Is Applied Digital Stock a Buy?

Applied Digital has secured $36 billion of AI data-center lease agreements over 15 years, with renewal options that could raise potential contract value to $86 billion over 30 years, but its shares have fallen about 50% from their May high of $50.73. Debt surged to roughly $5 billion as of May 31 from less than $700 million a year earlier, while fiscal Q4 revenue was only $258.7 million despite more than 400% year-over-year growth. Construction delays before rent commencement, memory-chip shortages, elevated energy costs, and a gradual ramp toward projected fiscal 2028 revenue of $1.45 billion leave the stock suited mainly to high-risk, long-term investors.

Analysis

APLD’s equity is effectively a long-duration, highly levered construction project rather than a mature AI-infrastructure cash-flow compounder. The central valuation variable is not signed backlog but the conversion rate of committed capacity into operational, rent-paying MW; delays simultaneously defer EBITDA, increase interest during construction, and can force incremental equity issuance. A high nominal contract value therefore does not protect shareholders if financing costs or capex per MW rise faster than lease economics.

The less obvious exposure is CHRN: APLD shareholders retain substantial look-through exposure to a more volatile GPU-cloud operator while also bearing the parent’s development leverage. That structure can create a temporary valuation opportunity only if the market assigns little value to APLD’s CHRN stake, but it also means any GPU-utilization or pricing downturn can impair both the tenant/compute layer and the landlord’s perceived asset value. MSFT is strategically important as a demand-validation signal, but it should not be treated as a material earnings lever absent disclosed committed capacity, credit support, and delivery milestones.

Over the next 1-3 months, financing announcements, construction commissioning dates, power-delivery confirmation, and debt-covenant disclosure matter more than AI sentiment. Over 6-18 months, the key risk is that abundant neocloud capacity turns take-or-pay contracts into weaker renewal economics once initial terms expire; power-constrained, fully energized sites will retain scarcity value, while delayed campuses may not. The consensus may be underestimating refinancing and dilution risk, but also over-discounting the shares if operational delivery begins to validate a repeatable project-finance model.

Key Decisions for Investors

• Keep APLD on a financing-and-commissioning watchlist rather than initiate a core long. Upgrade only after management discloses delivered MW, tenant rent commencement, and funding for the next build phase without a material equity raise; those data would reduce the main left-tail risk.
• For a tactical long, prefer a defined-risk APLD call spread dated 6-12 months out only after a confirmed facility acceptance or non-recourse financing event. Target at least 2:1 payoff; invalidate on a revised delivery schedule, higher capex per MW, or another unsecured-debt-funded expansion.
• Avoid owning both APLD and CHRN as separate full-size positions: this is concentrated exposure to the same AI-capacity cycle with different liquidity profiles. If CHRN’s market capitalization implies a discount to APLD’s reported ownership value after adjusting for holding-company costs, consider a small long CHRN / short APLD pair, subject to borrow availability and updated ownership data.
• Use MSFT or NVDA as cleaner AI exposure if the objective is secular demand rather than construction-finance beta. APLD should be sized as a special situation until interest expense, contracted revenue commencement, and free-cash-flow timing can be independently modeled.