TLDR
Reliability surveys do not show a dramatic rise in transformer failure rates. Vendors who claim otherwise are overstating, and technical audiences notice. The number that moved is consequence.
A failure in 2005 cost a replacement plus a few months of workaround. A failure today costs a replacement at sharply higher prices, plus a wait measured in years, plus the accelerated aging imposed on every neighboring unit carrying the load through that wait, plus whatever the interrupted load was worth. In a data center or a refinery, that last term dominates all the others. Same failure, same rate, far larger loss.
This is why the economics of prevention flipped without anyone's equipment getting worse. Prevention did not become cheaper. Failure became much more expensive. Any business case built on a claimed rise in failure rates is standing on the wrong number, and the right number is stronger because it survives scrutiny.
Two Camps. One big argument.
Ask the industry whether transformers fail more often than they did twenty years ago. You will get into an argument. One camp sells monitoring on the claim that failure rates are climbing. The other camp, the engineers who read the reliability data, know the claim does not hold up.
The most recent large study is CIGRE Technical Brochure 939, published in September 2024. It analyzes AC transformer reliability across a global fleet, and it does not show a dramatic rise in failure rates. A business case built on "transformers fail more now" starts on a number the audience can dispute. That is a weak place to stand.
There is a stronger number. It moved a great deal, and almost no one disputes it. The cost of a single failure.
Price the same event at two points in time.
A large power transformer fails in 2005. You order a replacement at 2005 prices. You wait a few months. You run a workaround and absorb the cost. The failure hurts, then it ends.
The same transformer fails this year. Four things happen, and each one is larger than its 2005 version.
The replacement costs far more. Wood Mackenzie put the rise in power transformer unit costs at 77 percent since 2019. A like-for-like unit is a different line item now.
The wait is measured in years. Lead times for large power transformers now run two to four years. The workaround is no longer a season. It becomes a multi-year operating condition.
The rest of the fleet pays. Every neighboring unit that picks up the transferred load runs hotter for the whole of that wait. Heat ages insulation. A two-to-four-year detour spends real life on assets that never failed.
The load itself has more at stake. A refinery train or a hyperscale compute cluster does not idle cheaply. Whatever the transformer was feeding stops, and that lost output often dwarfs the other three costs combined.
Same failure. Same rate. A far larger loss.
This is the point that reframes the whole prevention question. The math on prevention flipped, and no piece of equipment had to get worse for it to flip. Prevention did not get cheaper. Failure got much more expensive. The widening gap between the two is where the return on prevention now lives.
So what does a serious prevention case look like today? It prices the failure it avoids, in current dollars, current lead times, and current load value. That figure is large, and it holds up when a skeptical engineer checks it.
Prevention only pays if you can see the failure coming. Most transformer faults that end in replacement are mechanical or structural. Windings loosen. Clamping structures relax. Cores shift. These modes build over months and give off little or no gas, so an oil test can read clean until late in the process. Continuous condition intelligence reads the mechanical state of the unit directly and flags the change early, with time still on the table to plan. VIE builds that layer. The result is not a louder alarm. It is a multi-year, multi-million-dollar surprise turned into a scheduled intervention, with a replacement ordered early enough to matter.
Sell the avoided loss, not the fear of a rising failure rate. The avoided loss is documented, it is current, and it survives scrutiny from the exact audience worth convincing. That is the number worth building on.
References
- CIGRE Technical Brochure 939, "Analysis of AC Transformer Reliability," September 2024. https://www.e-cigre.org/publications/detail/939-analysis-of-ac-transformer-reliability.html
- Wood Mackenzie, "Transformer troubles: manufacturing and policy constraints hit US transformer supply" (power transformer unit costs up 77% since 2019). https://www.woodmac.com/news/opinion/transformer-troubles-manufacturing-and-policy-constraints-hit-us-transformer-supply/
- Lead times of 24 to 48 months for large power transformers: pv magazine USA, "U.S. transformer market faces severe supply constraints as lead times extend to four years," May 2026. https://pv-magazine-usa.com/2026/05/11/u-s-transformer-market-faces-severe-supply-constraints-as-lead-times-extend-to-four-years/