A Trio of Capital and Compute: The Rise and Paradigm Shift of Data Center Private Credit in the Age of Artificial Intelligence

Chapter 1: The Epochal Context – The Capital Chasm in New Digital Infrastructure

We are in an era of structural transformation driven by Large Language Models (LLMs) and Generative AI. The depth and breadth of this transformation far exceed mere software-level iterations; it is, in essence, a revolution in physical infrastructure. If the factories of the steam engine era were the physical vessels of industrialization, then in the age of artificial intelligence, data centers are the "digital factories" that house computational power and drive intelligence. These massive architectural clusters, deploying tens of thousands of high-performance computing chips and consuming enormous amounts of electricity, form the bedrock of the digital world. However, this exponential growth in demand for compute has directly given rise to an unprecedented need for capital—a capital chasm measured in the trillions of dollars.

Traditional financing channels—whether public debt issuance, equity financing, or commercial bank loans—have all revealed their structural limitations when faced with the unique demands of AI on data center infrastructure. The public markets demand high levels of standardization, transparent information disclosure, and lengthy issuance cycles, which are ill-suited to the rapid technological iteration and fleeting project opportunities characteristic of AI infrastructure development. While flexible, equity financing entails dilution and a high cost of capital that project sponsors are unwilling to bear for asset-heavy, high-leverage data center projects. Meanwhile, traditional commercial banks, constrained by strict capital adequacy regulations (such as the Basel III accords) and a relatively conservative risk appetite, often shy away from "Greenfield" construction projects that are structurally complex and lack initial cash flow. Their rigid credit decision-making processes are incapable of providing the customized capital structure solutions that these projects require.

It is against this backdrop that Private Credit, as an alternative asset class, has rapidly risen to prominence, becoming the key force to bridge this capital chasm by virtue of its unique advantages. Private credit funds, with their long-term, patient capital raised from institutional investors (such as pension funds, sovereign wealth funds, and insurance companies), can offer a degree of flexibility, execution speed, and structural customization that traditional financing channels cannot match. As a private credit investment team, we are not merely providers of capital; we are architects of transactions, managers of risk, and deeply involved participants in the success of the projects. Our role is to deeply understand the intrinsic value and risks of data centers as a special class of asset, and to design credit products that not only meet the sponsor's funding needs across different life cycle stages (construction, lease-up, and stabilization) but also deliver attractive, risk-adjusted returns to our investors.

This paper aims to systematically deconstruct the full life cycle of data center private credit investment in the age of AI from the practical perspective of a front-line investment team. We will eschew broad, macroeconomic discourse to focus on the details of micro-level execution, delving deep into every link of the chain: from deal sourcing, due diligence, financial modeling, and legal structuring, to the covenants and collateral details within credit agreements, and finally to post-investment management, risk control, and exit. The core thesis of this paper is that private credit is not just a "blood transfusion" for AI data center construction; rather, through its sophisticated transaction structuring and rigorous risk management, it is a profound participant in and shaper of the physical form and commercial landscape of this compute infrastructure revolution. We will reveal how, in this high-growth, high-capital-consumption arena, prudent credit principles can be combined with cutting-edge technological trends to capture this historic investment opportunity while keeping risks under control.

Chapter 2: Deconstructing the Data Center Credit Asset – Investment Logic and Risk Identification

As credit investors, our primary task is not to predict the future, but to manage downside risk. Therefore, the evaluation of a data center credit investment begins with a deep deconstruction of the core value of the asset as effective collateral. The credit value of a successful data center project stems not from the building itself, but from its ecological niche, its contracted long-term cash flows, and its intrinsic infrastructure attributes. We will articulate our investment screening framework through the following dimensions.

2.1 Geography, Power, and Network: The Trinity of Infrastructure

The value of a data center is rooted in its physical location. However, "location" here is far from the traditional commercial real estate concept of "prime real estate"; it is a three-dimensional coordinate system composed of power, network, and market demand.

Power Sovereignty is the primary cornerstone of our evaluation system. A data center is essentially a power conversion facility; its business model can be simplified as "buying electricity, converting it into compute via servers, and then leasing out that compute." Therefore, the ability to secure a long-term, stable, cost-controlled, and scalable power supply is the project's lifeline. In our due diligence, we dedicate significant resources to scrutinizing the project's Power Purchase Agreement (PPA). An ideal PPA should feature a long term (typically 10-20 years), a locked-in price or a clear pricing formula, and a supply guarantee from a reliable utility (usually a regional power giant). We engage independent power consultants to assess the capacity, stability, and redundancy of the local grid, even tracing back to the physical connections and upgrade potential of the substations. In the current context of increasing importance of ESG (Environmental, Social, and Governance), we increasingly favor projects that can secure green energy PPAs. This is not only for social responsibility but also because many large technology companies (the end tenants) have made carbon neutrality commitments, making green power a hard requirement for their data center selection, which directly translates into the project's commercial competitiveness.

