Applied Digital Just Hit 1GW Of Contracted Capacity

Applied Digital secured a 15-year, $7.5B take-or-pay lease at Polaris Forge 3, reinforcing its shift to a scaled, contracted AI infrastructure platform. The company now has over 1 GW of total contracted IT load, with repeat hyperscaler demand and 65% of revenue tied to investment-grade clients. Valuation is elevated at 32x forward EV/sales and 136x EV/EBITDA, but the large backlog supports expectations for strong future revenue and margin expansion.

Analysis

This is less about a single lease and more about APLD crossing a credibility threshold: once a platform has enough contracted load to look quasi-utility-like, the market starts underwriting financing optionality, not just project execution. That can compress the cost of capital and create a flywheel where additional hyperscaler demand is easier to win because counterparties prefer scale, power access, and delivery certainty over point solutions. The second-order winner set likely includes power, electrical, and buildout vendors tied to liquid-cooled AI campuses, while smaller colocators and speculative data-center developers face a tougher fundraising environment as customers increasingly anchor to de-risked, contracted capacity.

The key risk is that the equity is now pricing in a multi-year conversion story with very little room for slippage. At these valuation multiples, even modest delays in energization, interconnection, or capex inflation can cut the implied terminal value far more than the headline lease adds today. The near-term catalyst path is mostly operational: financing terms, incremental tenant signings, and whether management can keep turning backlog into cash flow without margin dilution over the next 2-6 quarters.

The contrarian view is that the market may be over-earning confidence from backlog size while underweighting concentration risk and the financing burden of growth. Take-or-pay contracts reduce demand risk, but they do not eliminate construction, power, permitting, or counterparty execution risk; if any one of those steps slows, the revenue back half of the model slides while fixed costs keep accumulating. The stock likely remains momentum-supported in the short run, but the cleanest upside is now from proof of delivery, not announcement alpha.

Key Decisions for Investors

• Stay long APLD only tactically on pullbacks, not as a core add here; use a 2-4 week window and require confirmation on financing/energization milestones before adding size. Risk/reward is still positive, but asymmetrically worse after the re-rating.
• Buy APLD downside hedges via put spreads 3-6 months out to protect against construction or capital-markets disappointment. The payoff improves if the market keeps assigning premium multiples before the first major execution miss.
• Pair trade: long beneficiaries of AI infrastructure buildout with less execution risk, short a basket of smaller speculative data-center names. The thesis is that capital will continue concentrating into the most bankable platform models over the next 6-12 months.
• For event-driven accounts, trade APLD only on post-announcement consolidation: add on 10-15% drawdowns after initial excitement fades, since the next catalyst set is likely slower-moving and operational rather than headline-driven.
• Monitor if the multiple compresses while backlog converts; if EV/sales stays elevated without visible free-cash-flow progress over 2 quarters, fade the name with a short against a higher-quality infra peer or via listed options.