Data center capacity has become one of the most sought-after resources in today’s digital economy. The rapid growth of cloud computing, artificial intelligence (AI), and data-intensive applications—particularly those driven by high-performance graphics processing units (GPUs)—has significantly increased demand for power, cooling, and physical infrastructure. As a result, a growing number of companies are turning to colocation arrangements as a faster and more cost-effective alternative to building and operating their own data centers.
In a typical colocation model, a company places its servers and hardware in a third-party data center facility that provides the underlying infrastructure, including power, cooling, physical security, and network connectivity. This allows organizations to retain control over their equipment while leveraging the scale, reliability, and connectivity of a specialized provider.
The agreements governing colocation arrangements are often complex and, if not carefully negotiated, can expose customers and providers to meaningful financial, operational, and legal risks. This blog post discusses some of the key issues in colocation arrangements.
1. Power
Power is one of the most critical and expensive components of a colocation arrangement, particularly with the rise of high-density workloads, such as AI and GPU deployments. Colocation agreements should clearly define allocated power to the relevant equipment (often referred to as “critical power”), how that power is delivered, power usage efficiency (PUE), and whether billing is based on reserved capacity or actual usage. The parties should also consider potential changes in billing during power disruptions or outages, as well as potential overage charges, derating practices, and any limitations on increasing density over time, as these factors can materially affect both cost and performance.
2. Connectivity
Connectivity is a key component of colocation arrangements. The agreements should address the availability and pricing for interconnect and cross-connect services, relevant Meet-Me Room (MMR) availability, and availability of and restrictions on carriers and network providers at the relevant data center. Over time, recurring cross-connect and interconnect fees can become a meaningful expense, particularly in highly interconnected environments.
3. Ready for Service Timelines
Provisioning timelines, often referred to as Ready for Service (RFS) dates, and consequences for failure to meet the provisioning timeline, either due to customer or service provider failure, are also important. A delay in the RFS may impair a customer’s operations and a provider’s revenue. Both parties typically incur significant expense in procuring equipment and services to run the data center (in the case of the service provider) and to fully utilize the colocation space by installing equipment and procuring remote hands and other ancillary services (in the case of the customer). Delays in RFS can disrupt downstream supply chain schedules and upstream commitments. Some of the potential structures that help align incentives with respect to RFS include tiered credits and termination rights for RFS delays.
4. Service Level Agreements
Service level agreements (SLAs) may define uptime guarantees, power availability, cooling standards, and response times, and remedies if those standards are not met, typically on a tiered basis. SLA negotiations in colocation agreements are uniquely technical in nature. The requirements of uptime and power availability depend on numerous technical and operational factors, such as the customer’s business model, servers that the customer plans to use in the colocation space, and technical features in the design of the colocation space (e.g., location and redundancy of monitoring systems), among others. Further, if performance or efficiency is expected to improve over time, the service standards could rise accordingly.
5. Pricing Structure and Escalators
Colocation pricing typically includes multiple elements that are usually broadly classified as monthly recurring fees (MRCs) and non-recurring fees (NRCs). MRCs may include charges such as base space charges, power and utility costs, connectivity fees, taxes, and ancillary services like remote hands. Annual escalators, whether fixed or tied to an index, should be considered.
6. Term Length and Renewal Provisions
Providers often seek multi-year commitments, sometimes extending up to a decade. These agreements may include automatic renewals with strict notice requirements and pricing that may reset at renewal. The parties should assess rights to renegotiate or exit at different phases of the relationship.
7. Scalability and Flexibility
Because infrastructure needs evolve, the parties should consider the feasibility of allowing for growth and contraction. This includes the ability to add power or space, as well as rights to adjacent capacity if applicable.
8. Security and Compliance
While most colocation facilities maintain strong physical security, contracts should clearly define access controls, monitoring protocols, and audit rights. They should also address compliance with relevant standards and allocate responsibility for maintaining those standards.
9. Liability and Indemnification
Colocation providers typically limit their liability, often capping damages and excluding consequential losses. The parties should carefully review the liability provisions and whether indemnification provisions adequately allocate risk, particularly with respect to property damage. Insurance and subrogation rights may be a critical component in the evaluation of risk with respect to damage to the colocation facility (for the service provider) and servers (for the customer). In addition to the customary risk allocation issues, the interplay between the limitation of liability, service level credits, and financial remedies for RFS delays are some of the most negotiated provisions in colocation agreements.
10. Disaster Recovery and Business Continuity
The colocation agreement should specify the level of redundancy in power, cooling, and network infrastructure, as well as the provider’s obligations during disruptions, including casualty and taking events. Force majeure provisions should be carefully reviewed for appropriateness and consistency, including in cases of prolonged force majeure.
11. Financial Stability (Provider and Customer)
An increasingly important issue is the financial health of the parties, particularly as customers deploy high-value GPUs and other specialized hardware. In the event of a party’s financial distress or bankruptcy, the other party may face risks related to access, potential liens, or delays in retrieving equipment. The parties should assess the importance of ownership rights, access continuity, and protection against creditor claims, such as waivers of liens and clear bailment language. Given the substantial value and scarcity of certain hardware, these risks are becoming more significant and warrant careful attention.
While much of the focus is typically on the provider’s financial health, customers should also consider how their own financial stability may be addressed in the agreement. Colocation providers often include protections such as credit checks, financial reporting requirements, security deposits, or the right to require prepayment if a customer’s credit profile deteriorates. In some cases, providers may have termination or suspension rights tied to perceived financial risk. These provisions can affect operational continuity, particularly for customers deploying critical infrastructure. As a result, customers should carefully review and, where possible, negotiate objective triggers, cure periods, and limitations on the provider’s ability to take action based on financial concerns.
12. Exit Strategy and Decommissioning
Exiting a colocation arrangement can be complex and costly if not properly planned. Agreements should address early termination rights and fees, timelines for equipment removal, requirements for data destruction, and the consequences (including associated fees) for holdover. A poorly defined exit process can lead to operational disruption and unexpected costs.
Next Steps
Colocation agreements are foundational to modern infrastructure strategy, particularly as power-intensive workloads and high-value hardware become ubiquitous. Evaluating these agreements through the lens of long-term flexibility, risk allocation, and operational resilience is essential to avoiding costly surprises.