USB 10-Gig Adapters Are Finally Small Enough to

Here's something that doesn't happen often in tech anymore: a category of devices getting cheaper, cooler, and smaller all at once. We're so used to the opposite—incremental improvements at inflated prices—that when it reverses, it feels almost suspicious.

Tech YouTuber Jeff Geerling just tested WisdPi's new USB-C 10-gigabit Ethernet adapter, an $80 dongle that promises to do what Thunderbolt adapters have been doing for years, but without the Thunderbolt tax, bulk, or heat output. For anyone who's ever needed serious networking speed on a laptop, this sounds too good to be true. So I wanted to understand: is it?

The Thunderbolt Problem Nobody Talks About

If you've ever shopped for 10-gig networking on a laptop, you know the options have been bleak. Thunderbolt dongles work, sure, but they're expensive (often $200+), require Thunderbolt ports that not every laptop has, and—this is the part people discover after buying—they run hot. Like, "don't touch this while it's working" hot.

"All my Aquantia-based 10 gig adapters turn into little ovens," Geerling notes in his testing. "That's a big reason they're so big. The enclosures are basically giant heat sinks."

That heat isn't just annoying—it's a design constraint that keeps these adapters bulky and makes them questionable companions for thin-and-light laptops. So when manufacturers started using Realtek's new chipset in compact USB-C form factors, it opened a question: could USB finally deliver 10-gig speeds without the Thunderbolt baggage?

What Actually Happens When You Plug It In

Geerling's testing reveals the answer is: it depends on your computer in ways that are genuinely difficult to predict.

On a Framework 13 laptop running Windows 11 with a standard USB 3.2 Gen 2 port (10 Gbps), he maxed out at about 7.4 gigabits per second one-way, or 9 Gbps bidirectional. That's respectable—clearly saturating the USB bandwidth—but not the advertised 10 gigs.

On a MacBook with USB 4, which theoretically runs at 40 Gbps, the result was... 6.3 Gbps. Wait, what?

The issue, it turns out, is architectural. USB 4 tunnels older USB standards over Thunderbolt, and for non-USB 4 devices, that means you're limited to a single 10-gig lane. With overhead, that ceiling is around 6-7 Gbps, regardless of the fact that USB 4 itself is much faster.

To actually hit 10 gigabits, you need a USB 3.2 Gen 2 2x2 port—that's the 20 Gbps variant. Geerling tested on a desktop with exactly that port and finally saw the full speed: 9.5 Gbps with overhead, sustained and consistent.

"If you want to get the full potential, make absolutely sure you have a USB 3.2 Gen 2 2x2 20 GB port," he warns, "which good luck trying to figure that out without a detailed spec sheet for whatever computer you're using."

And yeah, that's the problem. USB naming is famously terrible—"USB 3.2 Gen 2 2x2" sounds like a parody of tech jargon—and neither Windows nor macOS makes it easy to determine what your actual port capabilities are. Windows Device Manager just said "USB 3.0" for a port clearly doing better than that. macOS showed incorrect speed ratings in settings.

The Value Calculation Gets Interesting

Here's where this stops being just a tech specs story and becomes genuinely interesting for anyone considering an upgrade.

Even on the "slower" setups—the ones capped around 6-7 Gbps—this adapter still outperformed a 5-gig adapter by about 40%. That 5-gig adapter costs $30. This 10-gig version costs $80.

So you're paying 2.6x the price for 1.4x the speed, and only if your computer can't deliver the full 10 gigs. If it can, the calculation shifts dramatically—you're getting genuine 10-gig performance for less than half what Thunderbolt adapters cost.

But Geerling raises the question that actually matters: "How many of you even have 10 gig capable networks? If you don't have a 10 gig switch and a use case that needs a full 10 GB of bandwidth, you're better off with 2.5 or 5 gigs."

This isn't dismissive—it's practical. Most home networks run on 1-gig switches. Most internet connections don't approach these speeds. Most file transfers don't need them. The people who do need 10-gig are running media servers, doing video production work, or managing large data sets locally. They already know who they are.

For everyone else, a $30 adapter that delivers 4.6 Gbps might legitimately be the smarter buy.

The Heat Thing Actually Matters

One detail that surprised me: thermal performance. Geerling measured the adapter at 42°C (about 108°F) after sustained bidirectional testing. That's warm enough to notice, but not "this might damage something" warm.

Compare that to Thunderbolt dongles that legitimately become too hot to touch, and you start to see why size matters. The old adapters needed bulk because they needed thermal mass. This new chipset apparently just... doesn't generate that much heat.

It's drawing about 0.86 watts at USB 2 speeds (his measurement setup limited testing at full speed), which isn't nothing but also isn't alarming for a device doing this much work.

Where This Goes Next

The WisdPi adapter isn't the only option using Realtek's new chip. AliExpress already has multiple variants, and the chip is showing up in PCIe cards for desktops too. As manufacturing scales and competition increases, prices will likely drop further.

What's less clear is how quickly this reaches mainstream manufacturers outside China, and whether the driver situation improves. Right now, Windows requires manually downloading Realtek drivers—the built-in ones don't work. macOS handles it natively but displays incorrect information in settings. These are solvable problems, but they're friction points for non-technical users.

The bigger question is whether USB standards bodies will ever fix their naming catastrophe. "USB 3.2 Gen 2 2x2" is user-hostile in a way that actively harms adoption. When even Device Manager can't tell you what you're actually using, something is structurally broken.

Still, for the first time in years, Geerling's right about this: "It's nice to have something that's cheaper, faster, and better again."

That's rare enough to notice.

—Zara Chen