What a 100 MW Load Request Asks of the Grid

A large interconnection request reads, on paper, as one line item. Inside the planning process it behaves like a new municipality appearing on the system, with obligations for accredited capacity, deliverability, and multi-year transmission work attached.

A load request is a resource-adequacy event

When a prospective data center asks a utility for 100 megawatts of firm service, the request is easy to state and difficult to satisfy. It is useful to size the number before analyzing it. Roughly 1 gigawatt of continuous demand corresponds to about 700,000 Wisconsin homes, and about 500 megawatts of continuous load approximates a city the size of Milwaukee. A request in the hundreds of megawatts is therefore not a large commercial account in any conventional sense. It is a municipal-scale load arriving at a single interconnection point, and single-site peaks in this class range from a few hundred megawatts to well over a thousand.

The moment such a request enters planning, it stops being a customer question and becomes a resource-adequacy question for the whole balancing area. That reframing is the point most non-specialist discussion misses.

Capacity is not deliverability

The first instinct is to ask whether the region has enough generation, but that is the wrong stopping point. Adequate generation is only half the battle; delivering power to a specific site is a separate and often binding constraint. A region can hold sufficient nameplate capacity in aggregate and still be unable to serve a particular 100 MW node at peak because the wires between the resources and that node are insufficient.

This is why deliverability, not headline generation totals, governs whether a large request can actually be met. The physical path matters as much as the megawatts at the far end of it.

Accreditation: capacity only counts if it can reach the load

Wholesale planning encodes this directly through accreditation. MISO grants capacity credit only to resources that pass a deliverability test: a demonstration that their output can reach load during peak conditions. Where new transmission is required for a resource to be deliverable, its accredited capacity may be reduced until that transmission is in service.

The consequence for a large load request is concrete. A load-serving entity must carry accredited capacity at least equal to its peak demand plus a reserve, so adding a municipal-scale load raises the amount of deliverable capacity the entity must secure, not merely the amount of generation that exists somewhere on the system. Capacity that cannot reach the load does not count toward meeting it.

The planning reserve margin and the obligation a single request creates

Every load-serving entity must maintain a Planning Reserve Margin: accredited capacity equal to peak load plus a reserve buffer. A single 100 MW request pushes that obligation upward and does so against a system already under pressure. Regional planning anticipates that data-center and electrification load growth through the mid-2030s will exacerbate an existing capacity shortfall, and that many new loads will require additional firm, controllable resources rather than energy alone.

That phrase — firm and controllable — is doing real work. It signals that intermittent additions do not, by themselves, discharge the obligation a large new load creates; the system must be able to count on the capacity at peak.

Procuring firm capacity: what the request actually sets in motion

The obligation is not abstract; it shows up as physical resource plans. In one Wisconsin utility plan reflected in the Brief, roughly 3 gigawatts of new capacity intended for data-center and growth needs breaks down as approximately 1,120 MW of solar, 180 MW of battery storage, and 1,400 MW of natural gas, the latter as two plants of about 700 MW each. The presence of fast-start gas in that mix is not incidental. Utilities are leaning on dispatchable gas to guarantee round-the-clock capacity for large loads, precisely because a firm 100 MW commitment must be met at peak regardless of weather.

The scale of aggregated requests explains the magnitude. Two proposed data centers alone would roughly double one utility’s peak demand by 2030, from about 5 GW toward 10 GW. Each individual 100 MW request is a fraction of that, but each one carries the same structural demand: capacity that is firm, deliverable, and accredited.

Tariff design is beginning to reflect this. A proposed data-center tariff described in the Brief includes a 100 percent option, under which the data center funds a full dedicated resource and receives its output, and a 75 percent capacity-only option, under which it funds most of a new plant that counts toward system reliability. In both structures, day-to-day energy still comes from the grid mix at market rates; the dedicated plant functions as a reliability and deliverability backstop rather than an off-grid island. The mechanism is a direct answer to the question a single large request poses: who procures, and who pays for, the firm capacity it requires.

The transmission and substation tail

Meeting the request typically triggers physical construction with long, largely irreducible timelines. A new substation generally takes on the order of 18 to 24 months. A new high-voltage transmission line, once routing, permitting, and local opposition are accounted for, more commonly takes 5 to 10 years. In at least one Wisconsin case within the Brief’s scope, the regional transmission operator proposed new high-voltage lines and substation upgrades to serve concentrated new load, and one planned campus substation would be roughly twice the size of the largest existing substation in the state.

A distinction matters here. Large load additions generally move through utility planning rather than the generator interconnection queue. That queue is itself constrained, with entry-to-agreement timelines of several years and repeated schedule extensions. But the transmission required to deliver a load still needs regulatory approval and multi-state coordination, so the load path is faster than the generation path without being fast.

Why large-load screening is emerging

Because a single request can reshape a regional plan, some markets have begun to formalize how they absorb one. ERCOT (Electric Reliability Council of Texas) requires a mandatory screening study for loads above 75 megawatts, and comparable approaches exist elsewhere. At the federal level, regulators have been asked how to expedite generation and transmission for very large increments of new load. The direction of travel is toward treating large-load interconnection as its own planning category rather than an oversized version of an ordinary service request.

What this means for planners

A 100 MW request should be read as three obligations bundled into one number: accredited capacity to procure, a reserve margin to maintain, and a deliverability path to build. Each moves on a different clock, and the slowest clock — transmission — sets the true availability date.

As individual requests approach municipal scale, the open institutional question is where the screening threshold should sit and what a load-serving entity should be required to demonstrate before a single interconnection is allowed to reset a regional adequacy plan. What standard of deliverability should a request of this size have to meet before it is counted as served?

Source: the Wisconsin AI Infrastructure Readiness Brief. Primary references include MISO planning and accreditation rules, regional transmission planning, and Public Service Commission proceedings.