Biggest gaps to analyst fair value
A live scan of a curated universe, ranked by how far each trades below the analysts' average price target. Real data — a research starting point, not advice.
"Upside" is the analysts' average price target versus the current price, pulled live from public data. It reflects a curated watch-universe, not the entire market, and a high target gap is not a recommendation — always do your own research.
Undervalued, with room to run
Names screening below fair value or with outsized potential — and a few 2–3 year plans built around the one debate dominating markets: is AI a bubble?
Heading into 2026, an AI-driven valuation crash is the single most-cited market risk — roughly half to a majority of surveyed fund managers now call AI the market's biggest tail risk, with the mega-cap leaders seen as the most exposed. At the same time, value and small-cap names sit well below fair value. The plans below lean into that split: stay exposed to the real AI build-out, but with ballast underneath if the froth comes off. It's a debate, not a forecast — timing a bubble is notoriously hard.
The watchlist
A wide-moat consumer brand deep in a turnaround, trading at a fraction of its peak multiple. Margin recovery and inventory normalization are the swing factors — and it's barely correlated to the AI trade.
Screened ~50%+ below some fair-value estimates in 2026
Turnarounds take longer than hoped; weak consumer and China exposure can drag results for several quarters.
Medical-office and lab REIT — the most undervalued sector by Morningstar's read. Defensive cash flows and a fat yield make it ballast if growth multiples compress.
~40% below fair value, ~7% dividend yield (Morningstar, 2026)
Rate-sensitive; office/real-estate sentiment is poor and organic growth has been slow.
A boring, cash-generative telecom on a high yield — the kind of name that holds value when the market sells the froth and rotates to safety.
High single-digit yield; long-time value pick
Heavy debt, fierce price competition, and little growth; total return leans on the dividend.
A leveraged bet on the mortgage cycle — if rates ease, refinancing and origination volumes can snap back hard off a low base.
Below Morningstar's ~$17 fair value after a weak 2026 start
Highly rate-dependent and cyclical; a 'higher for longer' rate path keeps the core business under pressure.
AI datacenters need enormous, reliable power, and gas is the bridge fuel. As the biggest US gas producer, it's a physical-world beneficiary of the AI boom that doesn't depend on AI software multiples holding up.
Largest US natural-gas producer; powering datacenter demand
Commodity prices are volatile; a warm winter or weak LNG demand can gut earnings.
Builds the efficiency and grid projects utilities need as power demand surges. A 'picks and shovels' angle on electrification and the datacenter power crunch.
Energy-efficiency & grid-infrastructure play
Lumpy project revenue, balance-sheet leverage, and policy/interest-rate sensitivity make it a bumpy ride.
Precious-metals miners tend to catch a bid when equities wobble and macro uncertainty rises. A small ballast sleeve here can offset a growth-heavy book if the AI trade unwinds.
Silver/gold miner — leverage to metal prices
Miners are high-beta on the metal price; operational misses and costs can erase the hedge.
A gold miner with production growth — geared to a strong gold backdrop and a hedge against an equity drawdown or a weaker dollar.
Gold producer ramping output
Execution and cost risk at its mines; gold-price swings dominate the story.
The networking backbone of AI datacenters, with genuine earnings (unlike many AI names). The plan isn't to chase it — it's to have a price in mind and accumulate if an AI pullback drags it down with the rest.
Fair-value estimates raised on datacenter growth, 2026
Still an AI-cycle name: a datacenter capex slowdown or overbuild would hit it hard.
The dominant launch + Starlink franchise, finally public. A multi-year story on launch cadence and satellite connectivity — but priced for a lot of future success.
Newly public (June 2026); currently lossmaking
Rich valuation, negative earnings, thin trading history, and lock-up dynamics make early volatility likely.
2–3 year plans
The Barbell
2–3 yearsPair a few quality, profitable AI-infrastructure names with a sleeve of cheap defensive value. You stay exposed to the AI build-out without betting the whole book on lofty multiples holding.
If the AI trade corrects (analysts model 20–50% drawdowns in the mega-cap leaders), the value sleeve cushions you — and gives you dry powder to add to the AI names after they overshoot to the downside.
Hyperscaler capex guidance, datacenter utilization, and whether AI revenue actually scales with the spend.
If the bubble keeps inflating, the defensive sleeve lags and you underperform a melt-up.
Powering the AI Build-Out
2–3 yearsOwn the physical inputs AI can't run without — electricity, gas, grid and efficiency. Datacenter capex is projected past $500B in 2026, and all of it needs power.
Even if AI software valuations deflate, the power demand is already contracted and physical. This is the part of the AI story with the longest, least-hype-dependent runway.
Power-purchase agreements with hyperscalers, natural-gas demand, and grid-interconnection bottlenecks.
Commodity-price swings and project lumpiness; a genuine AI capex pause would eventually hit demand too.
Hedge the Froth
2–3 yearsA deliberate ballast sleeve — precious-metals miners and cheap value — sized to offset a growth-heavy portfolio if sentiment turns.
With ~45–57% of surveyed managers calling AI the market's biggest tail risk, a non-correlated hedge is cheap insurance. Gold and silver historically firm up during equity drawdowns and macro stress.
Real interest rates, the dollar, and equity-market breadth (narrow leadership is a warning sign).
If markets stay calm, the hedge is a drag; miners are volatile and not a clean hedge day-to-day.
Quality on Sale
2–3 yearsWide-moat compounders trading below fair value for company-specific reasons — not because of the AI cycle. Lower correlation to a tech-led drawdown.
These names can keep working even if AI multiples compress, because their stories (brand turnaround, mortgage cycle, pricing power) are their own.
Margin recovery, free-cash-flow trends, and whether the discount to fair value is closing.
'Cheap for a reason' is real — turnarounds can stay broken longer than you can stay patient.
Educational research only — not investment advice, and not personalized to your situation. Every name here carries real risk, valuations and fair-value estimates move daily, and the AI-bubble scenario is one view among many. Do your own work before acting on anything.