How to Evaluate Wind Turbine Manufacturers: A 7-Step Sourcing Checklist

I'm a procurement lead at a renewables project developer. Over the past 11 years I've run roughly 40 accelerated wind turbine sourcing cycles — some planned, some not. About a third were emergency re-sourcing jobs after a supplier missed a delivery slot, changed specifications mid-contract, or simply stopped returning calls.

This checklist is what I use when I'm evaluating wind turbine manufacturers from scratch, and it's the same list I hand to junior buyers on my team. It's written for project developers, utilities, EPC contractors, and distributors — anyone signing a supply agreement for a multi-megawatt wind turbine, or a fleet of them.

Seven steps, plus a short section on what usually goes wrong. One caveat before we start: my experience skews toward onshore projects in Europe and North America, with a few North Sea offshore jobs mixed in. If you're sourcing in Asia-Pacific or Latin America, the vendor landscape is different and some of the reference-checking logistics won't translate one-to-one.

The 7-Step Checklist

Step 1: Verify the manufacturing footprint, not the brochure

Ask which factory actually assembles the nacelle and which one makes the blades. In-house, or a third party? What's the annual output of that specific line? If they can't answer in five minutes, that's a data point.

Then ask for a third-party factory audit — DNV, UL Solutions, or TÜV — completed within the last 24 months. Not a self-declared quality certificate. An audit that names the site and the date.

From the outside, a polished technical brochure and a big booth at WindEurope look like manufacturing capacity. The reality is that a fair number of "manufacturers" are really integrators that outsource blades, castings, or even full nacelle assembly. Nothing wrong with that model. But you should know which one you're buying from, because your lead times and your quality escalation path depend on it.

Step 2: Check the certification stack, model by model

The baseline is a type certificate and design assessment from a recognized certification body, aligned with IEC 61400-1:2019 for onshore or IEC 61400-3-1:2019 for offshore. Component certificates for gearbox, generator, blades, and tower are separate documents — each with its own revision history.

The part people miss: certificates are model- and revision-specific. A type certificate issued for a 4.2 MW platform in 2018 doesn't automatically cover the 4.5 MW variant launched in 2023, even if the marketing name is unchanged. If you're comparing different models from the same wind turbine supplier — say a Vestas V164 offshore platform versus a V236-15.0 MW — you're looking at two separate certification files. Ask for both.

ISO 9001 is table stakes. It tells you nothing about whether the turbine will survive 20 years at your site's wind regime and turbulence intensity.

Step 3: Get five references you can actually phone

Not a logo wall. Five names, five site locations, five phone numbers. Same model, hub height within about 20 meters of your project, commissioned and operating for at least 24 months.

Why 24 months? Because the faults that matter tend to surface in the first two years. Gearbox issues, blade leading edge erosion, pitch system faults — these don't show up in the first six months of operation.

Of the five references I ask for, I usually get two who will talk openly. Those two are worth more than the other three combined.

Step 4: Look at model-level performance, not company reputation

The most frustrating part of turbine evaluation: the same OEM can have one model with strong availability figures and another with a chronic problem. Company-level reputation hides that.

Ask for availability data, capacity factor distributions, and main component failure rates for the specific model. If they won't share it, ask for the third-party energy yield reports from recent projects using that model. If they won't share those either, you know what you're dealing with.

Granted, some of this data is commercially sensitive. There's a difference, though, between "we can't share the raw data" and "we don't track it."

Step 5: Verify delivery slots — the real ones

Ask which factory, which quarter, and which manufacturing slot. Get it in writing. Nacelle slots and blade slots run on separate schedules, and the slower one sets your real delivery date.

In early 2024 I had to re-source a mid-size onshore order because the original OEM's blade supplier slipped by two quarters. By the time we found out, the only free slots in the region were for a different platform — which meant a layout re-permit and a new grid compliance study. The re-sourcing itself cost well into six figures, before we counted the schedule damage with the offtaker.

That's why this step matters more than it looks. Slots fill up. A "we can deliver in Q3" answer in Q1 can quietly become "Q2 of next year" by the time you've finished negotiating the contract.

Step 6: Service footprint and spare parts, over 20+ years

Where are the local service teams based? How fast can a main component exchange — a gearbox swap, a blade replacement — actually happen in your region? What's the crane mobilization lead time?

Then ask for spare parts pricing and lead times across the design life of the turbine (typically 20 to 25 years; sometimes 30 with an extension). Get this in the contract as a price book with an index formula, not as a verbal commitment. That single clause has saved me more arguments than any other.

Also worth asking: who reads your condition monitoring system (CMS) data? An in-house OEM team, a third party, or your own operations staff? Every answer has a different cost profile and a different escalation path.

Step 7: Contract terms — guarantees, LDs, and the exit

Four things to nail down before you sign:

• Availability warranty. Typically quoted in the 95–97% range, but the definition matters more than the number. Does it exclude grid outages? Planned maintenance? Curtailment for grid stability? Two contracts with "96% availability" can mean very different money over a decade.
• Power curve guarantee. Verified by a power curve test per IEC 61400-12-1, usually at your site, usually within the first 12–24 months of operation.
• Liquidated damages (LDs) for late delivery. With a cap that actually stings. A 5% cap on contract value doesn't motivate anyone on a two-year project.
• Performance bond and parent company guarantee (PCG). If the supply entity is a special purpose vehicle (SPV) and the parent won't sign a PCG, price that risk in.

And the exit clause. What happens if they don't perform? What triggers it, what's the notice period, and who keeps the tooling afterwards? I've seen supply agreements where the OEM retained the blade tooling post-termination. That's an 18-month problem hiding in plain sight.

What Usually Goes Wrong

Three mistakes show up again and again.

Comparing on $/MW alone. The lowest capex number is not the lowest cost. Sparsely available spare parts, slow crane mobilizations, and a lower-than-promised availability figure can wipe out a six-figure-per-MW saving inside two or three operating years. In my own tracking, the lowest quote has ended up costing more than the mid-range option in more than half the cases.

Trusting the OEM's own capacity factor figures. Use a third-party energy yield assessment — DNV, UL, or an independent consultancy. The number will usually come back lower, and the layout will usually need adjustment. Better to adjust the model than the wind farm.

Skipping the site visit. If the manufacturer won't let you walk the floor, ask why. Half the questions on this checklist answer themselves in the first twenty minutes of a factory tour.

Honestly, I'm not sure why some OEMs consistently hit their slot dates and others consistently miss. My best guess is it comes down to how much buffer they build into blade tooling schedules, and how honest their own procurement team is willing to be with their sales team. Whatever the cause, the ones that hit slots tend to be the ones whose factory you can actually visit.

Print this list. Bring it to the site visit. It's a short document, and most of it is just questions — which is the point.