What Is Applied Digital (APLD)? AI Data Centers, HPC Infrastructure, and Cloud Computing Explained

Applied Digital (NASDAQ: APLD) is a company focused on AI data centers and high-performance computing (HPC) infrastructure. It builds and operates data centers designed for high-density computing loads, providing the necessary power, facilities, cooling, and support for cloud computing and large-scale computing clients.

As generative AI models continue to scale, the value of data centers extends beyond server hosting and cabinet rentals. GPU clusters require continuous, stable, and large-scale power supply, while also demanding higher standards for heat dissipation, network connectivity, power redundancy, and infrastructure deployment speed. Operators that can provide large-scale power capacity and deliver facilities on schedule have become key players in the AI infrastructure supply chain.

Applied Digital's business transformation reflects the changing demands for computing infrastructure: transitioning from early data center operations related to digital asset mining to providing dedicated facilities for AI training, inference, and other high-performance computing tasks. Its long-term lease agreements, Polaris Forge and Delta Forge campuses, along with funding and debt arrangements, determine the company's ability to convert AI market demand into sustainable operating revenue and cash flow.

Key Points

• Company Positioning: Applied Digital (APLD) specializes in AI data centers and HPC infrastructure, primarily offering data center construction, hosting, and related services to large computing clients.

• Business Transformation: The company has gradually shifted from its early data center operations related to cryptocurrency mining to high-density AI computing infrastructure to meet the growing demands for GPU training and inference.

• Core Projects: Polaris Forge and Delta Forge are important campuses for the company's expansion into AI Factory operations, with varying project capacities, construction progress, and leasing arrangements.

• Revenue Model: The company generates revenue mainly through base rent from data center facilities, tenant construction services, and related fee recoveries, with total contract value not equating to current revenue or net profit.

• Key Risks: High capital expenditures, debt financing costs, project delivery delays, power supply constraints, and customer concentration can all impact the company's cash flow and long-term profitability.

What is Applied Digital (APLD)? An Analysis of the Company's Business Position and Development History

Applied Digital is a U.S. data center infrastructure company focused on providing dedicated data centers for high-performance computing and AI workloads. The company's stock ticker is APLD, and it is listed on NASDAQ. Unlike companies that directly develop large AI models, sell GPUs, or provide AI applications for individual users, Applied Digital's primary value lies in constructing infrastructure capable of supporting large-scale computing clusters and generating revenue through leasing and support services.

The company's development path is closely tied to changes in the digital infrastructure industry. Its early business involved providing data center space and related facilities for cryptocurrency mining, gradually expanding its infrastructure operations to serve AI and HPC clients. As GPU clusters demand higher power density, cooling capabilities, and operational stability, the company has begun to allocate more resources to facilities specifically serving AI training and inference.

In 2026, Applied Digital's business focus will further shift towards the AI Factory model. The term AI Factory does not refer to a factory that manufactures AI chips, but rather integrates power supply, computing equipment deployment space, cooling systems, network connectivity, and operational capabilities to provide clients with a usable high-density computing environment. The company's long-term lease agreements also indicate that its goal is not just to sell data center space, but to provide large clients with computing infrastructure that can be delivered in phases and operated long-term.

What Are the Core Businesses of Applied Digital? An Analysis of Data Center Hosting and HPC Infrastructure

Applied Digital's business can be understood from three levels: HPC Hosting, traditional Data Center Hosting, and cloud service-related operations, though these businesses have different positioning and development stages.

HPC Hosting is currently the core of the company's AI infrastructure strategy. The company is responsible for the planning, construction, and operation of data centers, configuring power capacity, cooling systems, and other infrastructure according to client needs. Clients typically require the deployment of large numbers of GPUs or other accelerated computing devices, so data centers must meet requirements for high power density, continuous operation, and large-scale expansion. Applied Digital generates revenue through long-term lease agreements and related construction services.