Connectivity is another lifeline for a data center. Data needs to be transmitted with low latency and high bandwidth. We assess whether the project site is close to backbone fiber optic networks and how many telecommunications carriers have connected to the facility. A "Carrier-Neutral" data center, by offering tenants a diversity of network choices and thus avoiding the risk of being locked in by a single provider, has a much higher asset value and leasing appeal than a single-carrier facility. We carefully review "Conduit Rights" and "Access Agreements" to ensure the project has the physical and legal rights to bring fiber into the building. The design and operational capabilities of the "Meet-Me Room (MMR)," where core network exchanges are located, are also a key focus.

Market Demand and Geographic Location ultimately determine the project's commercial value. We analyze the supply and demand dynamics of the data center market in the project's Metropolitan Statistical Area (MSA), including existing inventory, projects under construction, absorption rates, and vacancy rates. For Hyperscale Data Centers serving AI and cloud computing, the site selection logic is often "decentralized"—that is, not concentrated in traditional financial centers but located in areas with abundant renewable energy, low land and power costs, and minimal risk of natural disasters. However, these areas must still be able to serve major population and commercial centers with acceptable latency.

2.2 Tenant Credit and Lease Agreements: The Ultimate Anchor of Credit Risk

The risk assessment of data center private credit is, to a large extent, a Proxy Analysis of its tenant credit quality. This is especially true for projects that are fully or largely leased to hyperscale customers (such as top-tier cloud service providers and social media giants), where the stability of cash flow is almost entirely dependent on the tenant's ability to perform and the robustness of the lease agreement they sign.

Our analysis begins with a deep dive into the tenant's financial condition. We evaluate their balance sheet, cash flow situation, credit rating (if any), and their market position and long-term strategy in the cloud computing or AI space. The tenants we want to see are those who view the data center as "Mission-Critical"infrastructure for their core business. This means that even if the tenants themselves face financial distress, they will prioritize paying the data center rent to avoid the catastrophic consequences of business interruption.

Next, we scrutinize the Master Lease Agreement (MLA) line by line. This is typically a long-term "Triple-Net (NNN)" lease of 10-15 years or even longer. [NNN Lease] In the context of a specialized facility like an AI data center, an NNN lease is a critical risk-transfer mechanism. It means the tenant (often a well-capitalized tech giant) not only pays rent but also directly bears all operating costs, including property taxes, insurance, and—most importantly—maintenance expenses, particularly the highly technical upkeep associated with power and cooling systems. This makes the cash flow we rely on as creditors extremely pure and predictable, largely insulated from the volatility of the facility's day-to-day operations.

We pay special attention to the following clauses:

• Rent Composition and Escalation Mechanisms: Rent is usually denominated in "dollars per kilowatt per month." We look for whether the rent includes fixed annual escalation steps (typically 2-3%) to hedge against inflation.

• Renewal and Termination Rights: Tenants usually have multiple renewal options, but our focus is on whether the landlord has the right to refuse renewal under specific conditions. At the same time, we strictly scrutinize any clauses that could allow the tenant to terminate the contract early, and we strive to limit them to extreme situations (such as the complete destruction and inability to repair the asset). Any form of a "Termination for Convenience" clause is unacceptable.

• Bankruptcy Treatment: We analyze the legal standing of the lease agreement should the tenant enter bankruptcy proceedings. Under U.S. bankruptcy law, a commercial lease is an "Executory Contract." [Executory Contract] This legal concept is paramount for an AI data center creditor. It means that if a tenant (e.g., a cloud services behemoth) files for bankruptcy, it cannot unilaterally and casually suspend the lease. Instead, it must make a formal legal decision: to "Assume" the contract and continue performing all obligations (including paying rent), or to "Reject" it and terminate the lease entirely. Given the mission-critical nature of an AI data center to the tenant's core business—where a service interruption could cripple its entire cloud platform—the tenant will, with very high probability, choose to assume the contract. This legal mechanism provides a powerful protection for our source of repayment, even in the face of a tenant's extreme financial distress. This is a crucial safeguard for our credit security.

