The 2026 Value Landscape
Seven deep-value case studies where every pick sits in a sector or industry the engine has historically beaten on. AI infrastructure, defense, and the energy and materials complex, surfaced by the same multi-method screen running across thirteen markets and over 8,000 backtested picks.
Finpy Team · June 22, 2026
Key takeaways
- 1.Industries the engine has historically delivered a 100 percent win rate on include Semiconductors, Semiconductor Equipment, Electrical Equipment, Oil and Gas Equipment, Gold, and Independent Power Producers. Every case study below is anchored in one of those industries.
- 2.Seven mega caps surfaced by the screen as of June 22, 2026. Combined market value above $6.5 trillion. Margins of safety range from 65 to 96 percent.
- 3.Three macro threads tie the picks together: AI infrastructure (NVDA, ASML, ETN, CEG), defense (TDG), energy and materials (SLB, WPM). The engine surfaced all three at the same moment.
- 4.Healthcare is the worst sector in our backtest (+6 percent, 57 percent win). None of the seven picks live there. Where the engine's history says do not fish, the article does not fish.
Cross-market backtest, Top 25 portfolio
Traditional mode, Balanced preset, equal weight. Buy date set seventy-five days after the prior fiscal year end to model the ten-K filing window.
| Market | Avg return | Win rate | Sample |
|---|---|---|---|
| Germany | +87% | 84% | 572 |
| United States | +54% | 76% | 795 |
| South Korea | +49% | 92% | 43 |
| United Kingdom | +28% | 68% | 76 |
| Japan | +19% | 60% | 224 |
Seven mega caps. Three macro threads. One engine running across thirteen markets.
Every case study below sits in an industry the engine has historically beaten on at a 100 percent win rate, or in a sector with above 78 percent. The picks are not random; the same filters surface the next wave of opportunities when this batch reprices.
Open the live screenerThree macro threads
The engine surfaced seven names at the same moment. Four are AI infrastructure plays at different layers of the stack. One is defense. Two are the energy and materials complex.
AI infrastructure
NVDA at the compute layer, ASML at the manufacturing equipment layer, ETN at the data center electrical layer, CEG at the baseload power layer. Hyperscaler capex has not inflected; consensus models still anchor on a normalized run rate the orders book is running well ahead of.
Defense
TDG as the aftermarket cash machine. Defense spending is structurally elevated; the aftermarket parts model is private-equity returns in public-company clothing. Aerospace and Defense sits in Industrials, the strongest sector in the entire backtest.
What "Finpy Calculated Intrinsic Value" means
Multi-method. The number is the average of three independent calculations on the same fiscal data: Greenwald Earnings Power Value (current operating earnings capitalized at the cost of capital, plus the value of growth capital earning above that cost), a fifteen-year forward profit projection, and a fifteen-year forward free cash flow to the firm projection. Tangible book value enters as a floor. The proprietary part is in the gate-and-weighting layer the engine applies on top.
Long-term. This is a fund-grade fair-value estimate over a multi-year horizon, not a twelve-month consensus price target. The engine has no opinion on when the gap closes. It says what the operating economics are worth at the current cost of capital. Patient capital is the implicit holding period.
Uncapped. The math output is never clamped. When the engine reads a 26x ratio on ASML or an 8x ratio on NVIDIA, the bar visual shows that without truncation. Capping would falsify the read; the reader can interpret extreme outputs against the quality gates the same number passed.
NVDANVIDIA Corporation
A 60 percent operating margin business sitting at the choke point of every hyperscaler AI capex budget.
NVIDIA sits at the AI compute layer with operating margins consensus has not credited at scale. Even at a punitive 17.9 percent WACC, the math reads $42 trillion of fair value against a $5.2 trillion market cap. CUDA is the software moat the commentary undersells; the install base does not port without years of enterprise rewrites.
Bull case
- Hyperscaler 2026 capex guides stepped up, not down
- Networking attach (InfiniBand, Ethernet) lifts revenue per system
- Sovereign AI builds (Saudi, UAE, EU) are incremental to hyperscalers
- CUDA moat measured in years of installed tooling, not quarters
Bear case
- Hyperscaler capex cycles eventually peak; that quarter is the inflection
- Custom silicon (Google TPU, Amazon Trainium, Meta MTIA) takes share at the edge
- China export controls remain a real overhang
Highest-conviction AI infrastructure pick. Operating margin and industry backtest both point the same direction.