The Data Center Hosting business primarily provides facilities with power access for cryptocurrency mining clients. This business shares some infrastructure commonalities with AI data centers, such as reliance on power supply, data center space, and continuous operational capability, but the equipment density, cooling solutions, client demands, and contract structures differ. The expansion of AI operations does not imply that the original mining hosting business has completely exited.

Cloud services represent another business line. Applied Digital has initiated plans to spin off its cloud service business in 2026, with ChronoScale taking over related operations. When analyzing the company's financial performance, it is essential to distinguish between the contributions of HPC infrastructure, traditional data center hosting, and cloud services, avoiding simplistic comparisons of revenue and profit margins across different businesses.

Why Did Applied Digital Shift from Cryptocurrency Mining to AI Data Centers?

Both cryptocurrency mining and AI computing require substantial power, but their infrastructure needs differ. Mining operations primarily focus on power costs, equipment operating efficiency, and the cycles of the digital asset market; AI data centers must handle high-density GPU clusters, with stricter requirements for power redundancy, heat dissipation, network bandwidth, and facility delivery times.

Another characteristic of AI infrastructure is that clients may lock in capacity through long-term lease agreements. For data center operators, long-term contracts help clarify some demand before construction and provide a basis for financing and subsequent expansions. Compared to businesses that rely entirely on short-term market prices, this model offers greater revenue visibility, provided that facilities can be delivered on time, clients fulfill their obligations, and construction costs are controlled.

However, transitioning from mining hosting to AI infrastructure does not mean that existing assets can be converted without cost. High-density GPU environments may require redesigning power systems, cooling architectures, data center layouts, and network connections. Projects require significant capital from planning to operational readiness, so business transformation is accompanied by construction cycles, financing, and technical execution risks.

Applied Digital's choice can be seen as a response to the changing structure of computing demand: extending its capabilities in building and operating data centers to high-value, high-density computing scenarios. However, the growth in AI demand is merely a market opportunity and does not automatically guarantee project profitability; the ultimate performance still depends on contract prices, actual delivery capacity, capital costs, and operational efficiency.

How Does Applied Digital Build AI Factories? An Analysis of High-Density Computing Infrastructure

AI Factory focuses on integrating previously dispersed infrastructure into a complete system capable of supporting large-scale computing. Applied Digital is primarily responsible for constructing data centers that accommodate high-density computing loads, while GPUs, servers, and related computing devices are typically deployed by clients or their designated partners.

The construction process involves several interdependent stages. First, the available power capacity and grid connection conditions must be determined, followed by designing substations, power distribution, backup power, and cooling facilities, and then completing data center construction, equipment deployment, and system debugging. Any delay in one stage can affect the overall project delivery timeline.

In capacity planning, two indicators must be distinguished: critical IT load and grid-side power capacity. The former refers to the critical power capacity available for IT devices like servers, while the latter must also consider cooling, distribution losses, and other facility power needs. Therefore, a 300 MW critical IT load does not imply that the entire campus only requires 300 MW of grid power capacity.

Applied Digital's campuses adopt a phased construction approach, allowing different buildings to gradually reach a deliverable state. This model helps control the pace of one-time investments but also means that contracted capacity, in-construction capacity, and operational capacity must be tracked separately. Only capacity that enters operational status and begins providing services per contract can continuously contribute to base rent income.

How Does Applied Digital Serve Large-Scale Cloud Computing Clients Through Long-Term Lease Agreements?

Long-term lease agreements are a crucial component of Applied Digital's business model. Large cloud computing clients need to plan for AI computing capacity expansion over the coming years, and constructing dedicated data centers takes time, so both parties can lock in facility capacity, lease terms, and delivery arrangements through long-term agreements.

For example, in 2026, Polaris Forge 2 signed a lease agreement for approximately 200 MW of critical IT load, with a base lease term of about 15 years and a disclosed contract value of approximately $5 billion. Delta Forge 1 signed a lease arrangement totaling about 300 MW, also with a base lease term of about 15 years, corresponding to a contract value of approximately $7.5 billion.

These figures reflect the expected contract revenue over specific contract periods, not the cash already received at the time of signing, nor do they represent the project's net profit. Construction investments, financing costs, operating expenses, and potential delivery delays can all impact the final economic returns.