2.3 Technical Specifications and Future-Proofing: The Long-Term Value of the Collateral

Finally, we must evaluate the design and technical specifications of the data center itself to ensure its market competitiveness throughout the loan term and beyond, thereby guaranteeing its long-term value as collateral.

We assess its Tier Rating. Certified by the Uptime Institute, Tier I to Tier IV ratings represent the level of redundancy and fault tolerance of a data center in terms of power and cooling. Tier III and Tier IV facilities are favored by high-end customers for their high reliability and are also our preferred collateral.

Power Usage Effectiveness (PUE) is a key metric for measuring a data center's energy efficiency. Its formula is: PUE = Total Facility Energy / IT Equipment Energy. [PUE] This formula calculates the ratio of a data center's total energy consumption (including cooling, lighting, etc.) to the energy actually consumed by the IT equipment (servers, storage). A PUE value close to 1.0 means less energy is wasted on non-computing support facilities, resulting in lower operating costs. For an AI data center that consumes vast amounts of power, a minor improvement in PUE can save millions of dollars in electricity costs, directly boosting the project's Net Operating Income (NOI) and its Debt Service Coverage Ratio (DSCR). We review the design PUE and conduct sensitivity analysis on it in our financial model.

As the power density of AI chips increases dramatically, traditional air cooling methods are facing bottlenecks. We are paying increasing attention to designs that have or have reserved capabilities for liquid cooling. A facility that cannot adapt to the future demands of high-density computing may face the risk of technical obsolescence within a few years, thereby eroding its collateral value. We work with engineering consultants to assess whether the project's power supply, cooling pipes, floor loading capacity, and other design elements have reserved sufficient "Headroom" for future technology upgrades. This consideration of "Future-Proofing" is essential for prudent, long-term credit investing.

Through a systematic deconstruction of these three dimensions, we can form a comprehensive judgment on the intrinsic credit quality of a data center project. This process is data-driven, detail-oriented, and highly interdisciplinary, integrating finance, engineering, law, and market analysis. It constitutes the first and most important line of defense in our investment decision-making.

Chapter 3: The Practice of Due Diligence – A Journey of Verification from Paper to Reality

Following theoretical analysis and preliminary screening, we enter the most critical and time-consuming phase of the investment process: comprehensive Due Diligence. The objective of this phase is to take the assumptions established in our investment memorandum and subject them to rigorous, independent third-party verification and our own team's on-the-ground inspection, confirming or refuting them one by one. This is a process of transforming an "asset on paper" into a "tangible, quantifiable risk exposure."

3.1 Technical and Engineering Diligence: Delving into the Physical World

Data centers are technology-intensive assets, and the reliability of their physical attributes is directly linked to the security of our credit. We typically engage top-tier, independent engineering consulting firms with deep experience in the data center sector to conduct a full-spectrum technical review of the project.

For "Greenfield" or new construction projects, the focus of the review is on the design plans and construction schedule. The engineering consultant will review all design drawings, from the building structure, electrical systems, and mechanical cooling systems to the fire protection and security systems. They will verify whether the designed redundancy meets the claimed Tier rating. For example, for a Tier III data center, we would verify that it has N+1 redundancy for power and cooling, meaning that if any single critical component (such as a UPS, generator, or chiller) fails, a backup unit can immediately take over, ensuring no interruption of service. We review the "Commissioning Plan" to ensure all systems undergo rigorous stress testing before the data center is delivered.

The construction budget and timeline are another core area of review. We will compare the budget provided by the Sponsor against industry standard databases to identify any unrealistic cost estimates. We require the budget to include an adequate "Contingency" (typically 5-10% of hard costs) to cover unforeseen construction issues. We conduct background checks on the general contractor's qualifications, track record, and financial condition. The loan agreement will stipulate strict conditions for fund draws; we will only disburse funds proportionally based on the percentage of work completed as confirmed by our engineering consultant, and we require the Sponsor to contribute all of its equity capital first, ensuring their "skin in the game" is aligned with ours.

For "Brownfield" (redevelopment projects) or operating "Stabilized Assets," the focus of technical diligence shifts to assessing the condition of the existing facility, its remaining useful life, and its potential capital expenditure needs. The engineering consultant will conduct a comprehensive "Property Condition Assessment (PCA)," examining the operating records, maintenance logs, and replacement cycles of critical equipment. They will produce a 10-15 year capital expenditure forecast, which will be directly integrated into our financial model to ensure that cash flow can cover necessary maintenance and upgrade expenses.

