Three years ago, when artificial intelligence first captured our collective imagination, capital allocators and investors viewed it with the reverence of the Gods. Today, those same investors regard AI with the wariness of opening Pandora’s box, uncertain what manner of disruption might escape to infect the economy of margin-rich, knowledge-based businesses. This shift in perspective—from salvation to threat, from margin enhancer to margin destroyer—represents a reassessment of which forms of capital retain value.
The question was: How can artificial intelligence improve profit margins? The answers were: automate workflows, replace human labor, increase pricing power through data analysis.
Today, the question has flipped: How can artificial intelligence destroy margins?
There is a dawning recognition that technologies which automate knowledge for productivity may also threaten to commoditize products and lower barriers to entry for existing business models. Will AI reduce insurance brokerage to an algorithmic commodity? (AON insurance -14% in 7 days). Will wealth management advisors find themselves competing against robo-advisors charging basis points rather than percentage points? (Schwab -14% over the past 4 days). Will real estate brokerage relationships succumb to platform economics and artificial intelligence? (CBRE -28% over the past 3 days). The recent violent downward reactions in these formerly strong stocks represent not mere earnings adjustments but psychological shifts in the Wall Street Fashion Show.
Could knowledge move from asset to become a liability. Historically, knowledge-based businesses commanded premium valuations precisely because they converted asset-light intellectual capital into high returns on invested capital. Consulting firms, insurance brokers and wealth managers require minimal physical infrastructure, relying upon intelligent workers and confidence that their asset carrying customer walking out the door each evening will return the next day. These high margin, minimal capital intensity, scalable asset-light business models, once regarded as the most defensible form of capital, are increasingly viewed as vulnerable to AI focused substitution.
This growing sector rotation follows a typical pattern from exuberance over technological innovation with rapidly growing profits, followed by fears of overcapacity, cannibalization and the possibility that the technology’s disruptive power might be broader than initially imagined. If it can enhance advisory services, might it eliminate the need for human advisors entirely? When this fear arises, valuation compression arrives for businesses perceived as “disruption targets,” which eventually overshoots fundamental reality.
The leadership of the past three years has been notably narrow and capital-intensive among hyperscalers and their cousins. The mega-tech titans of cloud computing and artificial intelligence—committed hundreds of billions of dollars to AI data center capital expenditures with distant variable ROI’s. Now the market hunts for AI casualties – regardless of fundamental strength – with the enthusiasm of masses seeking revolution.
The Great Rotation: From Virtual to Physical
And rotation is precisely what current market action reflects. Small and mid-cap stocks, those perpetual laggards of the AI-driven rally, find themselves catching up. Equal-weighted indices firm. Industrial and transportation stocks outperform. Mega-cap concentration pauses at the verge of a breakdown.
Thus far, markets are not collapsing; they are redistributing capital: The more physical the output, the harder it is to disrupt. An artificial intelligence model, no matter how sophisticated, cannot:
- Transport freight
- Build a tractor
- Mine rare earth
- Pour concrete
- Harvest corn
- Repair an aircraft engine
At least not yet – unless one subscribes to a fraction of Elon Musk’s vision of 10 BILLION humanoid robots in 15 years. And the “not yet” matters considerably when one is attempting to forecast earnings two years forward.
Transportation and heavy industry require equipment, fuel, labor, regulation, and physical infrastructure. They operate within stubborn realities that remain indifferent to Silicon Valley’s ambitions. AI may improve margins and optimize logistics, but optimization is not substitution. A more efficient algorithm for routing trains does not eliminate the need for trains, tracks, or the diesel fuel that powers locomotives.
Mining, farming, manufacturing, railroad enterprises deal in tangible outputs subject to physical constraints. This helps explain why transportation stocks and industrial indices have found renewed sponsorship from investors wary of betting on businesses whose competitive advantages might evaporate as quickly as they materialized. (Transportation indices are up about 20 to 30% over the past 3 months while the AI kingpins -MAGS- are down 10%). Steel mills also appear defensible as best in breed Nucor steel has rallied 40% in recent months.
Capital Discipline Versus Hyperscaler Speculation
The hyperscaler AI buildout requires massive data center capital expenditures, multi-year payback periods, and uncertain monetization. By contrast, firms selling the proverbial picks and shovels of the AI gold rush monetize present demand rather than speculative future adoption. They benefit from the building of AI infrastructure without facing the risk of being replaced.
Taiwan Semiconductor Manufacturing Company fabricates the chips. ASML produces the lithography equipment necessary to make those chips. Lam Research supplies wafer processing equipment. Nvidia designs the GPUs that make AI training possible. Even Intercontinental Exchange, operating the infrastructure of financial markets, provides essential clearing and settlement services that AI can’t replace without regulatory approval that would require extremely complex legislative action. These enterprises monetize the present while the dreamers monetize the future. Future returns are assigned higher risk and lower valuation multiples.
A. AI-Exposed / Knowledge-Model Risk Cohort
| Company | Sector | Fwd P/E | FCF Yield | Narrative Risk |
|---|---|---|---|---|
| AppLovin | Ad Tech | 35–45x | Low | AI commoditizes targeting |
| Aon | Insurance Brokerage | 22–25x | ~3–4% | Automation compresses spreads |
| Raymond James | Wealth Management | 16–18x | ~5% | Fee compression risk |
| CBRE | Real Estate Brokerage | 14–17x | ~6% | Platform disintermediation |
These enterprises are asset-light business models, knowledge-based, margin-rich operations— now viewed as vulnerabilities to automation.