Some agreements adopt a take-or-pay structure, meaning that under the agreed conditions, clients may still need to pay the agreed fees even if they do not fully utilize the reserved capacity. Such arrangements enhance revenue predictability, but the specific coverage still depends on contract terms, delivery conditions, and related breach arrangements.

Additionally, client names may not all be disclosed. Even if clients have high credit ratings, customer concentration risk cannot be entirely eliminated. For Applied Digital, the quality of contracts should be assessed not only by their value but also in conjunction with client credit, contract duration, project progress, actual payment arrangements, and future renewal conditions.

What is Applied Digital's Business Model? An Analysis of Data Center Rent, Construction Services, and Operating Revenue

Applied Digital's revenue does not solely come from long-term rent. According to company disclosures, HPC Hosting revenue also includes tenant construction services, tenant fee recoveries, and other projects, so quarterly revenue growth may not entirely stem from new recurring rent.

• Base Rent Revenue: Revenue is recognized according to the lease agreement once the data center meets the delivery conditions and begins operations. This is an important indicator for observing long-term recurring revenue capacity.

• Tenant Construction Services: Construction services related to client onboarding and facility configuration may generate phased revenue, but they typically come with corresponding costs and cannot be directly viewed as rental profit.

• Fee Recoveries and Other Revenue: This includes recoveries of certain power or operating fees as stipulated in the contract, with specific amounts depending on contract structure and actual operating conditions.

From an operational analysis perspective, base rent better reflects the ongoing revenue capacity generated by operational assets; construction service revenue needs to be assessed in conjunction with project progress, costs, and gross margins. If a quarter sees a significant revenue increase, but the growth primarily comes from tenant construction services, it cannot be simply inferred that the same growth rate will be maintained in future quarters.

The company also needs to continuously invest capital in building new campuses. Before project delivery, related assets may have already incurred construction expenses and financing costs without generating complete rental income. Therefore, when analyzing APLD's business model, it is essential to observe revenue composition, project operational progress, operating profits, and free cash flow simultaneously.

How Does Applied Digital Address the Power, Cooling, and Infrastructure Needs of AI Data Centers?

Power is one of the most critical constraints in building AI data centers. Large-scale GPU clusters require stable power supply, and power capacity is influenced by local grid capacity, substation construction, grid connection approvals, and the progress of power infrastructure construction. Even if land has been secured, it does not guarantee that the project can immediately access sufficient usable power.

Cooling is another key issue. Traditional data centers can use air cooling within certain limits, but high-density GPU clusters generate more concentrated heat loads, making liquid cooling and other technologies potentially important infrastructure solutions. Cooling design must be coordinated with server configurations, cabinet density, and power systems, and cannot be resolved solely by increasing cooling equipment.

In its AI Factory projects, Applied Digital emphasizes high-density power supply, advanced cooling architectures, and facility designs suitable for large-scale computing. The company selects specific regions for campus construction based on power availability, climate conditions, and expansion space. Cooler climates may help reduce some cooling burdens, but actual energy efficiency still depends on facility design, local climate, and operational methods.

These conditions create competitive barriers for data center operators and pose project risks. Delays in power access, failures in cooling system debugging, rising construction costs, or delays in equipment delivery can all postpone client onboarding and affect revenue recognition.

What Are the Differences Between Applied Digital's Polaris Forge and Delta Forge Projects?

Polaris Forge and Delta Forge are two brands of campuses for Applied Digital's expansion into AI Factory operations. Both aim to provide dedicated infrastructure for large-scale AI and HPC workloads, but they differ in site selection, construction phases, planned capacities, and leasing arrangements.

Note: The capacities in the table reflect disclosed lease arrangements and do not represent the total planned capacity of the campuses, nor do they indicate that all facilities are operational. Project status may change with construction progress.

Polaris Forge 1 is one of the earlier operational HPC campuses, helping to observe the transition from the construction phase to the rental income phase. Polaris Forge 2 and 3 demonstrate the company's strategy of continuing to expand individual campus capacity and locking in demand through long-term contracts.