3.2 Commercial and Market Diligence: Validating the Source of Cash Flow

The core of commercial diligence is to validate the project's revenue assumptions. We independently research the micro-market in which the project is located. We subscribe to and analyze market reports from professional data providers (such as Cushman & Wakefield, JLL) to understand regional rental rates, absorption velocity, and new supply pipelines. We may even "Mystery Shop" competing data centers to obtain the most authentic leasing market intelligence.

The review of lease agreements goes beyond the legal text to verify the commercial substance. We will request a confirmation call with the end tenant to verify the authenticity and validity of the lease. We analyze the tenant's "Fit-out Costs." Typically, a data center is delivered as a "Powered Shell," and the tenant must invest significant capital to install server racks, cabling, and other IT equipment. High fit-out costs mean high "sunk costs" for the tenant, which greatly reduces the likelihood of them relocating at the end of the lease term, thereby enhancing the long-term stability of the cash flow we rely on.

3.3 Legal and Regulatory Diligence: Building a "Moat" of Risk Isolation

Legal diligence is led by the external law firm we retain. Its mission is to ensure that our collateral has clear and unencumbered title, is in full legal compliance, and that the transaction structure achieves maximum risk isolation.

Title and Land Diligence is fundamental. The lawyers will conduct a comprehensive "Title Search" to ensure the project land is free of any undisclosed liens, easements, or ownership disputes. They will review the "Survey" to confirm that the property boundaries are consistent with government records. For projects on leased land, we will review the "Ground Lease" to ensure its term is significantly longer than our loan term and that it does not contain any harsh provisions that could lead to early termination.

Permitting and Compliance Diligence ensures the legality of the project's construction and operation. The lawyers will verify that the project has obtained all necessary government permits, including zoning approvals, building permits, and environmental impact assessments. In today's world of increasingly strict data privacy and sovereignty regulations (such as the EU's GDPR), we also assess whether the data center meets the relevant compliance requirements at both the physical and operational levels, as this directly affects its attractiveness to certain customers.

Entity Structure Diligence is key to achieving "Bankruptcy Remoteness." We typically require the borrower to be a "Special Purpose Vehicle (SPV)." [SPV & Bankruptcy Remoteness] This legal structure is the bedrock of AI data center project finance. Given that projects can involve investments of hundreds of millions or even billions of dollars, we require the sponsor to establish a separate SPV to act as the borrowing entity for that specific project. This SPV conducts no other business apart from owning and operating the data center. The ingenuity of this structure lies in using strict limitations within the SPV's corporate charter (e.g., prohibiting additional debt, restricting asset disposals) and appointing an independent director to completely isolate the project's assets and liabilities from the sponsor's own financial risks. Even if the sponsor's other businesses or projects fail or go bankrupt, our collateral and source of repayment, as the data center project's creditor, remain unaffected.

Due diligence is an arduous and detail-oriented process. It requires the investment team to possess a skeptical mindset, an extreme attention to detail, and the ability to coordinate internal and external experts across various disciplines. A credit decision made without the validation of rigorous due diligence is akin to navigating in the dark. Only when every key assumption is supported by solid evidence will we advance the project to the next stage: structuring the final transaction and credit agreement.

Chapter 4: The Art of the Deal – Financial Modeling and Capital Structure Design

Once due diligence confirms the asset's fundamentals, our focus shifts to quantitative analysis and transaction structuring. The goal of this stage is to build a financial model that not only reflects the project's future cash flow potential but can also withstand various stress scenarios, and then, based on this model, to design a capital structure that balances risk and return.

4.1 The Dynamic Financial Model: Forecasting Future Cash Flows

The financial model we build is far from a static spreadsheet; it is a complex, multi-dimensional dynamic system used to simulate the financial performance of the data center throughout its entire life cycle, from construction and stabilization to its ultimate exit.

The Revenue Build is the starting point. For pre-leased projects, revenue forecasting is relatively straightforward, based on the rental rate ($/kW/month), leased capacity (kW or MW), rent escalation steps, and lease term stipulated in the contract. For any speculative development portion (i.e., not pre-leased), we forecast future revenue based on assumptions for market rent, lease-up velocity (Absorption Rate), and final stabilized occupancy, all derived from our commercial due diligence. We conduct sensitivity analysis on these key assumptions. For example, if the lease-up pace is 6 or 12 months slower than expected, how are the project's cash flow and debt service capacity impacted?

