Fake microSD Cards on Amazon: How to Spot and Test Them

Two habits filter out almost every fake microSD card: buy only listings sold by Amazon.com itself or the manufacturer’s official store, and run a full-capacity test with a free tool (h2testw or F3) the day the card arrives — while you can still return it. Counterfeits pass every casual check, because the lie lives in firmware: the card reports 1TB while containing a fraction of it, and your files quietly corrupt once they cross the real limit.

Here is how the fraud works, the listing red flags that mark a fake before you pay, and the ten-minute setup that proves what you actually received.

How a fake card lies

A microSD card has two parts that matter here: flash memory chips and a small controller that tells the host device what the card is. Counterfeiters take low-capacity flash — sometimes factory rejects — and reprogram the controller to report a much larger size. Every device downstream believes the controller. The card formats to “1TB,” accepts files, and shows them in the file browser.

The failure arrives later. When your data exceeds the real capacity, the controller typically wraps new writes back over old ones or writes into nothing. File names remain in the directory; the contents are gone or garbage. This is why fake cards are crueler than dead ones — a card that never worked gets returned the same week, while a fake works convincingly until you trust it with a vacation’s worth of photos.

The speed label usually lies too. Cards faking a SanDisk or Samsung identity routinely deliver a fraction of the genuine card’s write speed, which shows up as dropped frames in 4K recording and glacial file transfers — often the first symptom people actually notice.

The price floor test

Flash memory is a commodity with a real manufacturing cost, and nobody sells it far below that cost in volume. Genuine 1TB microSD cards from major brands sell in the $100 range — SanDisk’s 1TB Extreme, for example, has been a headline deal at $99.99 — and even aggressive legitimate discounts leave the real thing at several times the price of the suspicious listings.

So a “1TB” card at $12–20 is not a deal, a clearance, or a lucky find. There is no path by which that product profitably contains a terabyte of working flash. The same arithmetic scales down: a “512GB” card at $8 fails it just as hard. When a capacity-to-price ratio beats the best brand-name sale price by 4-5x, the capacity is the part that’s false.

Two honest exceptions to keep the test fair: small capacities (32–128GB) are genuinely cheap now, often under $20 from real brands — cheap 128GB is normal, cheap 1TB is not. And older-generation cards from real brands do get discounted deeply. The test is the floor, not the discount.

Reading the listing before you buy

On Amazon, the highest-signal check sits right under the buy button:

The “sold by” line deserves emphasis because of how Amazon’s marketplace works: many listings share a single product page across many sellers, and the seller can change depending on who wins the buy box. Tech outlets that have covered counterfeit memory for years converge on the same advice — on Amazon, buy first-party or official-store listings only; better yet, for memory cards specifically, retailers without third-party marketplaces (Best Buy, B&H) or the manufacturer’s own store remove the ambiguity entirely. Reviewers also document that warehouse commingling means even first-party stock is not a mathematical guarantee — which is why the test in the next section is not optional.

Check SanDisk microSD cards on Amazon

Check Samsung microSD cards on Amazon

Test every card the day it arrives

Free tools exist for exactly this job, and they are the only verdict that counts:

h2testw (Windows). The classic, published by Germany’s c’t magazine. Point it at the empty card; it fills the space with 1GB test files, then reads every one back and compares. The result is unambiguous: a genuine card reports all data verified with plausible write/read speeds; a fake reports corrupted data at almost exactly the card’s real capacity — the tool effectively prints the true size of the card in its error summary.

F3 — Fight Flash Fraud (macOS / Linux, Windows via Cygwin). The actively maintained open-source equivalent, split into f3write and f3read, which perform the same fill-and-verify cycle. The project’s own documentation describes it as inspired by h2testw, built to catch cards that “wrap around” writes beyond their real capacity. It also ships f3probe, a much faster destructive probe for cards you haven’t put data on yet.

The procedure that makes the test worth it:

1. Test before first use, not after. The write phase overwrites the card. Day one, before any photos.
2. Use a decent USB card reader — testing through a slow reader turns hours into a full day. Any reputable USB 3.x reader is fine; this is also your chance to sanity-check write speed against the card’s class rating (a V30 card sustaining 4MB/s writes is failing a different way).
3. Let it finish. Fakes usually reveal themselves early, but a clean pass requires the full run.
4. Keep the result inside the return window. A failed test plus a screenshot is a clean “item not as described” return case on Amazon. Return fakes rather than keeping them for small jobs — a card that lies about capacity has no honest size you can trust.

One tool clarification, because it trips people up: a speed benchmark is not a capacity test. Benchmarks like CrystalDiskMark write a few gigabytes to one small region of the card and time it — a fake with 64GB of real flash will benchmark perfectly, because the region it tests is real. Only a tool that touches the entire claimed capacity, the h2testw/F3 way, can catch the lie. Run a benchmark afterward if you care about speed; run the fill-and-verify first because it’s the one that catches fraud.

Signs a card you already own is fake

Plenty of readers arrive at this article after the purchase, when a card has started acting strange. The classic in-use symptoms:

• Old files corrupt as new ones are added. Photos from months ago suddenly won’t open, right after you copied a big batch on. That is the wraparound writing over them — the signature of a capacity fake.
• Videos truncate at the same point. Recordings stop or corrupt once total stored data crosses a consistent threshold.
• Copies to the card crawl. A card sold as V30-class sustaining single-digit MB/s writes is misrepresented even if its capacity happens to be real.
• The file system needs repair over and over. Devices keep asking to reformat a card that was fine last week.

If any of these fit, stop writing to the card immediately — every new write on a wrapped fake destroys existing data. Copy off everything that still opens, then run h2testw or F3 on the emptied card for a definitive verdict. If it fails outside the return window, it is still worth reporting the listing; capacity fraud violates marketplace rules everywhere, and reports are how listings get pulled.

The same firmware trick, by the way, is endemic to cheap no-name USB flash drives and “16TB portable SSDs” on marketplace sites — the identical fill-and-verify test applies, and the identical price-floor logic condemns them. A genuine multi-terabyte SSD costs what NAND costs, which is why the real ones look like the drives in our NVMe enclosure guide, not like a $29 miracle.

Physical tells (weaker, but free)

Packaging and print quality catch the lazier fakes: blurry or bolded fonts, missing serial numbers, absent certification logos, off-color branding, or a card whose printed label differs from the box. Genuine brand cards are consistent products; counterfeit runs are not. Treat these as supporting evidence only — good fakes have good packaging, which is why the capacity test remains the referee.

The same skepticism, incidentally, serves you across categories: memory cards are simply the easiest product to fake because the lie is invisible until data loss. When storage matters — game libraries on a Steam Deck, working files on an external NVMe drive, or the family photo archive on a NAS — the through-line is identical: buy from accountable sellers, and verify before you trust.

The 60-second version

• Never buy capacity that beats the best brand-name sale price by multiples — the economics don’t exist.
• Buy first-party or official-store listings; prefer no-marketplace retailers for memory cards when convenient.
• Test with h2testw (Windows) or F3 (Mac/Linux) before first use, every card, even from good sellers.
• Return anything that fails, with the test result attached.

A genuine card costs a few dollars more and several hours of one-time testing. A fake costs whatever the data on it was worth.

More storage and buying guides are on gadget-hub.jp English.