Business strategy for internet infrastructure and data centres
The data centre industry faces a fundamental disconnect between soaring AI compute demand and the physical limitations of power grids and cooling capacity. Cogliva synchronises these infrastructure constraints with commercial targets to build a resilient and actionable data centre strategy.
Industry snapshot
The internet infrastructure sector is currently defined by a massive bifurcation between legacy corporate data centres and high-density AI factories. Large-scale colocation providers and hyper-scalers dominate the landscape, owning the primary physical hubs that route the majority of global traffic. The structure is capital intensive, with multi-year development cycles and high barriers to entry based on land and power entitlements. Ownership is increasingly consolidated under massive real estate investment trusts and infrastructure funds seeking stable, long-term yields from digital growth.
Profitability in this sector is driven by the efficient conversion of megawatts into billable revenue. Margin is made through high occupancy rates, operational excellence in cooling, and long-term power purchase agreements that hedge against energy volatility. Margin is lost when capacity remains stranded due to poor floor layout or when unexpected cooling requirements exceed the facility's design limits. Rising construction costs and interest rates are currently squeezing internal rates of return on new builds, making asset light operations more attractive.
The current period is marked by the transition to AI-first infrastructure where traditional data centre designs are becoming obsolete before they are completed. Rack densities are leaping from 10kW to 100kW, necessitating a fundamental redesign of electrical distribution and mechanical cooling systems. Strategy is no longer just about floor space; it is about securing energy and water rights in a world where these resources are increasingly contested. The winners are those who can innovate in power generation and thermal reuse to satisfy both compute demand and environmental regulation.
Strategic pressures in this sector
The forces most likely to invalidate assumptions in a plan written last year.
Grid saturation and power scarcity
Electrical grids in major hubs are reaching their physical limits, forcing operators to look at secondary markets or on-site modular nuclear and hydrogen power.
Thermal management evolution
The rapid adoption of high-performance computing requires a shift from traditional air cooling to liquid-to-chip or immersion cooling systems.
Regulatory and sovereignty oversight
National governments are increasing scrutiny on data residency and land use, making the permitting process longer and more politically complex.
Specialised labour shortages激增
There is a chronic shortage of skilled mechanical and electrical engineers capable of managing next-generation high-density power systems.
Institutional sustainability mandates
Investors are demanding transparent reporting on water consumption and carbon intensity, shifting focus from pure profit to sustainable operations.
Long-Lead infrastructure delays
Supply chain delays for high-voltage switchgear and backup generators are extending the time between site commission and revenue generation.
What good strategy looks like in this sector
Power-First site selection
Strategy must be built from the ground up based on available megawatts and utility constraints rather than hypothetical demand forecasts.
Modular architectural flexibility
Operators should move away from monolithic builds toward modular designs that allow for liquid cooling retrofits as compute density increases.
Energy Self-Sufficiency integration
Executives must secure long-term energy supply through direct investment in renewable projects or microgrid technology to mitigate grid reliance.
Geographic diversification planning
Maintaining a diverse portfolio of Tier 1 hubs and Tier 2 edge sites ensures resilience against local regulatory shifts and regional power failures.
How the model is changing
Hyper-scale bespoke provisioning
Operators are shifting from standard wholesale leasing to hyper-scale custom builds where the tenant provides the hardware and the provider manages high-density cooling and power delivery. These contracts focus on long-term capacity reservations rather than square footage.
GPU-as-a-Service infrastructure
The industry is moving toward high-performance computing sites that trade low-latency network connectivity for access to renewable power sources and advanced liquid cooling. These sites specialise in AI model training where physical location is secondary to thermal efficiency.
Distributed edge computing
Edge computing providers are deploying micro-sites and modular facilities closer to urban centres to support real-time data processing for autonomous systems. Revenue models are shifting toward low-latency performance tiers and localized caching services.
Energy re-sale and thermal export
Traditional REITs are expanding into managed energy services where they act as microgrid operators and heat exporters to local municipalities. This model turns waste heat into a secondary revenue stream while reducing the net cost of cooling operations.
Signals worth monitoring
- National grit capacity alert level changes
- Wholesale power purchase agreement price volatility
- Hyper-scaler quarterly capital expenditure guidance
- Planning permission rejection rates in primary hubs
- Server hardware thermal design power trends
- Carbon credit spot price fluctuations
Typical challenges and the workflow that addresses them
| Challenge | How the workflow handles it |
|---|---|
| Regional power constraints are making it impossible to scale our footprint at the pace required by our Tier 1 tenants. | Use the Cogliva strategy diagnostic to evaluate geographic risks and identify alternative sites where the Management Copilot can model power availability versus construction timelines. |
| We are struggling to bridge the gap between our ESG decarbonisation targets and the massive energy demands of our new AI clusters. | The organisation context module captures environmental constraints while the Strategy Workbench allows executives to design a low-carbon transition plan with specific investment milestones. |
| The shift from air to liquid cooling requires a complete overhaul of our capital expenditure plans across twenty different locations. | The tactical plan feature decomposes site-specific mechanical upgrades into actionable workstreams with assigned budgets and timelines to ensure consistent multi-site rollout. |
| Our long-term lease structures are being threatened by volatile energy markets and unpredictable utility pricing. | Establish strategic signals within Cogliva to monitor global energy indices and national grid capacity alerts for an early warning system on margin compression. |
| I feel like our strategy is buried in static spreadsheets that do not reflect the rapid changes in the silicon roadmap. | Centralise all infrastructure planning in the Strategy Workbench where Liva can synthesise hardware trends into a dynamic roadmap that stays current with market fluctuations. |
KPIs that hold the strategy together
Power Usage Effectiveness (PUE)
PUE remains the primary metric for operational efficiency by measuring how much energy reaches the compute equipment versus the cooling infrastructure.
EBITDA per Megawatt
This financial metric tracks the profitability of power allocation and ensures that highly constrained energy resources are sold to the highest-margin tenants.
Design PUE vs Actual PUE gap
The variance between architectural intent and operational reality identifies systemic maintenance issues or sub-optimal facility management.
Stranded Capacity Ratio
This measures the amount of provisioned power or cooling that cannot be used due to physical layout constraints or unbalanced loading.
Carbon-Free Energy (CFE) Score
The percentage of operational power matched with local renewable generation is now critical for maintaining investment-grade ESG ratings.
Frequently asked
What is a data centre strategy?
A data centre strategy is a long-term plan that aligns physical infrastructure investments with digital demand and energy availability. It dictates where to build, how to manage power and cooling, and which customer segments to prioritise to ensure profitable growth in a volatile utility environment. Successful strategies must balance capital intensive builds with rapid technological shifts in server density.
How does data sovereignty affect site selection?
Sovereignty is addressed by localising data storage and processing within specific national jurisdictions to comply with evolving privacy laws. Operators must design flexible facility architectures that allow for physical and logical separation of data based on the legal residency of the information. This increases the complexity of multi-regional site selection and operations management.
Do we need a diagnostic before building the strategy?
It is the recommended starting point when the direction is contested, implicit or inherited. If your strategy is already settled, you can start in the Strategy Workbench and use the diagnostic later as a health check.
Put this into a strategy your team can run
Start with a diagnostic of your organisation, turn the findings into a business strategy, and keep it live with tactical plans and signals.