GaN Charger: Watts Per Cubic Inch, Not a Generation
Home
Yiwu, Zhejiang, China
Mon – Fri 9:00 AM – 6:00 PM · Sat 9:00 AM – 4:00 PM
[email protected]

What Is GaN Charger Technology, and Why the Box Shrank

Alan Xiao, Founder of Sourcing Pioneer
Alan Xiao Founder of Sourcing Pioneer
· Updated November 9, 2026 · 9 min read

A GaN charger uses a different switch. Everything else is still a charger.

The letters stand for gallium nitride, the material of the power transistor inside. It replaces the silicon switch that has done this job for decades.

That single swap changes how fast the circuit can switch, and the switching speed is what makes the box small. It does not change the charging protocol, the cable, or the phone.

Studio still life of a small white USB-C charger with a coiled cable on a beige surface
Small box, same job. The material decides the size, not the wattage alone.

The Short Answer

GaN is a wide-bandgap semiconductor. Its electrons need more energy to break free, and the crystal survives much stronger electric fields than silicon before it fails.

That property lets a GaN device be thinner for the same voltage rating. A thinner device holds less internal charge, so it can switch on and off faster without wasting the transition as heat.

Faster switching is the whole commercial story. A transformer only needs to store energy for one switching cycle, so a shorter cycle means a smaller magnetic core. Smaller core, smaller capacitors, smaller box.

Everything else in the charger stays: the bridge rectifier, the controller, the feedback loop, the output capacitors, the case, the plug. GaN changes one part of the bill of materials.

Where the Size Actually Goes

Power density is the honest way to compare two chargers. It is the rated output in watts divided by the volume of the case, in watts per cubic inch.

Published figures from a device maker put older silicon adapters at 3 to 5 W/in³. GaN designs reach 9 to 18 W/in³, which is close to a fourfold jump in the same space.

MeasureSilicon designGaN design
Power density3 to 5 W/in³9 to 18+ W/in³
Switching frequencyGenerally under 100 kHzHundreds of kHz, up to MHz
100 W two-port volumeReference40 to 60 % smaller
Same charger, weightReference40 to 55 % lighter

The weight figure is the one buyers notice. A lighter charger means less copper, less iron and less aluminium, and those are three of the costs the factory is trying to remove.

It also means less mass to absorb heat. That trade is the subject of a later section, and it is where cheap GaN designs fail.

Why Frequency Is the Lever

A switching supply moves energy in bursts. Each burst charges a transformer, then releases it to the output. The faster the bursts, the less energy each one has to hold.

Silicon MOSFETs are practical below about 100 kHz. Above that, the losses during each transition climb faster than the savings, and the device heats up.

GaN switches in the hundreds of kilohertz, and in some designs into the megahertz range. The transition time is short enough that the loss per switch stays small even at high rates.

Two circuit families make that useful. An active clamp flyback recycles the energy that a plain flyback burns in a snubber. It also holds the switch voltage at zero before the switch turns on. An LLC resonant stage does the same job with soft switching at both ends.

The published numbers for those stages reach 95 to 97 % or better at full current. That is the converter, not the whole charger, and the difference matters when a buyer reads a claim.

Flat bar chart comparing silicon charger power density with GaN charger power density
Three to five watts per cubic inch becomes nine to eighteen. The physics is the reason.

What Thermal Headroom Buys

Every watt lost inside the box becomes heat that has to leave through the case. A GaN design loses less, so it starts with an advantage.

Then the advantage gets spent. The same shrink that cuts the copper also cuts the surface that sheds heat. A small hot box is harder to keep cool than a large warm one.

So the win is not automatic. Thermal headroom can buy three different things, and only one of them is size.

It can buy size. Keep the loss the same, shrink the box, and accept a higher case temperature.

It can buy margin. Keep the box, cut the loss, and run cooler than the silicon version at the same output.

It can buy headroom for a second port. Spend the savings on another output, so two devices charge together instead of one.

A fifth option exists, and it is the one cheap designs pick. Spend the savings on the bill of materials, keep the box the same size, and sell it on the name alone.

A datasheet never says which of the three the designer chose. The case temperature at full load, measured after half an hour, is the number that reveals it.

One honest industry voice makes the point that gets buried in marketing. A GaN charger can put the same power into a smaller and hotter box. It may then wear out sooner than a bulky silicon one. The capacitors inside age faster when they run hot.

Measured Efficiency Is Not Stage Efficiency

This is the single most useful distinction on the page. Converter stage efficiency and wall-to-device efficiency are different numbers, and quotes mix them.

A test lab measured one 65 W ultra-thin GaN unit at 220 V. Across its load range, it landed between 84.6 % and 92.11 %.

Nothing is wrong with that result. It is simply measured at the wall socket and at the output plug, which includes everything between them.

The 95 to 97 % figure from the previous section belongs to the DC-DC stage alone, measured on a bench with instruments at both ends.

Between the two sit the input bridge, the power factor stage, the standby circuit, the output rectifier and the resistance of the cable. Each one takes a slice.

So when a listing claims 96 % efficiency, the useful question is: at what point, and over what load range. A number without those two answers cannot be compared with anything.

Flat panel contrasting converter stage efficiency with measured wall efficiency
Two honest numbers, two different measurement points.

The Part That Grows With Power

Above 75 W, supplies are expected to include power factor correction. That stage keeps the current drawn from the mains close to a sine wave, instead of letting it arrive in spikes.

That stage adds parts and heat. On a silicon design it usually means a bridge rectifier plus a separate boost stage, both of which sit in the path.

