Almost every wind project procurement failure I've seen didn't start with the turbine.
It started with the checklist.
I handle procurement for a mid-sized wind services company. Over the past 8 years, I've processed more than 200 emergency orders — mostly spare parts that nobody hoped to need, a few last-minute replacement requests on large projects. That's my lens. Not R&D, not sales. The person who gets the call when something breaks and the clock is already running.
I've watched developers lose real money on contracts that looked fine on paper. Not because the turbines were bad. Because the evaluation process was aimed at the wrong target.
Here's what I wish more buyers asked before they sign.
The question everyone asks — and why it's the wrong one
You're comparing turbines. You've got spec sheets open. You're looking at MW rating, rotor diameter, cut-in speed, $/kW. You ask the manufacturer: "What have you got?"
Fair. But that's the shallowest part of the procurement process, and most teams spend 80% of their energy there.
Last quarter I ran an emergency order for a client whose turbine was down for six days. The generation loss on that unit at current power prices was comfortably more than the cost of the part itself. The part? Around $18,000. The downtime calculus? A different conversation entirely.
See the problem?
They'd evaluated the turbine on MW rating. They hadn't evaluated the question that actually cost them money — how fast can a critical component reach this site?
Where the real problem lives
A wind turbine is a 25- to 30-year commitment. But most procurement documents put nearly all their weight on Day 0 and almost none on Day 3,000.
When I'm evaluating an OEM — Siemens Gamesa, Vestas, GE Vernova, Nordex, whoever — I don't care about the turbine's potential. I care about the operating reality. Three things specifically:
One: service radius. In a real failure, can the OEM get a technician to my site? Not "the region" — the actual road, the actual vessel window. I've seen developers pick a cheaper turbine and later discover the nearest service hub was 14 hours away, weather windows not included.
Two: spare parts strategy. Are critical components stocked in-region, or shipped from another continent? On offshore projects, waiting 10 weeks for a main bearing instead of 4 is the difference between a contained setback and a quarterly disaster.
Three: platform maturity. Prototypes are exciting. Mature platforms have traceable histories — known failure modes, actual operating hours, documented service procedures. The theoretical gains of a brand-new model often get eaten by the reality of the first projects.
That's the bit nobody wants to face: the newest turbine is rarely the cheapest turbine over the long haul.
What it actually costs when you get this wrong
Let me walk through a real calculation. Not theoretical — grounded in emergency orders we've processed.
Say your project is 50 MW with 15 turbines. Normal uptime is around 97%, annual production somewhere near 175 GWh. At $40/MWh, you're looking at roughly $7 million in revenue per year.
Now one turbine sits idle for 10 days instead of 4, because a component has to travel. Lost generation on that unit over six extra days: a few hundred MWh. At $40/MWh, that's not catastrophic on its own. Add expedited freight — $8,000 to $25,000 depending on the part and the route. Add emergency labor rates. Add whatever liquidated damages are sitting in your PPA.
Multiply that across 15 turbines over 25 years, in a market where every bidder is trying to shave service response times to win the deal. "Cheaper" starts to look very expensive.
And there's a second cost nobody puts in the spreadsheet: small projects get deprioritized in service queues.
When a large developer and an independent 50 MW project both need the same part, the OEM's dispatch team triages toward the bigger account. That's not malice — it's business. But if your procurement contract doesn't specify service priority, you're in line behind the utilities. I've watched it happen. It has nothing to do with how well your project was designed.
If you're on the smaller side, ask the OEM directly before you sign: Where does my project sit in your service priority process? If you can't get a straight answer, that's an answer.
So what should you actually look for?
Short version:
- Real service footprint in the region — not "we have a regional office," but specific locations and headcount you can verify
- Spare parts availability with quantified SLAs — committed days for critical components, not "commercially reasonable efforts"
- Platform operating history — fleet installations, years in service, published failure rates
- Contractual service priority — in writing, not verbally promised
- Rotor and power class that matches your site conditions — less glamorous, but verify it anyway
Take Siemens Gamesa, as one example. Their SG 3.4-132 platform has been in the field since around 2019, which gives you a service history you can actually check rather than a marketing number. Their offshore portfolio is similarly established at the MW scale the market is moving toward. But the point isn't which OEM wins. The point is: evaluate whether the platform has already solved the boring problems in your market, at your wind conditions.
One caveat worth stating plainly. My experience is mostly onshore, drawn from about 200 emergency procurement orders in the $50K to $2M range. If you're sourcing offshore in a new market, the timelines and risk profile are genuinely different. Don't port onshore assumptions onto an offshore plan.
And full disclosure — I once spent three weeks going back and forth between a new-entrant manufacturer with a cheaper $/kW contract and an established OEM that was 12% more expensive. On paper, the new entrant won. My gut said no. I'd been burned before by a deal that looked too good, so I went with the established OEM. In hindsight, that 12% bought me fewer sleepless nights than it bought me measurable savings. But I'm not 100% sure it would hold up on a different project with a different service map.
The spec sheet won't tell you when a turbine will fail. It won't tell you how long the fix will take. Your procurement checklist should.