ReviewGamers

Loading...

ReviewGamers
Handhelds

Handheld storage explained: microSD Express, UHS-I and internal SSDs

The cheapest storage mistake is assuming a card is a card. Two cards that look identical can differ by roughly ten times in throughput.

A handheld gaming console with its memory card slot open and two memory cards beside it.
Evidence: Tier C – Aggregated & AttributedFigures are normalised from named third-party measurement and linked to their source. They are not ours.How we test

Last verified 2026-09-13 by ReviewGamers Editorial

Storage is the cheapest place to make an expensive mistake, because two cards that look physically identical can differ by roughly ten times in throughput — and one of them may not work properly in your device at all.

The three things you might be buying

microSD (UHS-I). The familiar standard. The SD Association documents UHS-I at up to 104 MB/s. It uses a single row of contacts and has been the default for years.

microSD Express. A different architecture in the same physical shape. It uses PCIe and NVMe interfaces — the same technology as a laptop SSD — with ceilings up to roughly 985 MB/s, around ten times UHS-I.

Internal SSD. On handheld PCs, usually an M.2 drive in a compact form factor. Fastest, most expensive per gigabyte, and on most devices requires opening the machine.

The compatibility trap worth understanding

microSD Express is backwards compatible with UHS-I slots — but in those slots it runs at UHS-I speeds. You get the card’s price and the old card’s performance.

The reverse matters more. The Nintendo Switch 2 uses microSD Express as its removable storage standard. A UHS-I card you already own is not a substitute, regardless of how new or expensive it was. This catches a lot of people who reasonably assume their existing cards carry over.

Practical rule: match the card to the device’s stated standard, not to the physical slot. The slot will accept things the device cannot properly use.

When faster storage actually changes anything

Be sceptical here, because this is where storage marketing overclaims.

Faster storage reliably improves:

  • Load times, sometimes substantially
  • Texture streaming in large open-world games, reducing pop-in and stutter
  • Install and copy times

Faster storage does not improve frame rate in a game that is already loaded and running. If a game runs poorly because the chip cannot keep up, a faster card will not fix it. Storage is an I/O problem, not a rendering one.

Card or internal drive?

Card, if: you want capacity cheaply, you swap between devices, you mostly play smaller or older titles, or you do not want to open the device.

Internal drive, if: you play large modern titles, load times genuinely annoy you, or you are already comfortable opening the machine and it does not void anything you care about.

Both, realistically. Internal for the handful of games in rotation, card for the library.

Capacity: buy more than you think

Modern game installs are large and getting larger. The smallest configuration of any handheld is rarely the best value once you add the storage you inevitably need. Work out the total cost — device plus storage — before comparing devices at all, because that is the number you will actually pay.

Buying safely in Australia

Two specific risks:

Counterfeits. Fake memory cards are a long-running problem on marketplaces. They report the advertised capacity and fail once you exceed the real capacity, frequently taking your data with them. Buy from authorised retail rather than a marketplace seller with no accountability.

Consumer guarantees. Under the Australian Consumer Law, goods must be of acceptable quality and fit for purpose, and your claim is against the retailer. That is meaningful protection with a local seller and close to unusable with an offshore marketplace account.

What we have not done

We have not measured any card or drive. Real throughput differs from rated throughput, and sustained write speed is usually where cheap cards fall apart — which is exactly the measurement that requires owning the hardware and running a published protocol.

Those results will appear here as Tier A evidence once we have the devices. Until then this page explains the standards and stops there.