Load Ramp Schedules: How to Phase Power for a Data Center Campus

A data center campus does not switch on all at once. Buildings are delivered one at a time, data halls inside each building are fitted out in stages, and IT equipment is installed as tenants or internal teams need it. The power the campus draws from the grid grows in steps over several years.

A ramp schedule puts numbers on those steps. At minimum it gives peak demand in MW (and usually MVA) for each year of service. More detailed versions break the first years into quarters or months, show which building or phase drives each step, and separate the load the developer will build from the load it expects tenants to consume.

The figures are illustrative. Real ramps are uneven, and the gaps between steps depend on construction pace, equipment lead times and how fast the IT load fills each hall. Our guide to how much power a data center needs explains how to turn IT load and PUE into the facility demand that each step represents.

A utility has to have the capacity in place before each step arrives. The ramp tells it when transformers, substations, lines and network upgrades must be in service, and which of them can wait. It also tells the utility’s resource planners how much generation or capacity it needs to secure in each year.

• Study scope. Load studies test the system at specific demand levels in specific years. A different ramp is a different study.
• Upgrade sequencing. An existing substation might carry the first 30–75 MW while a new transmission source is built for the rest.
• Resource adequacy. Utilities and grid operators count future load when they plan capacity. A large ramp can move a utility’s whole resource plan.
• Cost exposure. The utility’s spending on the customer’s behalf rises with the ramp, and so does the security it will ask for.

This is why a 300 MW request with a credible five-year ramp is often easier to serve than a 300 MW request that wants everything in year two. The total is the same. The first one gives the utility time to build.

A credible ramp starts from the physical build plan and works outward to the meter. It should hold up when a utility engineer asks what is happening on site in each year.

1. 01Define the phases: buildings, data halls and any non-IT loads, with target construction completion dates.
2. 02Estimate how fast IT load fills each hall once it is complete. Halls rarely reach design load on the day they open.
3. 03Convert IT load to facility demand using the design PUE for peak conditions.
4. 04Add construction power, central plant and campus loads that arrive ahead of the IT load.
5. 05Check the result against equipment lead times, especially large power transformers and high-voltage breakers.
6. 06Decide what the campus can do if the grid is late: delay a phase, use bridge power, or accept a lower firm level for a period.

Keep two versions. One is the internal plan with its uncertainty showing. The other is the request submitted to the utility, which should be the number the project is prepared to commit to.

Utilities have learned that large-load forecasts are often optimistic. Many now tie the ramp to financial commitments in the electric service agreement. The usual pattern is that contract demand for each year follows the ramp, and a minimum bill charges for a share of that contracted demand whether or not the campus uses it.

That changes how a developer should think about the ramp. Overstating early demand is no longer a free option. It can mean paying for capacity that sits idle, posting collateral against load that does not exist yet, and facing exit charges if the project scales back. Our guide to large-load tariffs and electric service agreements explains these terms in more detail.

Some agreements allow the ramp to be adjusted within limits, deferred for a fee, or reduced with notice. Those flexibility terms vary widely by utility and are worth negotiating with the same care as the ramp itself.

Phasing is the main tool for matching a campus to what the grid can deliver on a given date. A few approaches come up often.

Whether a given approach is available depends on the utility, the regional market rules and the site’s own physical layout. A site with room for a future second substation and a generation yard keeps more of these options open. The bridge power option is covered in our guide to on-site generation and natural gas.

Ramps change. Tenants shift, designs move to higher densities, and a campus planned for 200 MW becomes a campus planned for 500 MW. Utilities generally treat a material increase as a new request, or at least a restudy, and some treat a significant acceleration the same way. A decrease can trigger contract adjustments or charges.

For site screening, the question is not just whether a location can reach the full-build number. It is whether it can deliver the first meaningful block of power on the date the project needs it, and then keep pace. When we screen a site, we look for the first 25–75 MW and the path to the rest as two separate questions.

• Is there existing substation or line capacity that could carry the first phase?
• What upgrades does each later step depend on, and how long do they usually take in that area?
• Does the land layout allow the substation, second source and any bridge generation to be added without redesign?
• Would a slower ramp materially change the date of first power?