At a glance
| Buyer | Investment firm sourcing on behalf of a client |
| Original request | 7.5MW total, units operating in parallel |
| Required timeline | Delivery inside six months |
| Core constraint | Paralleled generators must be matched, and matched multi-megawatt sets are scarce |
| New equipment lead time | Well beyond the buyer’s window |
| Closest available path | A fleet of matched 2MW portables coming out of a data center refresh, arriving in phases |
The call
The buyer got straight to it. They needed seven and a half megawatts, and the units had to work in tandem.
Then they asked a question most people at that end of the market are too proud to ask: is that a big load?
It is not a small one. But the honest answer is more interesting than the size, because seven and a half megawatts is not really a hard number to fill. It is a hard number to fill in six months, in matched units, that will parallel. Every word in that sentence removes options, and by the end of it there are very few left.
This is a conversation we have several times a month, and the shape of it almost never changes. So it is worth walking through what actually happens when someone asks for multiple megawatts on a compressed schedule, and why the constraint is never the one buyers expect.
The ask, and the one thing the buyer got right immediately
The request was 7.5MW, with the units required to operate in parallel. The buyer said they did not care whether the machines matched, then immediately reasoned their way to the answer anyway: they assumed matching was probably required if the units had to parallel.
Right the first time, and worth crediting, because it is the assumption that determines everything downstream.
Why paralleled units have to match
Generators sharing a load are not just generators sitting next to each other. They have to synchronize, agreeing on voltage, frequency, and phase angle continuously, and they have to divide load proportionally as it moves.
Mixing ratings breaks that in practice even where it is technically possible. The buyer floated the obvious workarounds: seven smaller units, or five of one size and one of another. Both can be engineered. Neither should be. Every mismatch pushes complexity into the control system and the switchgear, and control complexity at this scale is not something you want to be inventing. It is cost, it is engineering time, and it is one more thing to troubleshoot at three in the morning during an actual outage.
If the units have to parallel, buy them matched. That is the whole rule.
Why 7.5MW is not a number you can actually buy
Once matching is non-negotiable, 7.5MW stops being a target and becomes arithmetic. Total capacity has to divide evenly into whatever unit size you can source.
Working from 2MW machines, the honest answer is four of them and 8MW. From 2.5MW machines, three and exactly 7.5MW, assuming three matched 2.5MW units exist somewhere on the same schedule. There is no combination that lands neatly on 7.5MW using commonly available matched equipment.
This surprises buyers constantly, and it should not. Load studies produce precise figures, and precise figures feel like specifications. At this scale they are the floor, not the spec. You round up to the nearest buildable configuration and you accept the headroom, which is rarely wasted anyway once motor starting loads and growth are accounted for.
Six months does not buy new equipment
New generation at this size is not close to a six month window.
Current market lead times for new units run roughly 50 to 78 weeks in the 1500kW to 3000kW class, and 90 to 110 weeks in the 3500kW to 4500kW class. Matched multi-unit orders sit at the long end, because you are not waiting for a machine, you are waiting for a production slot big enough to hold several.
Paralleling switchgear compounds it, and this is the part that catches people. Medium voltage switchgear commonly quotes in the 52 to 80 week range. A paralleled installation is not finished when the generators arrive. It is finished when the gear that makes them cooperate arrives, and that gear is frequently the longer pole in the schedule.
We do have new 2MW units on order landing at the end of Q1 or the beginning of Q2 next year, which is genuinely fast for new equipment in this market. It was still outside what this buyer could use, and saying so early is more useful to them than a quote they cannot act on.
The used market rarely holds matched multi-megawatt sets
The obvious next move is the secondary market, and this is where multi-megawatt buyers get optimistic.
Used inventory from 20kW to 1000kW is deep and constantly turning. Above roughly 1.5MW it thins out fast, and the specific requirement of several identical units, same manufacturer, same rating, same vintage, available on the same schedule, is a completely different animal from finding one good large machine.
Matched multi-megawatt sets exist for essentially one reason: a facility that installed them together decommissions them together. Single units trickle onto the market continuously. Matched sets arrive as events.