Delta Forge 1 and Delta Forge 2 further extend the company's regional layout. Power supply, construction costs, climate conditions, and infrastructure conditions differ across regions, so it is not sufficient to judge which campus has higher economic returns based solely on project names or planned capacities. What truly matters are the actual investment amounts, deliverable capacities, rental prices, operating costs, and project commissioning timelines.

How to Interpret Applied Digital's Latest Financial Report? An Analysis of Revenue Growth, Contract Reserves, and Profitability Pressure

As of October 9, 2026, Applied Digital's latest financial report covers the first quarter of the 2027 fiscal year, ending August 31, 2026. The company's quarterly revenue reached $341.9 million, a year-over-year increase of 322%; the net loss attributable to common shareholders was $221 million, compared to a net loss of approximately $18 million in the same period last year. During this time, adjusted EBITDA was $64.4 million.

This data reflects two concurrent trends: on one hand, the company is experiencing rapid revenue growth, with the HPC Hosting business beginning to contribute more rental and support service revenue; on the other hand, the net loss has also significantly widened, indicating that revenue growth has not yet directly translated into GAAP profitability.

Particular attention should be paid to the revenue structure. In that quarter, HPC Hosting revenue was $262.6 million, with base rent at $65.8 million, tenant construction service revenue at $183.5 million, and tenant fee recoveries at $13.3 million. The high proportion of construction service revenue indicates that quarterly revenue is significantly influenced by project construction and client onboarding progress.

Contract reserves provide another perspective for observation. At the end of the 2026 fiscal year, the company disclosed that the five AI Factory campuses had a contracted critical IT load of approximately 1,410 MW, with an initial 15-year lease term corresponding to contract revenue of approximately $36 billion; if all related renewal options are exercised, the potential total contract value could reach approximately $86 billion.

However, the total contract value is a cumulative amount spanning multiple years, not the current asset value, current revenue, or cash flow already realized. When analyzing, it should be considered alongside operational capacity, the timing of revenue recognition, future construction funding needs, and actual cash recovery situations.

What Challenges Do Applied Digital's Capital Expenditures and Debt Structure Face?

Building AI data centers requires significant upfront investment before gradually generating rental income. Land, power systems, substation facilities, cooling equipment, construction, and client delivery preparations all incur capital demands. Therefore, even with long-term lease agreements, the company still needs to address funding gaps from signing to commissioning.

Applied Digital has continued to increase construction investments in the 2026 fiscal year. The company disclosed that some campuses are still in various stages of construction and are supported by debt financing for project development. For instance, the company disclosed the issuance of senior secured notes totaling approximately $1.59 billion with a coupon rate of 7%, intended to repay bridge financing and support the subsequent construction of Polaris Forge 1.

Such financing can help the company advance projects ahead of schedule but also increases interest expenses and debt obligations. Even if a project ultimately has strong rental income potential, if construction costs exceed budget, commissioning is delayed, or client payments are later than expected, the company may still face cash flow pressures.

Capital expenditure risks primarily manifest in three areas. First, the construction cycle and revenue recognition timing are inconsistent; expenditures occur before projects generate rental income. Second, debt financing costs can affect the final returns of projects, especially in high financing rate environments or when refinancing conditions tighten. Third, data center construction requires continuous expansion to realize large contract reserves, and new projects also mean new funding needs, making it insufficient to rely solely on signed amounts to assess financial safety.

Additionally, investors should pay attention to debt maturity timelines, secured assets, financing covenants, available cash, capital expenditure commitments, and potential equity financing needs. Long-term lease agreements can enhance revenue visibility but do not entirely eliminate construction risks and balance sheet pressures.

What Position Does Applied Digital Hold in the AI Infrastructure Supply Chain?

Applied Digital is positioned in the midstream of the AI infrastructure supply chain, connecting upstream resources such as power, land, construction, cooling, and networking with downstream cloud computing platforms, hyperscale cloud service providers, and AI computing clients.

