Prediction: This Artificial Intelligence (AI) Stock That Has Jumped by Over 5x in a Year Can Still Soar by 66% or More

Applied Digital reported Q3 fiscal 2026 revenue of $126.6 million, up 139% year over year, and swung to adjusted EPS of $0.09 from a $0.01 loss, both well above Wall Street estimates. The company says it has $16 billion of contracted lease revenue over 15 years from CoreWeave and another hyperscaler, with 250 MW of capacity expected to come online over the next year and a 300 MW campus targeted for mid-2027. The article argues this AI data-center pipeline supports revenue more than tripling from 2026 to 2028 and could justify a premium valuation.

Analysis

APLD is increasingly a scarcity asset on the capacity side rather than a pure software multiple story. The market is underappreciating the embedded call option in long-dated lease contracts: once a campus is financed, built, and partially occupied, incremental occupancy can translate into very high marginal returns on invested capital, but only if power delivery and tenant schedules stay intact. That creates a non-linear earnings path over the next 12-24 months where reported revenue can inflect sharply ahead of cash flow, keeping the stock volatile but supported.

The second-order winner is CRWV, which is effectively using APLD as external balance-sheet capacity to meet exploding customer demand. That should help CRWV scale faster without tying up as much owned infrastructure, but it also raises concentration risk if one builder becomes the bottleneck for multiple contracted deployments. For NVDA, this is a demand-intensity signal rather than a direct revenue catalyst: every incremental AI campus reinforces the need for accelerated GPU refresh cycles, networking, and power-dense systems, but the near-term read-through is more about sentiment than immediate orders.

The key risk is execution slippage on power interconnects, construction timelines, or tenant ramp rates. Because the bull case is priced off a multi-year ramp, any 6-9 month delay can compress the multiple quickly; the stock behaves more like a project-finance equity than a traditional software compounder. A second risk is valuation re-rating: if lease revenue visibility remains strong but the market rotates away from high-duration infrastructure names, APLD can still grow rapidly while underperforming the index.

The consensus likely underestimates how dependent the entire thesis is on a small number of hyperscale counterparties and on continued access to low-cost capital. The move is not over if new campus commitments keep landing, but it is overdone if investors assume every announced MW converts into on-time, high-margin revenue. The trade is less about next quarter and more about whether APLD can keep converting contracted pipeline into energized capacity faster than financing and construction costs rise.

Key Decisions for Investors

• Long APLD on pullbacks over the next 2-6 weeks; target a 6-12 month hold. Upside is driven by MW-to-revenue conversion, but size modestly because delays in commissioning can cut the equity multiple quickly.
• Pair trade: long APLD / short a high-duration software name with weaker cash conversion over 3-6 months. The goal is to isolate infrastructure scarcity value versus pure narrative-driven growth exposure.
• Buy CRWV selectively on weakness as an operating leverage beneficiary of outsourced AI capacity buildout; 3-9 month horizon. Best risk/reward if investors expect another capacity-announcement cycle, but watch counterparty concentration.
• Use NVDA as a hedge rather than the primary long: add upside exposure via call spreads into any APLD/CRWV-led AI infrastructure rotation. This captures broader capex enthusiasm without overpaying for single-project execution risk.
• If APLD rallies toward a materially richer sales multiple before additional campus financing is secured, trim 25-30% and redeploy into names with nearer-term cash flow visibility.