions: Ceasefire in Persian Gulf; AI productivity breakthrough validates tech capex; central bank easing cycle revives growth stocks.
Investment implications: Reduce HALO overweight; selectively re-enter quality software names with defensible moats (Salesforce Agentforce, ServiceNow).
Chapter 7: Investment Implications — Positioning for the Physical
Winners in the HALO regime:
| Sector | Rationale | Key Names |
|---|---|---|
| Energy Infrastructure | Hormuz premium + AI power demand | Exxon, Shell, Chevron, NextEra |
| Defense & Aerospace | Global rearmament supercycle | Rheinmetall, BAE, Hanwha, L3Harris |
| Mining & Metals | Copper/lithium scarcity + AI infra | BHP, Freeport-McMoRan, Albemarle |
| Utilities | Regulated returns + data center load | Duke Energy, Southern Co, Iberdrola |
| Heavy Industrials | Reshoring + infrastructure rebuild | Caterpillar, Deere, Siemens Energy |
Losers in the HALO regime:
| Sector | Rationale | Key Names |
|---|---|---|
| Enterprise Software | AI functional replacement | SAP, Salesforce (legacy), ServiceNow |
| IT Services | Billing model disruption | Accenture, Infosys, TCS |
| Digital Advertising | AI-generated content flood | Meta (ad revenue), Alphabet |
| Fintech/Payment | Regulatory + AI disruption | Block, PayPal, Adyen |
The most important metric has changed. For a decade, investors rewarded companies that could scale revenue without proportional capital investment. The new regime rewards companies that have already built physical assets that would take competitors years and billions of dollars to replicate. As Goldman puts it: "Asset intensity has become a key driver of valuations and returns."
Conclusion: The Revenge of the Physical World
The HALO trade represents something more profound than a typical sector rotation. It is a philosophical reckoning with the limits of the digital economy. For fifteen years, the market's highest valuations were reserved for companies that could grow without touching the physical world — software platforms, digital marketplaces, cloud services. The AI revolution was supposed to be the ultimate extension of this logic: pure intelligence, weightless and infinitely scalable.
Instead, AI has revealed the opposite truth. The companies building AI need more physical infrastructure than any industry in history. The companies threatened by AI are precisely those whose products can be replicated digitally. And the companies most insulated from AI are those whose value is literally embedded in the earth, in steel, in concrete, in the irreducible physical reality of grids, pipes, and machines.
The Hormuz crisis has amplified this lesson with brutal clarity. When missiles fly over oil shipping lanes, no amount of software engineering can conjure barrels of crude from server racks. When data centers consume 17% of the national grid, the owners of that grid hold the real leverage.
Wall Street has a new holy word. It is not "disruption" or "scale" or "platform." It is HALO — the halo that glows around assets you can touch, build, and defend. In the AI era, the physical world is having its revenge.
Sources: Goldman Sachs Global Investment Research, "The HALO Effect" (Feb 24, 2026); Morgan Stanley Trading Desk (Kunal Sodha); Motley Fool; The Guardian; TradingKey; Saxo Market Commentary
Leave a Reply