Its business model differs from that of GPU manufacturers. Chip companies primarily generate revenue through the sale of computing hardware, while Applied Digital earns revenue by building, delivering, and operating data center facilities capable of supporting this hardware. It is not entirely equivalent to traditional cloud service providers, as its core business leans more towards physical infrastructure and capacity leasing rather than directly providing complete cloud computing products to end clients.

This positioning allows Applied Digital to benefit from the growing demand for AI computing power, but it also means it must contend with the real constraints of the infrastructure industry: power capacity scarcity, long construction cycles, high initial investments, and project returns dependent on lease contracts and operational efficiency.

When comparing with peers, it is not sufficient to look at how many MW of campus capacity or how many billions of dollars in contract value a company has announced. More meaningful indicators include construction costs per MW, operational capacity, the proportion of rental income, campus operating profit margins, net debt levels, client credit quality, and capital return rates.

In the long term, Applied Digital's competitive advantage will depend on its ability to continuously secure available power, deliver high-density data centers on schedule, and convert contracted capacity into stable operating cash flow at reasonable financing costs. The expansion of the AI industry provides market space, but project execution and financial discipline remain key factors determining operational outcomes.

Summary

Applied Digital (APLD) is a company centered on AI data centers and high-performance computing infrastructure, with its business transformation reflecting the trend of the data center industry shifting from traditional hosting to high-density AI computing facilities. The company is expanding its business scale through campuses like Polaris Forge and Delta Forge, as well as long-term lease agreements, and is beginning to generate more HPC infrastructure revenue.

However, contract reserves, quarterly revenue growth, and ultimate profitability are three distinct concepts. Tenant construction services may drive rapid short-term revenue increases, but the long-term quality of operations still needs to be validated by base rent, operational capacity, and cash flow. Meanwhile, large capital expenditures and debt financing also expose the company to significant construction and financial risks.

Understanding APLD requires simultaneously assessing market demand, project delivery, contract quality, rent growth, and balance sheet health. Only when contracted capacity gradually converts into operational assets and generates cash flow sufficient to cover operating expenses, interest, and future capital needs can the expansion of AI infrastructure truly translate into sustainable operational value.

FAQ

What Kind of Company is Applied Digital (APLD)?

Applied Digital is a data center infrastructure company listed on NASDAQ under the ticker APLD. The company primarily builds and operates data centers for AI and high-performance computing (HPC), while also retaining data center hosting operations for cryptocurrency mining clients. Its AI infrastructure business mainly generates revenue through long-term leases and related construction services.

How Does Applied Digital Differ from CoreWeave?

The two companies have different positions in the AI infrastructure supply chain. Applied Digital focuses more on the planning, construction, operation, and leasing of data centers, while CoreWeave centers on GPU cloud computing services, providing clients with computing resources and related cloud services. The two can form collaborative relationships at the infrastructure and computing service levels, but their business models are not the same.

What Are Polaris Forge and Delta Forge?

Polaris Forge and Delta Forge are campus project series used by Applied Digital to expand its AI data center business. The former includes campuses in North Dakota, while the latter includes campuses in the southern U.S. Each project varies in capacity, lease contracts, construction progress, and expected commissioning timelines and should be tracked separately.

Where Does Applied Digital's Revenue Primarily Come From?

The company's revenue includes base rent from HPC data centers, tenant construction service revenue, related fee recoveries, and income from traditional data center hosting operations. The sustainability and profit margins of different revenue sources vary, so when analyzing its operational performance, special attention should be paid to the growth of base rent and the costs associated with construction service revenue.

What Risks Should Investors Focus on Regarding APLD?

Key risks include delays in data center construction, insufficient power access, capital expenditures exceeding budgets, debt and interest burdens, customer concentration, and changes in AI infrastructure demand. Even with long-term lease agreements, the company must complete construction and delivery to gradually realize contract revenue. When assessing operational quality, it is essential to consider operational capacity, cash flow, debt structure, and capital return rates comprehensively.