Applied Digital Corporation (APLD): AI Data Center Infrastructure Pioneer at an Inflection Point

Applied Digital Corporation designs, builds, and operates next-generation digital infrastructure, primarily for the AI and high-performance computing industries. The company operates through three main business segments: Cloud Services (offering GPU compute capacity), Data Center Hosting (providing colocation and powered shell space for large-scale AI workloads), and a legacy Crypto Mining segment that is being wound down or repurposed.

The company generates revenue through long-term lease agreements and service contracts. Its primary customers are hyperscale cloud providers, AI labs, and large enterprises seeking reliable, cost-efficient compute capacity. Applied Digital differentiates itself by securing sites with access to low-cost, often renewable or stranded power, enabling significant operational cost advantages. The company's current operational and development footprint spans North Dakota (Ellendale), Texas, and other strategic locations across the United States, with a total potential capacity pipeline exceeding 1.5+ gigawatts. With approximately 284 million shares outstanding and a market capitalization of $8.1 billion, the company is a significant but early-stage player in the AI infrastructure buildout.

• Near-Term (0-12 Months): The primary growth driver is the commencement of revenue recognition from the Ellendale campus. The first phase of the CoreWeave contract is expected to become operational in the current calendar year, driving a substantial sequential increase in quarterly revenue. Management has guided that the second half of the fiscal year will see the majority of this revenue inflection. Additionally, the company's GPU cloud services segment is expected to see continued demand as enterprise clients seek to reserve scarce H100/H200 capacity.
• Medium-Term (1-3 Years): Growth will be driven by the continued buildout of Phases 2 and 3 of the Ellendale campus and potential expansions at other sites. The company is actively marketing its powered-shell solutions to other hyperscalers and sovereign AI funds, which could lead to additional contract announcements. The medium-term trajectory points toward a company scaling from tens of millions in revenue to potentially $1 billion+ in annualized revenue run-rate, driven by the contracted pipeline.
• Long-Term (3+ Years): The long-term opportunity is tied to the sustained secular growth of AI compute demand. As power becomes the primary constraint on AI expansion, Applied Digital's portfolio of power-secured sites becomes increasingly valuable. The company's potential to sell or develop additional campuses, coupled with the growth of its own cloud platform, positions it to be a top-tier independent infrastructure provider.

The financial trajectory illustrates a company in a heavy investment phase. Historical losses reflect the development of its data center portfolio and the wind-down of crypto operations. The dramatic improvement in gross margins projected for FY2026 and FY2027 is predicated on the operationalization of the Ellendale campus, which carries high contracted rates with minimal incremental power costs. The company is expected to require significant capital raises or debt financing to complete its construction pipeline, which will likely lead to shareholder dilution in the near term. However, the projected EBITDA inflection in FY2026 is the critical metric that the market is underwriting, transitioning the company from a development-stage entity to an operating infrastructure REIT-like business.

The AI data center market is experiencing explosive growth, with the global market size projected to exceed $100 billion by 2028, growing at a ~20%+ CAGR. The primary constraint is not demand but the availability of power and the speed of construction. This environment favors companies with pre-secured power and construction expertise.

Applied Digital competes in a bifurcated landscape:

• Hyperscalers (In-House): Microsoft, Google, Amazon, and Meta are building massive in-house capacity but still rely on external providers for overflow demand and specialized needs.
• Specialized Data Center REITs and Operators: Companies like Digital Realty (DLR) and Equinix (EQIX) focus on traditional colocation and enterprise data centers but are increasingly pivoting to AI workloads.
• Emerging AI-Native Infrastructure Players: Companies like CoreWeave and Nebius (NBIS) operate on a similar model to Applied Digital — building large-scale, GPU-centric clouds. These are direct competitors but also potential customers or partners.
• Powered-Shell Developers: Specialists like Talen Energy (TLN) (through its Cumulus Data segment) and Lancium compete for hyperscale contracts on greenfield sites.

Applied Digital's competitive positioning relies on speed-to-market and power access. Unlike traditional REITs, which focus on dense urban interconnection, APLD targets rural areas with low-cost power, offering significant total-cost-of-ownership advantages for massive AI training clusters. Compared to CoreWeave, APLD is earlier in its development cycle, but its diversified pipeline of sites and focus on wholesale leasing (rather than solely operating its own cloud) provides a different risk profile.

DCF Discussion: A discounted cash flow analysis on Applied Digital is highly sensitive to assumptions regarding construction costs, financing rates, and the long-term contract renewal rates. Using a base case scenario that assumes the full contracted Ellendale campus becomes operational by FY2027, generating ~$700M in annual EBITDA, and applying a 10% discount rate with a 2% terminal growth rate, we derive an intrinsic value per share in the $35-$45 range. However, a bear case scenario, where construction delays push revenue out by 12 months and financing costs rise, yields a fair value closer to $15-$20. The wide divergence highlights the extreme optionality and risk embedded in the current share price. The stock trades at a premium to traditional data center REITs on a forward basis, reflecting the hyper-growth nature of its contracted backlog, but at a discount to pure-play GPU cloud providers like CoreWeave, which have more operational history.

• Execution and Construction Risk: The company's entire valuation hinges on delivering complex, multi-phase construction projects on time and within budget. Any significant delays in power interconnection (with grid operators like MISO or ERCOT), supply chain issues for transformers or cooling equipment, or cost overruns could severely impair the financial model and delay revenue recognition.
• Customer Concentration: A substantial portion of the company's contracted revenue is derived from a single customer (CoreWeave) for the Ellendale campus. If this customer faces financial distress, renegotiates terms, or defaults, Applied Digital would face a catastrophic loss of its primary revenue driver and would be left with a partially built, highly specialized asset.
• Financing and Dilution Risk: The company's expansion plans require billions in capital. Given its negative cash flow, it is reliant on debt markets and equity raises. A tightening of credit markets or a decline in its stock price could force the company to issue equity at dilutive levels or abandon certain projects, limiting its growth potential.
• Technological Obsolescence: The AI hardware cycle is rapid. While APLD primarily provides the physical infrastructure, its GPU Cloud segment is exposed to the risk that its deployed GPUs (e.g., H100s) become obsolete as newer architectures (e.g., Blackwell) are released, potentially reducing utilization and pricing power.
• Industry Competition and Power Costs: The intense competition for power-secured sites could drive up land and power costs, eroding the company's cost advantage. Additionally, any regulatory changes regarding energy pricing or renewable energy mandates could alter the economics of its North Dakota and Texas sites.