Applied Digital CEO: AI Industry Will See “Pretty Significant Delays Through 2026 and 2027”

Applied Digital reported fiscal Q3 2026 revenue of $126.64 million, up 139.29% year over year and 61.37% above consensus, while adjusted EPS of $0.09 beat the -$0.21 estimate and adjusted EBITDA rose to $44.14 million from $6.26 million. The company also signed a new 210 MW Delta Forge 2 lease, taking contracted revenue to roughly $36 billion across five campuses, and priced $1.59 billion of 7.000% senior secured notes due 2031 to fund a 150 MW expansion. Management argues labor shortages and modular buildouts are supporting execution, though concentration risk remains elevated with about 70% of contracted revenue tied to CoreWeave or another single hyperscaler.

Analysis

The market is still underestimating how much AI infrastructure is becoming a labor-arbitrage story, not just a power story. If onsite skilled labor stays tight, the real moat shifts to operators that can standardize, prefab, and replicate buildouts faster than the market can price in; that should compress the relative advantage of conventional EPC-dependent peers and favor names with repeatable deployment playbooks. Second-order beneficiaries extend beyond the obvious hyperscaler landlords to industrial modular fabrication, electrical gear, and logistics firms that can scale offsite assembly without exposing projects to the same wage inflation and schedule slippage.

The key takeaway is that execution optionality has value when capacity is scarce and pre-leased. When contracted revenue is concentrated, schedule reliability becomes the main determinant of whether leverage is a feature or a bug: a few quarter delays can defer cash conversion, but on-time delivery can force faster renewals and better pricing economics into 2027. That makes the next 2-4 quarters more about proof of construction discipline than about absolute demand, and it also raises the bar for competitors trying to announce similar megawatt pipelines without an operating track record.

The contrarian view is that the stock may be pricing in too much perfection on both build speed and customer concentration. If financing markets tighten or one hyperscaler pauses expansion, the downside is not just lower growth but a much lower multiple on a highly levered, single-tenant cash flow stream. The bigger reversal catalyst is not demand destruction; it is a modest increase in project slippage combined with rising rates, which would hit both the duration of the equity story and the cost of capital that underpins future expansion.

Key Decisions for Investors

• Long APLD on any 8-12% pullback over the next 2-6 weeks, but size as a tactical growth/execuation trade, not a core compounder; target a 15-25% rebound if the next project milestone lands on time.
• Buy APLD call spreads 3-6 months out to express upside from de-risking of build schedules while capping premium if the market re-rates the name too aggressively on financing or concentration concerns.
• Pair trade: long APLD / short a more execution-fragile AI infrastructure peer with a heavier EPC dependence over the next 1-2 quarters; thesis is that prefab, repeatable delivery should outperform in a labor-constrained cycle.
• Monitor senior secured debt spreads and secondary equity supply in APLD over the next 30-60 days; widening spreads or an equity raise would be the earliest warning that the market is starting to discount schedule risk.
• If hyperscaler capex headlines keep rising but APLD fails to confirm on-time commissioning over the next 1-2 earnings prints, reduce exposure quickly: the stock’s upside is tied to evidence, while downside from execution misses can be abrupt.