The OpEx Model is a critical variable. Power cost is the largest operating expense. We model it precisely based on the pricing terms in the PPA and our PUE expectations. We will separate power costs into fixed and variable components, with the latter being directly correlated to the IT load factor. Other operating costs, such as staff salaries, maintenance contracts, insurance, and property taxes, are modeled based on comparable project data and third-party consultant reports. We also conduct stress tests: for instance, if electricity prices rise by 20% or property taxes are reassessed, to what level will the DSCR (see below) fall?

The CapEx Model must account for the entire loan cycle. In addition to the initial construction costs, the model must include Maintenance CapEx for the periodic replacement of equipment (e.g., UPS batteries, chiller fans) as well as Leasing Commissions & Tenant Improvements (LC/TI), which are cash outflows that may occur when attracting new tenants or retaining existing ones.

Based on these inputs, the model generates a complete set of pro-forma financial statements (Income Statement, Balance Sheet, Cash Flow Statement) for the next 10-15 years. From these, we calculate the core metrics for assessing credit risk:

• Debt Service Coverage Ratio (DSCR): This is the most central metric for measuring the project's ability to cover its loan principal and interest payments with its cash flow. Its formula is: DSCR = Net Operating Income (NOI) / Total Debt Service (Principal + Interest). For a stabilized data center, we typically require the DSCR to be maintained above 1.25x to 1.40x. In the model, we test the revenue decline or cost increase scenarios under which the DSCR would fall below 1.0 (i.e., cash flow cannot cover debt service). This point is known as the "break-even point."

• Loan-to-Cost (LTC) / Loan-to-Value (LTV): LTC is used to measure the leverage level of a construction loan, while LTV is used for a loan on a stabilized asset. For senior secured loans, LTC is typically controlled between 60-70%, and stabilized LTV between 55-65%. These ratios determine our "Equity Cushion," i.e., the percentage by which the asset's value can decline before our loan principal begins to suffer a loss.

• Debt Yield: This metric measures the project's Net Operating Income relative to the loan principal, without regard to loan amortization or interest rate. Its formula is: Debt Yield = Net Operating Income (NOI) / Loan Amount. Debt Yield provides a "purer" perspective on the asset's return, unaffected by the loan terms. We typically require a stabilized debt yield in the 9-11% range.

4.2 The Layering and Pricing of the Capital Structure

Based on the analysis from the financial model, we begin to design specific credit products. The capital structure of a data center is typically layered:

Senior Secured Loan: This is the product we provide most often. As senior creditors, we hold a first-priority lien on all of the project's assets. This type of loan has relatively low risk and is therefore priced more conservatively, typically at a floating benchmark rate (like SOFR) plus a Credit Spread. This spread has a wide range, from 250 to 550 basis points, depending on the project's risk profile (Greenfield vs. Stabilized), leverage level, tenant credit quality, and the competitive market environment.

Mezzanine Debt: This sits between the senior loan and the equity. The collateral for mezzanine debt is typically a pledge of the equity interests in the project company (the SPV), and its repayment is subordinated to the senior loan. Because it carries higher risk, its pricing is also higher, often in the form of a 10-14% fixed rate or PIK interest.

Payment-in-Kind (PIK) Interest: This is a structure that requires special caution in our discussion. PIK interest allows the borrower to defer cash interest payments for a certain period (usually during the cash-constrained construction phase) by adding the accrued interest to the loan principal. While this structure provides the project with breathing room, it causes the loan principal to grow rapidly, thereby increasing the project's overall leverage and our ultimate risk exposure. When we observe PIK structures being used excessively in the market, or used to mask an asset's underlying cash flow deficiency, it is often a danger signal of an overheating market and declining credit standards. In our own transactions, we strictly limit the use of PIK, typically only allowing it during the first 12-24 months of construction, and we charge a higher interest rate on loans that include a PIK feature to compensate for the additional risk we are taking.

In addition to the interest rate, our return also comes from various Fees, including a 1-2% Upfront/Arrangement Fee charged at closing, and a potential Exit Fee or Prepayment Penalty if the loan is repaid early. These fee structures are designed to ensure we can achieve our targeted minimum return even if the loan is refinanced ahead of schedule.

The final transaction structure is the art of seeking the optimal balance between risk, return, and market acceptability, all based on rigorous quantitative analysis. It requires a profound understanding of the asset's underlying drivers and the ability to translate that understanding into concrete, enforceable financial terms.

