Applied Digital: Emerging Leader in AI Data Centers?

Applied Digital is a digital infrastructure company, primarily focused on building and operating energy-dense data centers for workloads such as artificial intelligence and high-performance computing. What differentiates it from classic technology operators is that its business model is not geared towards software development or training AI models, but towards providing the physical base needed to operate these systems on a large scale.

Concretely, APLD develops campuses consisting of data centers without proprietary servers or GPUs, but designed to host high-performance hardware. These facilities feature high-capacity connectivity, advanced cooling systems, and infrastructure optimized to support the latest generation of GPUs.

Customers, which include AI companies, hyperscalers, and cloud operators, use these data centers to run their workloads, directly installing their own hardware equipment, such as GPUs, memory, and chips. The company monetizes through long-term contracts based on energy capacity, infrastructure lease, and ancillary services related to configuration and operational management.

Business Model Positioning

This model places Applied Digital in the infrastructure provider segment, which makes it more like a real asset operator rather than a traditional technology company.

In fact, the difference compared to a cloud provider is substantial: while the latter also offers the computational layer, APLD focuses exclusively on infrastructure. From this, it can be deduced that its competitive advantage lies not in its ability to offer a complete service, but in the speed with which it can build AI-ready data centers, integrating energy and advanced cooling, which are emerging as the real bottlenecks in the sector.

Strategic Developments

In addition, management began a process to spin off the Applied Digital Cloud division and combine it with EKSO Bionics, resulting in ChronoScale, a platform dedicated to accelerated computing.

This transaction allows APLD to maintain its focus on the core infrastructure business, allowing the cloud segment to follow an independent growth path, with access to dedicated capital and a strategy more consistent with the compute market. At the same time, the relationship between the two entities remains strategic, as the cloud will benefit from privileged access to data centers developed by APLD.

Financial Performance

From a financial point of view, the data show a society in transition. Until 2024, revenue growth was subdued and still partly tied to the legacy crypto hosting business.

2025, on the other hand, represents the first real turning point, with revenues increasing sharply, up to about 319 million TTM, equal to a growth of over 120% on an annual basis. This increase reflects the entry into the HPC segment.

A share of the initial revenues derives from tenant fit-out activities, i.e. configuration and preparation of data centers for customers, including the installation of the infrastructure necessary to host the GPUs. These are low-margin, non-recurring revenues, which represent a preliminary phase to the full monetization of assets.

The true economic value of the model emerges with the activation of long-term leases, which begin to contribute only progressively with the entry into operation of data centers.

Margin and Cost Dynamics

This change is also reflected in the dynamics of margins. The gross margin rose from around 22% in 2024 to over 45% on a TTM basis, signalling a significant improvement in the business mix.

However, the company remains at a loss at the operating level, with a negative result of about 40 million, while operating expenses grow significantly, from about 83 million to over 180 million. This is consistent with a phase in which production capacity is built before recurring cash flows become fully operational.

Capital Intensity and Risk

The main element of attention remains the financial profile. Applied Digital is a highly capital-intensive business, characterized by very high investments, with capex exceeding 1.7 billion in the last period and strongly negative free cash flow, around -1.8 billion.

Added to this is a significant increase in debt, which reaches around 2.8 billion, and a strong equity dilution, with the number of shares outstanding more than doubling between 2024 and 2025. In essence, the company is converting capital into infrastructure, anticipating investments that should generate returns in the long term.

Contracts and Growth Visibility

Conference calls allow these numbers to be contextualized and show how the company has entered a new operational phase.

Management sees the real turning point in signing multi-year contracts with top-tier customers, including CoreWeave and an investment-grade hyperscaler, for hundreds of megawatts of capacity and a total value in the order of about $16 billion over 15-year horizons.

These agreements transform the model from an early build logic to a platform with visible and contracted future revenues, although full monetization will take time, with a progressive ramp expected between 2026 and 2027.

Execution Constraints

A particularly relevant element that emerged from the calls is that demand is not the main constraint for growth.

The management repeatedly emphasizes that the interest from hyperscalers is very high and that the pipeline includes several gigawatts of capacity under development. The real limit is rather related to the ability to execute, in terms of energy availability, construction time, supply chain, and specialized workforce.

Financing Strategy

At the same time, the company has developed a financial strategy based on project financing and partnerships with large institutions, with the aim of supporting growth while reducing the cost of capital over time.

In the current phase, debt is high because it finances asset construction; in the long term, the intention is to refinance on more favorable terms once the data centers are operational and backed by solid contracts.

In this sense, a possible interest rate reduction cycle could help improve the financial sustainability of the model, even if it remains linked to macroeconomic variables such as inflation and energy prices.

Long-Term Vision

Finally, the strategic vision of the management is clearly oriented towards the long term. Applied Digital aims to build a platform capable of generating multi-decade cash flows, typically between 15 and 30 years, moving closer to an infrastructure model similar to energy or real estate than to a cyclical technology business.

In this sense, the company is positioned as one of the players that provide the physical foundation necessary for the development of artificial intelligence on a global scale.

Conclusion

In conclusion, Applied Digital presents itself as a fast-growing company, well positioned on a long-term structural trend, also growing strongly, but still in an early stage from a financial point of view.

Success will depend on being able to execute projects on schedule, keeping customers of high quality, such as hyperscalers and CoreWeave, and optimizing the capital structure over time. If executed effectively, the business could evolve towards a stable platform that generates predictable cash flows; otherwise, the high level of investment and leverage is the main risk factor.

In light of the analysis, APLD appears to be an investment with an asymmetric risk-return profile, characterized by a significant upside but also by a still high level of risk.

For this reason, it can be included as a minority position in the portfolio, to be progressively increased upon the occurrence of operational confirmations in the next conference calls, in particular on capacity activation and contract monetization.