AI training clusters don’t draw power like traditional servers. They surge. Synchronously. And the transformer sitting inside your PDU was probably never designed for it.
In November 2025, the Uptime Institute published research documenting what facilities engineers had been seeing on their monitoring dashboards: GPU racks in AI training clusters swing from 60–70kW to over 150 kW in seconds, with the “difference between the low and high points of power draw during training program execution” exceeding 100%. Every rack in the cluster does it at the same time, because every GPU is working on the same training job. Those power swings create two distinct problems for your transformer. The first is a capacity problem: whether the transformer can handle the total load at peak. The second is a power quality problem: whether it can handle the harmonic characteristics of that load. This article covers the second. For the capacity problem — including how to size transformers for correlated GPU burst loads — see our companion article, “AI Data Center Transformer Sizing: The Capacity Problem Behind the GPU Boom.”
That synchronized demand creates electrical conditions that standard K-13 transformers were never built to handle. And if you’re deploying AI infrastructure on K-13 rated PDU transformers, you’re running on equipment that could fail at critical moments with shortened lifespans.
Traditional data center loads are well-understood. A mix of servers, storage, and networking equipment draws relatively steady power with moderate harmonic content. Schneider Electric’s White Paper #38 measured real-world data center branch circuits and found that modern server power supplies with active power factor correction (PFC) draw “almost perfectly sinusoidal” current, with the majority of branch circuits measuring less than20% current THD. At the panel and PDU level, diversity effects — different loads drawing at different times — further reduce the aggregate harmonic impact.
This is the environment K-13 transformers were designed for. Mixed loads. Natural diversity. Moderate harmonic content that averages out across phases and over time.
AI infrastructure breaks every one of those assumptions.
A rack of NVIDIA GPUs is not a mixed load. It’s dozens of identical, high-power switching supplies running the same workload at the same time. When a training batch starts, they all ramp simultaneously. When a checkpoint runs, they all drop simultaneously. A joint research paper from Microsoft, OpenAI, and NVIDIA confirmed that“ due to the large and synchronous nature of the job, participating nodes are co-located to form a majority of a datacenter, or even multiple datacenters in the same grid, making the power swings visible at the rack, datacenter, and power grid levels.”
These swings happen fast. Research from the University of Alberta found that “modern AI accelerators can exhibit power variations exceeding 50% of their thermal design power (TDP)within milliseconds.” The GPU motherboard currently experiences “abrupt, multi-ampere surges” while the power supply AC input lags these transitions by30–50 milliseconds.
The problem for your PDU transformer isn’t just the magnitude. It’s the combination of factors:
The Uptime Institute put it plainly: these step changes “may cause power quality issues such as significant harmonics and sub-synchronous oscillations that distort the sinusoidal waveforms in AC power systems.”
NVIDIA themselves acknowledged the severity of this problem. Their GB300 NVL72technical blog explains that “thousands of GPUs operate in lockstep and perform the same computation on different data. This synchronization results in power fluctuations at the grid level.” Their engineering response was to add 65joules of electrolytic capacitance per GPU to the GB300 to smooth power demand by 30% at the grid input level — a hardware fix to an infrastructure-level problem.
K-factor isn’t a marketing label. It’s a quantifiable rating defined in UL 1561 (Section3.5) that describes how much additional harmonic-induced heating a transformer can safely handle. The formula:
Where Ih(pu) is the rms current at harmonic “h” (per unit of rated rms load current) and h is the harmonic order. In plain terms, each harmonic current is squared, multiplied by its harmonic order squared, and summed across all harmonics. This captures a physical reality: because winding eddy current losses increase with the square of frequency, higher-order harmonics cause disproportionately more heating in transformer windings.
Specifically, winding eddy current losses increase with the square of the harmonic frequency. A 5th harmonic current (300 Hz) creates 25 times the eddy current heating that the same magnitude of current would cause at the fundamental 60 Hz. The 7th harmonic creates 49 times. The 13th creates169 times. These losses concentrate in the winding conductors, particularly at their surfaces due to skin effect and between adjacent conductors due to proximity effect.
K-factor ratings correspond to the percentage of non-linear loads a transformer is designed to serve:
The physical differences between a K-13 and K-20 transformer are not subtle. K-20 requires:
A K-13 transformer has these features too — but sized for a load profile where at least a quarter to half of the connected equipment draws relatively clean power. Strip away that linear-load cushion, and the thermal margins that made K-13 adequate for traditional data centers disappear.
Harmonic-induced overheating doesn’t trip an alarm and shut things down cleanly. It degrades transformer insulation gradually and then fails abruptly.
Transformer insulation aging follows the Arrhenius relationship: for dry-type transformers, approximately every 10°C above the rated hot-spot temperature cuts insulation life in half. A transformer designed for 20 years of service that consistently runs 30°C above its thermal design point won’t last five years. The failure mode isn’t gentle — it’s thermal protection tripping your PDU and taking down every rack downstream, or worse, an insulation breakdown that requires emergency replacement.
And replacement isn’t quick. Dry-type PDU transformers in the 75–500 kVA range currently run 12–30 week lead times industry-wide. CustomK-20 configurations take longer. Add emergency premium pricing, and you’re paying substantially more for a transformer you should have specified correctly the first time.
But the transformer cost is the smallest number in the equation. For a GPU-as-a-service provider like CoreWeave — now public and operating 33+ data centers with racks drawing 120–132 kW each — a PDU failure during a training run means lost compute hours that can’t be recovered, potentially restarted training jobs, and SLA penalties to customers paying premium rates for GPU time. A single NVIDIA GB200 NVL72 rack draws approximately120 kW. An unplanned outage affecting even a few racks during a multi-week training run is an expensive lesson in transformer selection.
Not every data center needs K-20. Here’s an honest framework:
K-13 is the right specification when:
K-20 is the right specification when:
If you’re deploying NVIDIA GPU infrastructure and you’re not sure what your harmonic profile will look like, specify K-20. The cost difference between K-13 and K-20 is a fraction of the replacement cost if you spec’d wrong.
Not every transformer manufacturer offers K-20 as a standard product. Schneider Electric does not offer K-20-rated transformers at all. Hammond Power Solutions offers K-20 only through their Synergy line; their standard Sentinel K series caps at K-13. For most suppliers, K-20 is a special order that adds weeks or months to delivery.
When evaluating a K-20 supplier for AI infrastructure, the specifications that matter go beyond the K-rating itself. Ask whether K-20 is a catalog product or a custom build. Ask about neutral bus sizing — 200%+ is the minimum for GPU environments where triplen harmonics drive neutral current above phase current. Confirm delta-wye configuration for triplen harmonic isolation, Class H insulation for thermal headroom, and 100%factory testing under load. And ask about lead times: PDU-class dry-type transformers (75–500 kVA) should ship in weeks, not months. The 120-week lead-time figures cited in trade press refer to utility-scale power transformers — a different product category entirely.
Quality Transformer & Electronics manufactures K-20 dry-type transformers as standard catalog products across the75–500 kVA PDU range, with typical lead times of 4–6 weeks from facilities in the Bay Area, Greater Los Angeles, and Nevada.