Chapter 5: The Bedrock of the Credit Agreement – Covenants, Collateral, and Control

If the financial model is the "soul" of the transaction, then the Credit Agreement is its "skeleton." This legal document, often running hundreds of pages, precisely defines the rights and obligations of the lender and borrower. It is our ultimate legal weapon as creditors to protect our interests and control project risk. The core of the agreement lies in constructing an interconnected system of protection composed of collateral, covenants, and control rights.

5.1 The Watertight Collateral Package

Our loan must be fully secured by all of the project's tangible and intangible assets. The design of this collateral package aims to ensure that, in the event of a borrower default, we can smoothly and completely take over and dispose of the entire project to recover our principal and interest.

• First-Priority Mortgage/Deed of Trust on Real Property: This is the most fundamental piece of collateral, covering the land and the buildings upon it. By recording this instrument in the local government's real property records, we obtain the highest priority claim on the real estate.

• First-Priority Security Interest in all Personal Property: A significant portion of a data center's value comes from its critical equipment, such as generators, uninterruptible power supplies (UPS), cooling systems, and power distribution units. We perfect our security interest in this "personal property" under the framework of the Uniform Commercial Code (UCC) by filing a UCC-1 Financing Statement. The scope of this security interest is extremely broad, even including the project's intellectual property and bank accounts.

• Assignment of Leases and Rents: This is a critically important component. We require the borrower to assign all of its lease agreements with tenants, and the rental income thereunder, to us as collateral. This means that upon a default, we have the right to bypass the borrower and collect rent directly from the tenants. We notify the tenants and execute a "Subordination, Non-Disturbance and Attornment Agreement (SNDA)." [SNDA Agreement] This tri-party agreement is the key legal instrument for locking down AI data center cash flows. Subordination means the tenant's leasehold rights are junior to our mortgage lien. Non-Disturbance is our promise to the tenant that, as long as they are not in default, we will not disturb their tenancy even if we foreclose on the property. Attornment is the tenant's promise that, should the landlord default, the tenant will recognize us (the creditor) as their new landlord and continue to pay rent. For high-value AI tenants, an SNDA ensures the stability of the lease, preventing any interruption to our core source of repayment, regardless of what happens to the property owner.

• Pledge of Equity Interests: We require the project's direct owner (typically a holding company set up by the sponsor for this project) to pledge 100% of its equity interests in the project company (the SPV) to us. This arrangement provides an alternative, and often faster, remedy to a real estate foreclosure. In a default scenario, we can enforce the equity pledge to quickly gain control of the project company, thereby taking over the project as a whole and avoiding the time-consuming and procedurally complex traditional foreclosure process.

5.2 Covenants: Setting the "Guardrails" for Behavior

Covenants are a series of promises and restrictions in the credit agreement that the borrower must adhere to. Their purpose is to maintain the project's risk profile at a level no worse than when we originally underwrote the loan.

Positive Covenants stipulate what the borrower "must do," such as:

• Pay property taxes and insurance premiums on time.

• Maintain and operate the data center in accordance with the highest industry standards.

• Provide us with detailed financial statements and operating reports on a regular basis (typically quarterly).

• Comply with all applicable laws and regulations.

Negative Covenants stipulate what the borrower "must not do without our prior written consent." These are our core tools for controlling risk. For an AI data center, common negative covenants include:

• Debt Restrictions: Prohibiting the borrower (SPV) from incurring any additional indebtedness.

• Lien Restrictions: Prohibiting the creation of any new liens on our collateral that could rank senior to our claim.

• Asset Disposition Restrictions: Prohibiting the sale, transfer, or disposal of any material assets.

• Change of Business Restrictions: Prohibiting a change in the use of the property from a data center.

• Material Contract Amendment Restrictions: Prohibiting the amendment or termination of key commercial contracts, especially the lease agreements with major AI tenants and the Power Purchase Agreement (PPA).

• Affiliate Transaction Restrictions: Strictly limiting non-arm's-length transactions between the borrower and the sponsor or other affiliates to prevent value from being siphoned out of the project company.

Financial Covenants are quantitative performance requirements that typically come into effect after the project reaches stabilization and serve as an early warning system.

• Minimum DSCR Requirement: Requiring the project's DSCR at the end of each quarter not to fall below a pre-set level (e.g., 1.25x). A breach of this level, even if the borrower is still making payments on time, constitutes a technical default, giving us the right to intervene.

