This is one where the standard advice — “add correction capacitors to fix power factor” — is actually the wrong fix for a VFD specifically, and it’s worth explaining why, because getting this wrong on a real installation can make things worse, not better.

The distinction that matters: displacement vs. true power factor

Power factor has two components, and they need different fixes:

Displacement power factor is the phase shift between voltage and current caused by inductive or capacitive loads — a motor running directly across the line, for example, draws current that lags voltage. This is what correction capacitors are designed to fix, by supplying leading reactive current to cancel the lag.
True (total) power factor also accounts for harmonic distortion — current drawn in a non-sinusoidal shape, which happens with any load that draws current in pulses rather than a smooth sine wave. This includes VFDs.

A VFD’s input stage is a rectifier — it charges the DC bus by drawing current in short pulses near the peak of each voltage half-cycle, not smoothly across the whole cycle. That pulsed current is rich in harmonics (mainly the 5th, 7th, 11th, and 13th). The result: a VFD’s displacement power factor is often already close to unity (commonly 0.95–0.98) even with no correction at all, but its true power factor as seen from the supply can look considerably worse — sometimes down around 0.7–0.8 — purely because of harmonic distortion, not phase lag.

This is the key point: capacitors fix displacement power factor. They do essentially nothing for the harmonic-driven true power factor problem a VFD actually has. Worse, adding capacitors sized for a “normal” inductive load onto a VFD-heavy system can create a resonance between the capacitor bank and the system’s inductance at a frequency close to one of the harmonics the VFD is generating — which can amplify that harmonic rather than suppress it. This isn’t a minor caveat; it’s the actual reason you don’t just bolt PFC capacitors onto a VFD installation without checking for it first.

What actually improves a VFD’s power quality

AC line reactors or DC link chokes — inductors placed on the drive’s input or DC bus that smooth the current pulses the rectifier draws, directly reducing harmonic content. This is the standard, inexpensive first step, often bringing total harmonic distortion down meaningfully (roughly from 80%+ unmitigated down to somewhere in the 30–40% range, depending on the reactor’s impedance) — check the specific reactor’s spec rather than treating this as a fixed number.
Passive harmonic filters, tuned to the specific harmonics a given drive or system produces, for installations needing tighter compliance with harmonic limits (e.g. under IEC 61000-3-2/3-12 or a utility’s connection requirements).
Active harmonic filters or 12-pulse/18-pulse rectifier configurations, for larger installations or where multiple VFDs on one supply push aggregate harmonic distortion high enough to affect other equipment on the same network.
An active front end drive, which replaces the diode rectifier with an actively controlled bidirectional one — this genuinely achieves both near-unity displacement power factor and low harmonic distortion simultaneously, at higher cost, and is the same technology used for regenerative braking on the electrical side.

One more correction worth making: DC bus capacitors aren’t power factor correction

The original framing that “modern VFDs come with built-in power factor correction capacitors” conflates two different things. The capacitors inside a VFD’s DC bus are there to filter and smooth the rectified DC voltage — they’re essential to how the drive works internally, but they are not performing grid-side power factor correction. If a drive’s datasheet claims a high displacement power factor, that’s usually inherent to the rectifier design, not an added correction feature.

What’s actually worth checking before doing anything

Check what your utility actually bills or penalizes on — displacement power factor, true power factor, or a harmonic-distortion-specific charge. Switzerland and most of the EU increasingly look at total harmonic distortion limits on larger connections, not just the classic cosφ. This determines what problem you’re actually solving for.
Measure before correcting. A power quality analyzer reading at the point of connection tells you whether you have a displacement problem, a harmonic problem, or both — guessing leads straight into the capacitor-resonance issue above.
Size any harmonic mitigation to the actual drive and system, not a generic percentage — a qualified electrical engineer or the drive manufacturer’s application support can confirm reactor or filter sizing for your specific load.

The short version

Power factor correction capacitors are the right tool for lagging reactive power from motors and other inductive loads running directly across the line — they are largely the wrong tool for the power quality problem a VFD actually creates, which is harmonic distortion from its rectifier front end, not phase lag. Line reactors, DC chokes, harmonic filters, or an active front end are what actually address that. Measure what you’re dealing with before choosing a fix — adding capacitors to a VFD system without checking for resonance risk can make harmonic problems worse, not better.

If you’re seeing a power factor penalty or harmonic compliance issue on an installation with several VFDs, that’s worth measuring properly before assuming a capacitor bank is the answer.