ASMLASML Holding
A monopoly on the only machine that prints leading-edge logic, trading as if the cycle is ending instead of restarting.
EUV lithography is a single-source product line. ASML is the only company in the world that ships these machines. The 26x ratio looks extreme but it is supported by a 34.6 percent operating margin, a 52 percent return on equity, and a recurring service revenue base growing in the high teens attached to a multi-decade install base.
Bull case
- No competition on EUV; Nikon lost decades ago on cost-per-feature
- High-NA EUV machines in initial deployment at TSMC, Samsung, Intel
- Service revenue compounds on multi-decade install base
- China DUV demand growing despite restrictions
Bear case
- China export controls cut a permanent slice of TAM
- Semiconductor cycle inflects down before High-NA delivers
- Canon nanoimprint lithography is a slow-moving but real threat
The cleanest monopoly in the public market by the engine's read, with the 100 percent win backtest as the receipt.
ETNEaton Corporation
The electrical layer of the AI infrastructure stack, with a backlog the consensus is still catching up to.
Operating margin at 19 percent and rising. Return on equity at 21 percent. The data center vertical has been the standout growth driver for six straight quarters; backlog in Electrical Americas grew double digits year over year. Aerospace is steady; vehicle segments are the drag the consensus already knows about.
Bull case
- Data center backlog growing double digits Y/Y
- Industrial reshoring lifts the whole sector
- Acquisitions in data center cooling and power management
- eMobility business approaching profitability
Bear case
- Hyperscaler capex inflects down in 2026 or 2027
- Chinese industrial automation pricing competition
- Vehicle segment continues to drag operating income
The cleanest electrical infrastructure pick, anchored in the strongest sector in the backtest.
CEGConstellation Energy
The largest US nuclear fleet, suddenly the cheapest way for hyperscalers to buy 24/7 carbon-free baseload.
The structural shift came with the Microsoft Three Mile Island restart deal: long-duration PPA pricing nuclear baseload at a premium to grid average. Amazon followed with the Talen Energy deal. PJM capacity auction prices have spiked. The merchant baseload economics has been rewritten in eighteen months.
Bull case
- Microsoft and Amazon nuclear PPAs at premium pricing
- PJM capacity prices keep stepping up
- Largest US nuclear fleet, scarce asset
- Bipartisan policy support for nuclear permits
Bear case
- Hyperscaler power deals plateau and merchant prices revert
- Regulatory risk on existing nuclear permit renewals
- Small modular reactor competition (years out, but a real ceiling)
The cleanest way to own AI-driven baseload power demand at a value entry.
TDGTransDigm Group
Forty-seven percent operating margin on aftermarket aerospace parts. Private-equity returns in public-company clothing.
TransDigm acquires niche aerospace parts businesses with sole-source positions on specific aircraft platforms. Aftermarket revenue (roughly 55 percent of total) is the cash machine: parts are spec-locked to the platform, pricing power is structural, and the install base grows every year as the fleet expands.
Bull case
- Sole-source positions on thousands of platform-specific parts
- Aftermarket pricing discipline held through three administrations
- Defense budgets continue to expand globally
- Aftermarket fleet install base grows with airline expansion
Bear case
- Defense spending inflects down (geopolitical de-escalation, fiscal tightening)
- OEMs push back on aftermarket pricing
- Leveraged balance sheet; refinancing rates are a recurring overhang
A public-market version of the private equity playbook, in the strongest sector in the backtest.
WPMWheaton Precious Metals
A 68 percent operating margin gold royalty model. Not a miner. The defensive position the rest of the seven cannot give you.
Wheaton is not a mining company. It pre-pays mining operators for the future right to buy a percentage of their production at a fixed low price. That model produces a 68 percent operating margin and a 21 percent return on equity. Through the cycle, WPM outperforms the underlying miners because it carries none of the operating cost inflation.