• Cash Sweep Mechanism: If the DSCR falls below a higher threshold (e.g., 1.40x) but remains above the minimum requirement, or if other specific "trigger events" occur, the agreement may require all "excess cash flow"—that is, cash remaining after covering operating expenses and debt service—to be deposited into a lockbox account controlled by us. These funds are then used to prepay the loan principal, thereby rapidly reducing our risk exposure.

5.3 Intercreditor Agreement

If the capital structure involves multiple tiers of creditors (e.g., us as the senior lender and another institution providing mezzanine debt), then an Intercreditor Agreement is essential. This agreement clearly defines the rights and obligations of each party regarding payment waterfalls, voting rights, and post-default enforcement actions (e.g., who has the right to initiate a foreclosure). It prevents creditors from becoming deadlocked due to conflicting interests when problems arise, which would harm the interests of all creditors.

The drafting and negotiation of a credit agreement is a nuanced process filled with strategic considerations. As an investment team, we must work closely with our lawyers to ensure every clause is clear, unambiguous, and capable of providing the most robust legal protection for our capital amidst future uncertainties. This document will ultimately serve as the ballast that allows us to navigate steadily through volatile markets.

Chapter 6: Post-Closing Management and Exit Strategy – From Safeguarding Value to Realizing Returns

The funding of a loan is not the end of the investment process, but merely the beginning of long-term, active post-closing management. Throughout the life of the loan, our primary mission is to continuously monitor the project's operational and financial health, promptly identify and respond to potential risks, and ensure the safety of our credit asset. Concurrently, we must clearly plan and execute an exit strategy to ultimately realize our investment returns.

6.1 Proactive Monitoring: The "Radar System" for Risk

Our asset management team acts as a "risk radar," employing a series of mechanisms to continuously monitor every project in our portfolio across multiple dimensions.

Construction Monitoring is of paramount importance for greenfield projects. We regularly receive and review construction progress reports and budget draw reports from our independent engineering consultant. Each time the borrower requests a new construction draw, we verify that the requested amount matches the work completed and that all payments made to date are properly documented. We conduct periodic Site Visitsto personally confirm the progress and quality of the work. Any significant schedule delays or cost overruns will trigger our early warning system and require the sponsor to provide an immediate solution (which usually involves injecting additional equity).

Operational Monitoring focuses on financial and commercial performance. On a monthly or quarterly basis, we review the financial reporting package submitted by the borrower, comparing it against the projections from our original underwriting model. We closely track key operational metrics such as Power Usage Effectiveness (PUE), occupancy rates, and rental income. We hold regular post-closing management meetings with the sponsor's management team to discuss the project's operational status, market dynamics, and any challenges it faces.

Covenant Compliance Testing is a routine part of post-closing management. At the end of each reporting period, we calculate the financial covenant metrics (like DSCR) and check whether the borrower is in compliance with all positive and negative covenants. For any covenant breach, regardless of its severity, we will issue a formal notice demanding that the borrower cure the default within a specified period. This strict enforcement discipline is key to maintaining our credibility and control as a lender.

Tenant Risk Monitoring is equally vital. We continuously track the public financial reports, credit rating changes, and news flow of our major tenants. Any sign of deterioration in a tenant's financial health will prompt us to reassess the project's cash flow risk and may trigger stricter cash management measures (such as activating a cash sweep).

6.2 Workouts and Restructuring

When a risk warning becomes a reality and a material default occurs, our asset management and legal teams immediately intervene to initiate the workout process. Our objective is always to maximize the recovery of our capital.

The first step is typically to issue a "Reservation of Rights Letter," formally notifying the borrower of its default while stating that we reserve all legal remedies available to us as we consider a path forward.

Next, we may negotiate a "Forbearance Agreement" with the borrower. Under this agreement, we agree to temporarily refrain from exercising our default remedies (like foreclosure) for a specified period (e.g., 3-6 months) in exchange for the borrower taking a series of corrective actions, such as hiring a restructuring advisor approved by us, injecting new equity, or accelerating marketing efforts to find new tenants. This agreement usually comes with more stringent reporting requirements and tighter default triggers.

If negotiation and restructuring efforts fail, we will be forced to take the ultimate legal step: enforcing our collateral. As previously discussed, we typically have two options: the traditional, lengthier real estate foreclosure process, or the often faster and more efficient equity pledge foreclosure via a UCC sale. In either case, our goal is to gain full control of the asset. We would then, either through our own operational capabilities or by hiring a professional asset manager, stabilize the asset, improve its operations, and sell it when market conditions are favorable to maximize its value. The value of an experienced private credit team is demonstrated not only by its accurate pre-investment judgment but also by its ability to calmly and professionally manage complex situations and minimize losses when problems arise post-closing.