GaN makes a different arrangement practical. A totem-pole power factor stage replaces the diode bridge with switches, which removes a fixed voltage drop from the main current path.

The published claim is that this and the lower switching loss together shrink a multi-port high-power block by 30 % or more.

For a buyer the consequence is simple. Below 75 W, GaN is mostly about size and weight. Above that line, GaN pays for itself in the power factor stage too. That is why the 100 W to 240 W tier is where it earns its price.

The 75 W line also splits the market. A 65 W design can skip that stage. A 100 W design cannot.

So two chargers of the same size can differ by a whole circuit. Ask which side of the line the design sits on, and whether the stage is built.

What GaN Says Nothing About

The protocol list is a separate specification, and it decides whether the charger is useful to a given laptop.

USB Power Delivery negotiates a voltage and a current. Programmable Power Supply adds fine voltage steps for phones that ask for them. The extended range reaches 48 V at 5 A, which is the 240 W tier.

A GaN charger can support all of that, or none of it. A silicon charger can support all of it too.

So the two questions are independent. What material is the switch, and which protocols does the port speak. A listing that answers only the first has told you about the size, not the compatibility.

Which Claims Can Be Checked

Marketing has reached the material. Some names describe a product family, some describe a brand, and one is a trademark that reads like a generation.

Treat any “GaN II” or “GaN III” as a label until a document appears behind it. There is no public generation numbering for charger switches, and the numbers change between sellers.

Six things can be checked instead, and each one is either on the datasheet or measurable on a bench.

ClaimHow to check it
SmallerPower density in W/in³, from dimensions and rated watts
LighterGrams, on a scale, against the model it replaces
EfficientWall-to-output measurement, with load points named
CoolerCase temperature at full load, after 30 minutes
CompatibleThe protocol list, port by port
SafeThe safety standard and the certificate number

Start with the number the seller never prints. Divide watts by volume, and the claim either survives or it does not.

Two of those six can be settled with a ruler, a scale and a thermometer. A buyer does not need a laboratory to check the first, second and fourth items in that list.

How to Read the Label

A charger label is short, and most of it is worth reading. Four lines tell you what the box can do.

Watts. That is the most the port can give. A laptop will only take what it asks for.

Volts and amps. These come in pairs. Look for the pair your device needs, not the top one.

Protocols. PD, PPS, or both. A port can give 65 W and still refuse the fine steps a phone wants.

Safety mark and number. A mark with no file number behind it is a sticker.

One more line is worth a look. The input range tells you whether the charger works on 110 V as well as 230 V. That decides whether one part number can ship to both markets.

If a line is missing, the box is not lying. It is simply not telling you, which is a different problem and usually a cheaper one.

Ten Minutes With a Ruler

Three of the claims on this page can be checked at a desk, with no lab and no parts.

Measure the box and multiply the three sides. Divide the rated watts by that number. You now have watts per cubic inch, and it can be compared with anything else on the market.

Put the charger on a kitchen scale. Then put the model it replaces on the same scale. The gap is real, and the listing rarely states it.

Run it at full load for half an hour. Touch the case, then read the number on a $10 infrared gun. A case that climbs past the point of comfort is spending its margin on size.

None of this needs a test house. It needs ten minutes and the will to write a number down.

Five Lines for the Order

An order that says “GaN, 65 W” leaves every decision above to the factory. Five lines close it, and they fit the same brief you send a factory for any other part.

Flat card listing five lines for a GaN charger order
Five lines. Each one replaces an adjective with a number.

One. Name the switch material and the topology, since the topology is what harvests the material’s advantage.

Two. Give a power density target or the case dimensions, not the word “small”.

Three. Ask for wall efficiency at named load points, and the case temperature at full load after half an hour.

Four. List the protocols required, port by port, including whether programmable supply is needed.

Five. Name the safety standard and require the certificate number with the shipment.

The electrical side of the same product has its own file. A charger is checked for insulation before it ships. The way that test is set up is covered in this piece on the hipot test. The paperwork for a European sale is separate again, and the CE guide sets out what goes in it.

If the same product also carries a network port, the supply budget question in PoE power budgets is the next one to settle.

WATTS PER INCH

Send me the listing or the drawing you were given. I will work out the power density, say whether the efficiency claim names its load points, and list the five lines the order is still missing. No cost, no obligation.

Check My Charger

Questions Buyers Ask

Does a GaN charger charge faster than a silicon one?

Not by itself. Speed comes from the protocol and from what the device accepts. A 65 W GaN charger and a 65 W silicon charger deliver the same power to the same laptop.

Why is a GaN charger smaller?

Because the switch can run at a much higher frequency, so the transformer and the capacitors that store energy for each cycle can be smaller. The box follows the components.

Is GaN more efficient than silicon?

The switching stage usually is. The whole charger depends on the design, the load point and the power factor stage. Measured wall efficiency on a small 65 W unit can sit in the high eighties at part load.

Does a hot charger mean a bad one?

Not automatically, but it is the trade to watch. A smaller case has less surface to shed heat, and the capacitors inside age faster when they run hot.

What does GaN have to do with USB-C PD?

Nothing directly. Power Delivery is a protocol, and GaN is a switch material. A charger needs both, and they are specified separately.

Can I get 240 W from a GaN charger?

Yes, with the extended power range, at 48 V and 5 A. It needs a cable and a device that also support that range, so the charger alone does not guarantee it.

Questions & Comments

We read every one and reply within 24 hours

If the listing says GaN and stops there, send me the line. I will work out the watts per cubic inch and name what is missing.

Subscribe
Notify of
guest
0 Comments
Most Voted
Newest Oldest
0
Would love your thoughts, please comment.x
()
x