B. AI Infrastructure & Physical Monetizers
| Company | Sector | Fwd P/E | FCF Yield | Moat Character |
|---|---|---|---|---|
| TSM | Foundry | 20–25x | ~4% | Scale dominance |
| ASML | Lithography | 30–35x | ~3% | EUV monopoly |
| Lam Research | Wafer Equipment | 18–22x | ~5% | Installed base |
| Nvidia | AI Chips | 28–35x | ~3% | CUDA ecosystem |
| ICE | Exchange Infrastructure | 20–23x | ~4% | Regulatory network |
These firms sexcel due to scarcity (ASML’s extreme ultraviolet lithography monopoly), switching costs (Nvidia’s CUDA software ecosystem), and infrastructure economics subject to regulatory oversight (ICE’s exchange operations).
C. Tangible Asset Leaders
| Company | Sector | Fwd P/E | AI Risk |
|---|---|---|---|
| Caterpillar | Heavy Equipment | 16–18x | Low |
| Deere & Company | Agriculture | 20–23x | Low–Moderate |
| Union Pacific Railroad | Rail | 19–22x | Very Low |
| Delta Air Lines | Airlines | 10–12x | Very Low |
| McDonald’s | Consumer | 23–25x | Brand moat |
Tangible output and limited substitution risk protect these businesses operating in the physical realm where competitive advantages derive from scale, regulatory barriers, and brand equity built over decades rather than evolving algorithms.
Consumption Sector in Secular Uptrend
If artificial intelligence truly posed an existential threat to economic activity in the near term, one would expect the American consumer to telegraph distress through altered spending patterns. Yet the evidence suggests otherwise.
Bank of America reports approximately 5% year-over-year spending growth across all income cohorts. Higher-income households show growth around 2.5%, while even lower-income segments manage 0.3% increases—modest perhaps, but hardly suggestive of panic. Cassandra’s love to quote the very elevated credit card and auto loan delinquencies, but they deliberately obfuscate reality. Aggregate credit card delinquencies remain near 3%, while subprime 90-day delinquencies exceed 12%. This latter figure sounds alarming until one recalls that subprime borrowers represent only 14–20% of the total borrowing population. Systemic caution historically intensifies when overall delinquencies approach 4–5%; we remain well below such thresholds.
Travel and leisure spending remains robust. The American consumer continues to fly on Delta, queue at McDonald’s, and spend on Latin America and Nordic travel themes. Consumers also have a taste for culinary pleasures immune from AI substitution. An AI assistant may recommend dinner options but it cannot eat the meal, take the vacation, board the plane or experience the theme park. Physical experiences remain stubbornly resistant to digital substitution, a reality that investors are embracing as these stocks are becoming oversubscribed recently and due for a pause.
Tangible Capital Portfolio: 2026 “AI-Resilient Basket”
Core Infrastructure
- ASML
- Taiwan Semiconductor (TSM)
- Lam Research
- Nvidia
These firms provide essential inputs to AI development without facing replacement risk from AI itself.
Market Toll Collectors
- CME Group
Infrastructure providers operating under regulatory oversight, collecting fees – immune from AI.
Tangible Industrial Beneficiaries
- Caterpillar
- Deere & Company
- Union Pacific Railroad (IYT=transport ETF)
Heavy industry and transportation firms whose physical operations benefit from AI optimization without facing displacement risk.
Consumer Experience
- McDonald’s
- Delta Air Lines
Businesses providing physical experiences that resist AI substitution.
These eligible portfolio candidates emphasize monetization of present activity rather than speculation on future disruption, but lack potential benefit from AI adoption.
Afterburners, Not Apocalypse
The broader economic trajectory remains constructive. Earnings continue growing, profit margins persist at elevated levels, and wages outpace inflation. Capital rotates rather than retreats— suggesting reallocation rather than capitulation.
Artificial intelligence technology separates businesses with defensible competitive advantages from those whose profitability rested on information that algorithms can eliminate. This represents Schumpeter’s creative destruction, essential to capitalism’s dynamism, not apocalypse.
The early AI cycle was narrow and hyperscaler-dominated by the MAG 7 and its brethren. The present phase is broader, with sector rotation lifting participation across value-small-mid-caps, industrials, and tangible asset businesses that had been left behind during the initial wave of AI enthusiasm.
The transition from viewing AI as Promethean savior to fearing it as Pandora represents the maturing investor cycle understanding which business models benefit from technological change and which face displacement.
The stock market, taken as a whole, remains resolute in its optimistic posture, with valuations continuing to press higher on aggregate. Since the 2022 post-Covid bull market began, U.S. equities have expanded from roughly 140% of GDP—total economic output—to approximately 220% today. U.S. household net worth has climbed from an already robust $135 trillion in 2022 to nearly $185 trillion, underscoring the wealth effect supporting spending and investment alike, with figures such as Elon Musk even approaching an unprecedented trillion dollar milestone in personal net worth.
Our fourth-quarter outlook anticipated a short-term peak in mid-February, followed by a 5–10% corrective phase into quarter-end. On the S&P 500 Index, we continue to monitor key technical support levels in the 6,700s and just under 6,500 in the weeks ahead. With the former mega-cap leaders entrenched in a multi-month consolidation and several “picks-and-shovels” and rotation beneficiaries now extended into overbought territory, the case for a tactical pause remains intact.
An increase in cash allocations, paired with continued selective rotation into portfolio holdings and the sectors discussed above, appears prudent for this quarter. As the exaggerated software panic exhausts itself and the newly crowned winners cool from overextension, the broader market should be positioned to reassert a more balanced advance as the second quarter unfolds.