6.3 Exit Strategy: Completing the Capital Cycle

Our funds are closed-end vehicles with defined terms, so every investment must have a clear exit path.

Refinancing is the most common exit. Once a project is completed and achieves stabilized operations, its risk profile decreases significantly, making it attractive for lower-cost, long-term financing from sources like commercial banks, insurance companies, or the Commercial Mortgage-Backed Securities (CMBS) market. The sponsor will use the new, cheaper loan to pay off our relatively more expensive construction or transitional loan. Our loan agreements typically include a prepayment protection period (e.g., 2-3 years) during which early repayment incurs a penalty, ensuring our investment achieves a minimum holding period return.

Asset Sale is the second common exit path. The sponsor may sell the stabilized data center to a core investor seeking long-term, stable cash flows, such as a Real Estate Investment Trust (REIT), an infrastructure fund, or a large pension fund. As the senior lender, our loan will be repaid in full at the closing of the asset sale.

Repayment at Maturity is the final option. If neither of the above scenarios occurs, the borrower is required to repay all outstanding principal and accrued interest on the loan's maturity date as stipulated in the contract.

A successful exit depends on our accurate judgment of market liquidity and the future capital environment at the time of underwriting. We must ensure that the asset we are financing will remain attractive to the next round of capital providers, whether debt or equity, when our loan is due to be repaid.

Chapter 7: Conclusion – Navigating with Prudence at the Intersection of Compute and Capital

We stand at the dawn of an era driven by artificial intelligence, an era characterized by an explosive growth in the demand for compute infrastructure. This transformation has opened up an unprecedented, historic market opportunity for the private credit industry. Data centers, as the physical entities that host this revolution, are a natural fit for the investment appetite of private credit due to their asset-heavy nature, high capital consumption, and their linkage to long-term, high-quality contracted cash flows. As capital providers in this space, we have the opportunity to not only generate superior risk-adjusted returns for our investors but also to participate deeply in and promote the construction of the digital economy's foundation.

However, immense opportunity is often accompanied by latent risks, masked by the euphoria of the market. As we have repeatedly emphasized in our analysis, signs of an overheating market are emerging—such as the loosening of credit standards, an over-reliance on Payment-in-Kind (PIK) structures, and spread compression resulting from too much capital chasing a limited number of quality projects. History has proven time and again that the seeds of future risk are sown during any asset class's period of rapid expansion.

Against this backdrop, a return to the first principles of credit—prudent risk management, rigorous due diligence, sophisticated structuring, and proactive post-closing monitoring—is more important than ever. The comprehensive investment framework systematically articulated in this paper, from qualitative analysis to quantitative modeling, from legal architecture to operational oversight, is our core methodology for navigating through cycles and finding certainty in the midst of uncertainty.

Looking ahead, we foresee several trends that will profoundly impact data center credit investing:

• The Power Challenge: The immense energy consumption of data centers is subjecting them to increasing social and environmental pressure, while also creating physical supply bottlenecks. In the future, credit deployment will increasingly favor projects that can secure long-term, renewable power supplies. The due diligence on grid stability and energy contracts will become more critical than ever before.

• Technological Iteration: The emergence of new technologies like liquid cooling and the continuous evolution of computing architectures are placing new demands on data center design. As long-term creditors, we must be forward-looking and able to assess the technological "moat" of our collateral, avoiding investments in assets that risk technical obsolescence within the term of our loan.

• Supply Chain Risk: From generators to switches, data center construction relies on a global, complex supply chain. Geopolitical tensions, trade barriers, and other factors could lead to delays in the delivery of critical equipment and cost increases, posing greater challenges for the risk management of construction loans.

Ultimately, the union of artificial intelligence infrastructure and private credit is a profoundly symbiotic relationship. The former provides the "digital engine" that will drive future economic growth, while the latter supplies its indispensable "financial fuel." On this path, filled with both opportunity and challenge, the ultimate victors will not be the speculators swept up in market frenzy, but rather the long-term value investors who adhere to credit discipline, respect fundamentals, maintain a healthy reverence for risk, and pursue excellence in every detail. Our mission is to compose our own prudent and solid chapter in this grand symphony of compute and capital.