What Drives the Price of a Power MOSFET
Die area, voltage class, package and test level set a power MOSFET price before anything else. Why two similar 650 V parts can differ threefold in cost.
OCENME Engineering · Published November 14, 2026 · 8 min read
Power MOSFET pricing is set by die area, voltage class, package and test level, in that order. Silicon accounts for roughly half to two thirds of unit cost, so a 650 V part at 0.19 Ω typically costs 40–60% more than a 0.45 Ω part on the same process, and an isolated package adds 10–25%.
Key points
- Die area scales with RDS(on), so the on-resistance you specify is the biggest single price lever
- Higher voltage class costs more per amp — a 650 V die is physically larger than a 500 V one at the same resistance
- Package and test add 10–30% — isolated tabs and 100% surge or leakage screening all appear in unit price
- A quote spread above roughly 30% usually means a different die generation, not a different margin
If two suppliers quote power MOSFETs that both say 650 V, 0.45 Ω, TO-220F on the front page, the price difference between them can still be a factor of two. The headline specification does not set the price. The price is set by four things in a fixed order — die area, voltage class, package and test level — and then modified by volume and by where the industry sits in the wafer supply cycle. Understanding that order is what turns a confusing quote sheet into something you can negotiate against.
Silicon is the price
The dominant cost in a power MOSFET is the die itself. Silicon typically accounts for half to two thirds of the unit cost at power device volumes; package, leads, assembly and test share the remainder. And die area is not a free variable — it is set directly by the resistance you specify.
For a given process and voltage class, on-resistance is inversely proportional to die area. Halve the RDS(on) and you roughly double the die. Since yield falls slightly as die gets larger, the cost curve is actually a bit steeper than linear. This is why the resistance number you pick is the single biggest price decision you make, and why a part specified at 0.19 Ω cannot be expected to cost the same as one at 0.45 Ω even from the same supplier on the same process — the die difference is on the order of 2.3 times, and 40–60% of that reaches the unit price after yield and packaging.
The practical consequence: before asking a supplier for a better price, check whether you actually need the resistance you asked for. Every 20% of RDS(on) you can give back is roughly 20% off the die cost. A loss budget done honestly at your operating temperature — where RDS(on) is 40–70% higher than the 25 °C datasheet number — often shows that the cheap part was the right part all along.
Voltage class and process generation
At the same resistance, a higher voltage class costs more, because blocking voltage requires a thicker, more lightly doped drift region. A 650 V die at 0.45 Ω is larger than a 500 V die at 0.45 Ω, which is larger than a 30 V die at 2 mΩ. This is also why price-per-amp comparisons across voltage classes are meaningless: a 30 V trench MOSFET and a 650 V super junction MOSFET are different products with different cost structures, not two grades of the same thing.
Within one voltage class, the process generation matters more than the brand. The shift from planar to super junction construction at 600–650 V cut RDS(on) per unit area by 3–5 times — we cover the technical side in super junction vs planar MOSFET. For pricing, the meaning is simple: a supplier still buying planar die for a 650 V adapter design is paying for roughly three times the silicon, and no amount of margin trimming will close that gap against a super junction quote. A quote spread above roughly 30% between two suppliers for a similar part number usually means a different die generation, not a different profit margin.
The same logic applies to silicon carbide. A 650 V SiC MOSFET carries a substantially higher price than a super junction silicon part at the same resistance, and the premium is justified only where switching frequency or thermal headroom converts into system savings. Below roughly 30 kHz at moderate power, the premium usually does not pay back.
Package and test
Package choice moves unit price by 10–30%, and it is the part of the quote most often compared carelessly. A TO-220 with a bare tab is the cheapest construction. A full-mould or isolated package (TO-220F, TO-263 with isolated tab, D-PAK-F) costs more because the moulding process is more complex and thermal resistance is worse, which sometimes forces the next resistance class down — a hidden price increase inside the package choice. PDFN-style surface mount packages save assembly cost at volume but require board copper to do the heatsinking, which shifts cost onto your PCB rather than eliminating it.