Bull case
- Streaming model carries zero operating cost inflation
- Central bank gold buying continues quarter after quarter
- Higher silver mix than Franco-Nevada (cycle lead indicator)
- Selective deal pipeline, capital-light
Bear case
- Gold price reversion compresses stream economics
- Macro overlay flips (sovereign debt resolution)
- Stream contract renegotiation pressure from miners
The defensive sleeve of the seven, in a 100 percent win backtest industry.
SLBSLB (Schlumberger)
The largest oilfield services franchise in the world, sitting on an international upcycle the US-centric narrative keeps missing.
Roughly 80 percent of SLB revenue is international. The Middle East (Aramco, ADNOC), Offshore (Brazil, Norway, Guyana), and West Africa are the growth engines. US land has been the drag the headlines focus on; it is roughly 15 percent of total. Digital services growing faster and at higher margin than the consolidated business.
Bull case
- International capex cycle in motion, especially deepwater
- NOC budgets (Aramco, ADNOC, Petrobras) plan multi-year capex
- Digital services (Delfi platform) growing above company average
- ChampionX acquisition accretive to EPS run rate
Bear case
- Oil price reversion compresses E&P capex
- Energy transition narrative compresses long-term capex
- US land softness continues
The engine's best read on the international energy upcycle, anchored in a 100 percent win industry.
What links these seven
A trillion-dollar AI compute leader sits next to a nuclear utility and a gold royalty company. The thread is that the same Greenwald engine, run with the same Balanced preset on the same June 2026 fundamentals, surfaced all of them at the same time.
- ·Hyperscaler capex is the through-line for four picks. NVIDIA, ASML, Eaton, Constellation Energy - different layers of one capex cycle that consensus is modeling flat when the orders book is running up.
- ·Capital efficiency, not size, is what the engine selects. 47% op margin at TDG, 60% at NVDA, 68% at WPM all read the same way to the math.
- ·Backtest receipts sit next to every pick. Every industry is one the engine has historically delivered on.
Find similar names yourself
Open the live screener with the same Balanced preset that produced the case studies above. Switch the country dropdown to surface picks in Germany (where the strategy posts +87 percent), Japan, Korea, or any of the other markets in the universe.
Open the live screener with the same filtersThe same engine is accessible via the MCP integration for queries from Claude, an API client, or your own portfolio tooling. Documentation is on the Factor Markets product page.
How the engine works under the hood
Earnings Power Value treats a company as the present value of its current operating earnings at the current cost of capital, plus the value created by growth capex earning above that cost. A fifteen-year profit projection compounds net income from the latest fiscal year forward; the FCFF projection does the same for free cash flow to the firm. The three are averaged with tangible book value as a base floor.
Seven institutional quality gates confirm the math is reading a real operating business. Operating Margin and Operating Efficiency screen for profitability at the operating line. Net Cash and Price-to-Book are the Graham-style balance sheet filters. Revenue Growth and Return on Investment confirm capital is being deployed at attractive rates. Share Dilution and Accrual Quality close the most common ways earnings get inflated on paper. REITs and insurance are excluded at the model level because the math breaks structurally on those balance sheets.
What could go wrong
- ·Mispricing persistence. The market can hold a 5x undervaluation in place for years. The engine flags the gap; it does not say when it closes.
- ·Cost of capital assumption. A two-point WACC shift moves a fair value by 15 to 25 percent. The case studies use default assumptions; serious investors will re-run the math with their own.
- ·Single-snapshot timing. Numbers are as of June 22, 2026. The screener will produce a different slice when the next fiscal year close lands.
Built by a team that lives in the screener
Factor Markets is the platform behind this guide. Greenwald earnings power value, a fifteen year FCFF and profit projection, seven institutional quality gates, walk-forward backtest tables across thirteen markets. Built for fund-grade decisions, opened to anyone who wants to see how the math reads a real business.
Explore the platformAuthor positioning
No position in any of the seven companies at the time of writing. Factor Markets is the Finpy platform behind this analysis. The screener view shown in the examples is publicly accessible; the API and MCP integration are part of the platform tier.
Nothing on this page is investment advice. The numbers shown were pulled live from the Factor Markets production database on June 22, 2026 and will move with the market. Greenwald math is sensitive to the cost of capital assumption. Substituting a different equity beta, market risk premium, or risk-free rate changes the fair value materially. Do your own research before any trade.