Where matched sets actually come from
Which is exactly what produced the closest thing to an answer on this call.
We bought a fleet of used 2MW portable generators out of a data center that is replacing its entire backup plant with larger machines. We know the units well, because we sold them to that data center in the first place. When the site decided to upgrade, we bought them back. Our sales engineer and our CEO drove out and inspected the fleet in person before we committed to it, which is not a courtesy, it is how you avoid buying a yard full of expensive problems.
Two things about that story matter to anyone chasing megawatts.
This is now the dominant source of matched large sets. Data centers are refreshing backup plants at pace as capacity requirements climb. That refresh cycle is the reason matched multi-megawatt equipment reaches the used market at all, and it is why the good sets are usually spoken for before they are ever listed.
They come out in phases. A live data center cannot pull its backup power all at once. Units release one at a time as each replacement is commissioned. One is already in our yard. The rest arrive as the swap progresses, on the site’s schedule rather than ours.
That phasing is why we would not commit to a six month delivery on the full quantity. We could have. It would have been a guess, and a guess on a multi-megawatt delivery date is not a quote, it is a liability. The buyer is better served by a hard maybe than a soft yes.
The smartest question the buyer asked
Partway through, the buyer reframed the problem entirely: could they take a portion now and contract for the rest as it released, out of the same set?
That is the right instinct, and most buyers never get there.
Buying in phases from a single decommissioned fleet is not the same as piecing a system together from whatever surfaces on the market. The units are already matched to each other by definition. They share manufacturer, rating, vintage, control platform, and service history. A buyer reserving against a known fleet is doing something structurally different from a buyer hoping the market coughs up three more of something.
It changes the sequencing too. Install and commission what you have, size the switchgear and the pad for the full build from day one, and add units into a system that was engineered to receive them. That is a very different project from retrofitting capacity into gear that was never scoped for it.
The tradeoff is real and it should be said out loud: partial capacity is partial capacity. If the load genuinely requires the full figure on day one, phasing does not solve the problem, it just distributes the disappointment. If the load can ramp, phasing is often the only path that fits a compressed schedule at all.
The question the buyer did not know to ask
Two of the units under discussion were 2011 and 2013 builds. We asked whether the end client needed to be emissions tier compliant.
The buyer’s response was to ask what that meant.
That is not a knock on them. It is one of the most common gaps in the market and one of the most expensive to discover late, and the fact that a sophisticated buyer sourcing seven figures of equipment had not encountered it says something about how rarely it gets explained properly.
The short version for large diesel gensets in the United States:
- Emergency standby engines get meaningful relief. Under EPA’s stationary compression ignition rules, engines certified for emergency use are held to standards that do not require add on aftertreatment. That is why an older large diesel can still be entirely legal and entirely appropriate for standby duty.
- Emergency status has limits attached. Operation outside actual emergencies is restricted, with maintenance and testing generally capped around 100 hours per year. Exercise cycles and load bank testing fit comfortably inside that. Running the machine as a routine power source does not.
- Non-emergency and prime power is a different regime. If the engine will carry load outside of outages, for peak shaving, for grid support, or as a primary source, the exemption stops applying and the compliance picture changes completely.
- State and local rules stack on top. Some air districts are considerably stricter than the federal baseline, and permitting is local.
For this buyer the answer was no, tier compliance did not matter, so build year stopped being a factor. A buyer who answers that question wrong at the sourcing stage finds out at permitting, which is the worst possible time to find out.
What “used” actually means at this scale
The portables in question sat in place at the data center for roughly a decade. The customer pulled the wheels off years ago and the machines never moved again. The trailer chassis underneath them were never designed to sit that long and are not reusable.
So the scope of work is not a wash and a coat of paint. The units come off the old chassis, which stay on site. They ship back to Colorado. They are serviced, load bank tested, and put through a full functional inspection. New chassis go underneath them.
