This piece covers what he found, why the numbers work out that way, and how to run the same test on your own setup.
What the author found
The report is a single setup, not a lab test. It doesn't name the SSD, hub or PC, and it gives no measured speeds, so I can't give you a slowdown percentage.
What it does say:
- With the SSD plugged straight into the PC, it behaved normally. CrystalDiskMark showed the advertised sequential read and write speeds.
- A real-world check, copying large files and reading them back, also showed no notable problems.
- The direct-connection result cleared the drive and its cable. The hub was the one remaining variable.
- The drive supports USB 3.2 Gen 2x2. The hub in the path was only USB 3.2 Gen 1. The hub itself plugged into a USB4 port on the PC.
- The fix was to replace the hub. The author says that cost less than replacing the 2TB drive.
Why a Gen 1 hub hurts a Gen 2x2 drive
USB links run at the best speed that both ends support. The names map to signaling rates like this:
| Name | Signaling rate |
|---|---|
| USB 3.2 Gen 1 | 5 Gbps |
| USB 3.2 Gen 2 | 10 Gbps |
| USB 3.2 Gen 2x2 | 20 Gbps |
One comparison source puts it plainly: a 10 Gbps Gen 2 device connected through any 5 Gbps link will run at the slower 5 Gbps tier because the lowest common capability becomes the bottleneck. The same logic applies to a 20 Gbps drive behind a 5 Gbps hub.
In practice, one source that tested drives reported roughly 465/464 MB/s read/write on USB 3.2 Gen 1 versus about 2087/2010 MB/s on Gen 2x2. That was Kingston's test of its own XS2000, so treat it as a vendor figure for one product. It still shows the scale of the gap. A Gen 1 path can plausibly cost a fast SSD most of its speed.
Remember what these numbers are. They are link signaling rates, not guaranteed file-copy speeds. Real throughput is lower because of protocol overhead.
A USB4 port doesn't guarantee a fast hub
The author notes the PC's USB4 port had plenty of headroom. That's true of the host, but it doesn't fix a slow hub downstream. The USB Implementers Forum (USB-IF) says USB4 is backward compatible and that the resulting connection scales to the best mutual capability of the devices being connected. A USB4 port feeding a Gen 1 hub gives you a Gen 1 hub.
Two cautions:
- USB4 branding doesn't promise a Gen 2x2 downstream port. A hub's upstream link, each downstream port and the cable all matter.
- A headline bandwidth figure doesn't describe each port. Check the specs for the exact port you plan to use for the SSD.
Shared bandwidth and power: possible factors, not proven here
The author also points to two other hub limits. I'm treating both as things to investigate, not findings from his setup.
Bandwidth sharing. His hub also carried dongles for a wireless keyboard, mouse and controller, and sometimes a phone. Microsoft's USB driver documentation confirms that competition for bandwidth can come from several hardware and software sources and is hard to predict. It also says hub transaction translation can use bandwidth, depending on the connected devices and the topology.
That doesn't show that idle input dongles noticeably slow an SSD. The author gives no evidence on that. Keyboard and mouse traffic is tiny next to SSD transfers. In his case the Gen 1 ceiling is the clearly supported cause.
Power. Hubs differ in how much power they supply, and a bus-powered hub may not feed a hungry SSD enough. The author warns this can mean slowdowns or unexpected disconnects that risk data errors. He didn't report seeing disconnects or corruption, and he didn't test power. Still, it is worth checking if a drive drops out or stutters through a hub.
How to test your own setup
- Connect the SSD directly to a PC port you know supports the speed it needs. Skip the hub, dock or monitor passthrough.
- Benchmark it. Run a sequential test such as CrystalDiskMark, then copy a large file to confirm. Use the same cable and the same file, and don't run other heavy transfers at the same time.
- Repeat through the hub under the same conditions and compare.
- If the hub run is slower, check the whole path. That means the host port, the hub's upstream link, the downstream port you used, and the cable. Look up the actual data-rate specs. Don't go by the USB-C connector shape.
- Check power if the drive is flaky. Find out whether the hub is bus-powered or has its own adapter, and what each port is rated for.
- Rule out contention. Unplug other devices, or leave them idle, and retest to separate a hard speed ceiling from competition for capacity.
Choosing a replacement hub
If you want the drive to stay connected through a hub, look for these before buying:
- An upstream connection that matches or exceeds the drive's speed.
- A downstream port rated for the drive's mode, such as 10 or 20 Gbps.
- External power if you plan to run several devices.
- A short, appropriately rated cable.
Another option is to reserve a fast direct port for the SSD and put slower accessories on the hub. That's roughly what the author did, except that he also swapped the hub.
Bottom line
The author's A/B test was simple: plug the drive in directly and compare. It cleared the SSD and its cable, and it pointed at a Gen 1 hub as the bottleneck. That's cheap to try before you blame a drive or buy a new one. This is one reported setup without published speed numbers, so don't assume any hub with a few dongles will cut your SSD's speed. Check the weakest link in the path first.