Test level is the other quiet cost driver. Standard production test covers the datasheet limits at room temperature. If your application needs 100% high-temperature leakage screening, 100% surge or avalanche testing, or lot-based HTRB monitoring with certificates, each of these adds handling, time and fall-out to the supplier’s line, and it arrives in the price. That premium is worth paying — a screened batch is cheaper than a field return — but it should be a conscious decision, visible in the quote as a line, not assumed to be included.
Volume and the market cycle
Below roughly 50,000 units per year, price is dominated by lot economics: setup, handling and the distributor’s margin. Between that and a few million units, price falls with volume in predictable steps. Above that, pricing becomes a capacity conversation with the die supplier.
The other variable is the wafer cycle. Power devices are built largely on 200 mm (8-inch) foundry capacity at mature nodes, and that capacity comes and goes in cycles. When it tightens, lead times stretch and prices rise across the industry, hardest in commodity voltage classes where suppliers have no pricing room. When it loosens, spot prices fall, sometimes below the levels a manufacturer can sustain, which is precisely when supplier consolidation happens. A manufacturer that controls its own packaging line can smooth some of this cycle by holding work in progress; a trading company reselling die cannot. This is one of the structural reasons buying from a manufacturer with in-house assembly behaves differently than buying from a reseller, even at the same part number.
What the cost drivers look like side by side
| Driver | Typical price effect | What you control |
|---|---|---|
| RDS(on) (die area) | Roughly proportional — halving resistance doubles die cost contribution | Pick resistance from a loss budget, not from habit |
| Voltage class | Higher class costs more per amp at the same resistance | Match class to the real bus voltage plus margin, not one class up “for safety” |
| Package | 10–30% between bare-tab and isolated or full-mould | Isolate only if your design needs isolation; check the thermal penalty |
| Test level | 5–20% for 100% HT or surge screening | Ask for screening as an explicit option, not bundled |
| Volume | Falling steps below ~1M units/year | Consolidate part numbers to move up a step |
| Market cycle | Double-digit swings over 1–2 years | Fix prices with the manufacturer, not the spot market |
When to shop on price — and when not to
Price-first selection is right when the application is a commodity power stage at moderate stress: an adapter, an auxiliary supply, a motor start circuit — anywhere the operating point sits well inside the safe operating area, the topology is proven, and the datasheet limits have margin at your worst case. There, the parts are genuinely interchangeable, and the quote sheet is the right tool.
Price-first selection is wrong when any of the following is true: the design sits near a limit (thermal, avalanche, surge), the part carries your safety or warranty exposure, or the volume is large enough that a silent die change would be expensive to detect. In those cases the qualification cost — which we describe step by step in how to qualify a second-source power device — dwarfs the unit price saving, and the correct comparison between suppliers is not price but change control: who notifies you when the die source moves, and whether the reliability data you were shown applies to the part you will actually receive.
A middle case worth naming: high volumes of a stressed part. There, price pressure is real but the answer is usually to relax the specification honestly — one resistance class up, verified to fit the thermal budget — rather than to push a borderline part through qualification on the strength of a cheap quote.
Limit and scope
This breakdown describes discrete silicon power MOSFETs at industrial volumes, sold directly or through distribution. It does not cover integrated modules, GaN products, or parts sold with automotive documentation flows, where documentation and traceability requirements change the cost structure entirely. The percentages given are typical ranges observed across quotes and should be treated as orientation, not as a quotation model — individual suppliers’ cost positions vary with their die source and packaging capacity. Nothing here is a substitute for a loss calculation on your own board; the numbers in this article explain price, not fitness.
We list 500–700 V high voltage types and 650–1200 V SiC types with their own datasheets on the high voltage MOSFET page and the SiC MOSFET page. If you are comparing a quote against a datasheet, the checklist in how to read a MOSFET datasheet covers which of those numbers were measured under conditions that match yours.
This article describes general manufacturing practice for machined and fabricated components. It is not a substitute for a drawing-specific engineering review — tolerances, material selection and process capability depend on the individual part.