If you're about to pay for a big batch of microSD cards, hold off until a sample has been tested for its true capacity and for write speed over a long copy. Check that the cards in one lot match each other, too. The supplier should also put the NAND grade, the controller and a locked parts list in writing, and those few steps stop most of the trouble that turns a cheap deal into a costly one.
We'll walk through each test, how many cards you should check and the questions to raise with a supplier.
Why Bulk microSD Orders Go Wrong
Bad orders tend to fail in a handful of ways. Some cards claim a bigger size than they really have. Then there are cards built on downgraded or recycled NAND, which seem fine at the start and give up months later. The speed printed on the label may also be more than the card can keep up for long.
Mixed batches cause the rest. A first shipment can pass every test while the next one comes from a different factory run with different flash inside. If your product writes constantly, standard cards may not be the right choice at all, and high endurance cards are worth comparing.
Capacity Checks: Catching Fake Cards
Fake capacity is the most common bulk problem, and the easiest to catch.
1. How Fake Capacity Works
Inside the card, the controller has been set to show a size it doesn't really have, for example 128GB on a card with only 16GB of flash. Your computer takes that at face value. As soon as you write beyond the true limit, new files vanish or older ones get overwritten, and quite often no error ever pops up.
2. Running a Full Write and Verify Test
A quick file copy won't catch this. On Windows, H2testw is the tool to use, and on Mac or Linux it's F3. Both fill every last bit of the card with test data, then read the whole thing back. Format the card beforehand and pick all of the free space. A 128GB card can take an hour, sometimes more, so it helps to run several cards side by side on separate readers.
3. What a Failed Result Looks Like
A good card reads back every byte with no errors. A fake shows errors that start at a certain point, and the tool reports a real capacity far below the label. When even a single sample fails like that, turn down the whole lot. Fakes almost never show up alone.
Speed Tests That Match the Label
|
Mark
|
What it promises
|
|
Class 10 / U1
|
At least 10MB/s sustained write
|
|
U3 / V30
|
At least 30MB/s sustained write
|
|
V60 / V90
|
At least 60MB/s or 90MB/s sustained write
|
|
A1
|
At least 1,500 random read and 500 random write IOPS
|
|
A2
|
At least 4,000 random read and 2,000 random write IOPS
|
These marks come from the SD Association, and they're minimums the card has to hold over time, not peak numbers. Why does that matter? Plenty of cards write quickly for the first few gigabytes, then slow right down once their cache is full. So test with a big file, 10GB or more, and keep an eye on the speed from start to finish rather than for the first few seconds.
How Many Cards to Test: Sampling a Bulk Order
Nobody can test every single card, which is why buyers rely on sampling tables such as ISO 2859 or ANSI Z1.4. Under the general level II plan most people use, a lot of 1,201 to 3,200 cards needs 125 samples. For 3,201 to 10,000 cards, it's 200.
Take samples out of different cartons, not just one box. Hang on to any cards that fail, along with their packaging, as you'll need them if you make a claim.
|
⚠ A simple sampling plan
Every lot: full capacity test and sustained speed test on the sample.
First order: add a long write test on a few cards to check for slowdown over time.
Any failure: hold the lot and ask the supplier for batch records before accepting it.
|
What to Ask Your Supplier Before You Order
Get these answers in writing before you pay a deposit:
-
NAND grade: is it original factory grade, or downgraded flash that didn't pass full testing?
-
Which controller is on the card? Ask whether it will stay the same.
-
Will the CID and manufacturer ID match across every lot?
-
A locked bill of materials, so the cards in production match the samples you signed off.
-
How they'll arrive: bulk trays, blister packs or custom retail boxes.
-
Batch test reports, plus the warranty and return terms.
-
Lead time, and how much warning you'll get before any part changes.
A supplier that can't answer these clearly is a risk at any price. Digiera supplies bulk microSD cards for OEM orders, with CID customization and speed class sorting, from Digiera's production facility.
Reading the Card: Markings, CID and Batch Consistency
Inside every SD card there's a CID register. It holds a manufacturer ID, a product name and a serial number. On Linux you can read it from the system files when the card sits in a direct reader, and some USB readers pass it through too. Many cheap USB readers don't.
Compare the CID and the printed markings across your samples. Mixed manufacturer IDs or different markings inside one lot are a strong warning sign. It usually means the supplier pulled cards from more than one source.
Loading Data and Duplicating Cards for Products
When the cards ship inside a product, you'll most likely load an image onto each one. A duplicator can write a stack of cards in one go. It should always verify them afterwards, though, because a weak card often only fails once it's been written in full.
Some suppliers can preload content or set a custom CID during production, which saves a step on your line.
Bulk microSD Pricing in 2026
Card prices have gone up all year along with NAND costs. TrendForce still expected NAND contract prices to keep climbing in the fourth quarter of 2026, with AI server demand behind it. With genuine flash costing more, any offer well below the going rate has a higher chance of being downgraded or fake.
Where you can, plan orders ahead and fix prices for a set period. Send your capacity, speed class and quantity to Digiera and request samples and a quote.
FAQs
How do I know if a microSD card is fake?
Run a full write and verify test. H2testw or F3 will do it. A fake card throws errors partway through and shows a real capacity far below the label, and very low prices or missing package details are common warning signs as well.
How can I test for the actual capacity of the microSD card?
Format it first and fill all the available space with H2testw using Windows or F3 using Mac/Linux. If there are any errors, the capacity is lower than advertised. Spend at least an hour testing a 128GB card.
What is the best method to check the speed of the microSD card?
Transfer one big file, preferably over 10 GB. Pay attention to the speed throughout the whole process, not just at the beginning. Most of the cards slow down after filling the cache, but a good one must keep up with its promised minimum.
How many cards should I test from a bulk order?
Use a sampling table such as ISO 2859. Under general level II, a lot of 1,201 to 3,200 cards calls for 125 samples. Pull them from different cartons, and test every new lot.
Are bulk microSD cards lower quality than retail?
Not always. There are many bulk microSD cards that are manufactured in the same facilities as retail microSD cards, but without packaging. Quality depends on the NAND grade and the vendor; therefore, one must get confirmation about both in advance.
What does A1 and A2 mean on a microSD card?
They're application performance classes, meant for running apps straight from the card. For A1, the minimum is 1,500 random read and 500 random write IOPS. A2 raises that to 4,000 and 2,000, though the host device has to support A2 to get the full benefit.
What is a CID on an SD card?
It is a register within the card. It contains the manufacturer ID, product name, and serial number. Verification of CIDs among a bulk consignment is useful to verify that the cards are sourced from the same place, and some suppliers provide a customized CID.
Why are some bulk microSD cards so cheap?
Most of the time it's fake capacity, downgraded NAND or recycled flash. Genuine NAND got more expensive through 2026. A price far below the market should make you suspicious, so test samples before paying for the full order.