That is worth understanding as a buyer, because it is where the real difference sits between a reconditioned unit and a used one. A machine bought sight unseen off a listing is a set of photographs and an hour meter reading. A machine that has been pulled, inspected, serviced, load banked, and re chassised is a known quantity with a documented history. At multiple megawatts, that difference is not a preference. It is the entire risk profile of the project.
What multi-megawatt buyers should take from this
- Decide early whether the units must parallel. That single answer constrains everything downstream: unit size, quantity, switchgear, budget, and schedule.
- Treat your load figure as a floor, not a specification. You will round up to a buildable matched configuration. Plan the headroom rather than resenting it.
- Price and schedule the switchgear alongside the generators. Paralleling gear routinely quotes longer than the machines it controls.
- Ask where matched sets come from before you go looking for one. They come from decommissioning events. If you are not positioned when one happens, you are waiting for the next one.
- Answer the emissions question at sourcing, not at permitting. Standby duty and prime duty live under different rules, and the answer determines which build years are even on the table.
- Consider phasing against a known fleet. Partial capacity from a matched set beats full capacity that does not exist yet, provided the load can ramp.
- A nameplate megawatt figure is the start of due diligence, not the end. Fuel type, voltage, configuration, physical location, and condition all determine whether a unit is actually a candidate.
- Start earlier than feels necessary. Every constraint in this article is a calendar constraint. Money does not fix a single one of them.
Frequently asked questions
Do generators have to be the same size to run in parallel? For any practical installation, yes. Paralleled generators must synchronize on voltage, frequency, and phase angle and share load proportionally. Mismatched ratings can be engineered around, but doing so drives significant additional complexity and cost into the control system and switchgear, and it adds failure modes at exactly the moment the system needs to be dependable. Matched units are the standard approach for a reason.
How do you get 7.5MW of generator capacity? In practice you do not buy exactly 7.5MW. You buy the nearest buildable matched configuration, such as four 2MW units for 8MW total or three 2.5MW units for 7.5MW if a matched set of that rating is available. Because paralleled units must match, total capacity has to divide evenly into whatever unit size you can actually source on your timeline.
What is the lead time on a new multi-megawatt generator? Current market lead times run roughly 50 to 78 weeks for new units in the 1500kW to 3000kW class and 90 to 110 weeks for the 3500kW to 4500kW class. Matched multi-unit orders trend toward the long end. Paralleling switchgear frequently quotes in the 52 to 80 week range and can be the longer pole in the schedule.
Where do used matched multi-megawatt generator sets come from? Almost entirely from decommissioning events. Facilities that installed several identical units together, most commonly data centers refreshing their backup plants for higher capacity, release them together. Single large units appear on the used market continuously, but matched sets only appear when a site retires one. They typically release in phases as each replacement is commissioned.
Can an older generator still meet emissions requirements? For emergency standby duty in the United States, often yes. EPA rules for stationary compression ignition engines certified as emergency units do not require add on aftertreatment, which is why older large diesels remain legal and appropriate for standby service. Operation is limited outside actual emergencies, with maintenance and testing generally capped around 100 hours per year. If the engine will run as prime power or for non emergency purposes, the exemption no longer applies and the compliance requirements change. State and local air rules can be stricter than the federal baseline.
Can I buy part of a generator set now and the rest later? Yes, and when the units come from a single decommissioned fleet it is often the smartest available path on a compressed schedule. The units are matched to each other by manufacturer, rating, vintage, and control platform, so later additions integrate cleanly. Size the switchgear and the site for the full build from the beginning. The limitation is straightforward: if the load requires full capacity from day one, phasing does not solve it.
Talk to us before you go looking
Generator Source is the largest buyer and seller of new and used industrial generators in North America, with inventory from 20kW to 4MW+ and branches in Brighton CO, Jacksonville FL, Pensacola FL, and Austin TX. We buy decommissioned fleets directly, which is how matched multi-megawatt sets reach the market at all, and it is why we frequently know a set is coming before it is available to anyone else.
If you are scoping several megawatts, the useful conversation is the one that happens before the deadline exists.
Browse used industrial generators or call to talk through your requirement.