Patriot Memory is a name that I’ve been aware of since the early days of this project. I think I came across them thanks to CameraMemorySpeed.com’s page of manufacturer and OEM IDs. They make a variety of memory products, including DRAM, SSDs, flash drives, and SD/microSD cards.
I bought these cards as a 5-pack. As you can see by the pictures, they didn’t bother with trying to put these in retail packaging — they just put them in a cardboard envelope. Inside were two hard plastic microSD holders: one with 4 cards, and one with a single card. (I guess nobody makes a cheap plastic holder that will hold 5 microSD cards?)
I’ve only tested one of these cards so far — so the information below is going to talk about the results that I got from that one card. (I’ll come back and update this once I’ve had a chance to test the others.)
Looking at this card’s CID information reveals something a little odd — the manufacturer ID is set to hex 12, and the OEM ID is set to hex 3456. This means that the first three bytes of the CID are hex 123456. I don’t believe for a minute that the SD Association assigned this to Patriot Memory…or anyone else for that matter.
I was originally going to consider these to be name-brand cards…but if I go strictly off of the criteria that I set out for determining what’s considered a name-brand card, it fails for a few reasons:
- It didn’t come in retail packaging
- The manufacturer can’t be easily determined by looking at the manufacturer ID/OEM ID
- The CID register bears signs that they were attempting to conceal the card’s true origin (e.g., the first three bytes of the CID being set to
123456)
If anyone from Patriot Memory is reading this…sorry. Do better next time.
These cards were actually quite skimpy, coming in at 5.48%. This is quite high, considering that the average for authentic cards — as of the time I’m writing this — is 1.76%. This means that for a 64GB card, you’re only getting about 60.5GB of usable space. (What are you doing with that extra 3.5GB, Patriot??)
But on the other hand, these cards did rather well on price — coming in at $0.066 per gigabyte, which is far below average (which is at $0.252 per gigabyte as of the time I’m writing this) — so just in terms of space, they’re a good value.
How did they do on performance? Again, I’ve only tested a single card so far — so this may be subject to change in the future:
- Sequential read speeds were just slightly above average, coming in at the 43rd percentile.
- Sequential write speeds were above average, coming in at the 70th percentile.
- Random read speeds were slightly below average, coming in at the 37th percentile.
- Random write speeds were slightly below average, coming in at the 48th percentile.
Overall, this card’s performance was just…average. On my scoring system, this card scored a 0.01. (My scoring system puts the average of all cards at 0 — so the closer to 0 a card is, the closer it is to average.)
These cards carry the Class 10, U3, V30, and A1 performance marks. Measurements were good enough for it to qualify for the Class 10, U3, and V30 marks, but not the A1 mark. However, I’ll throw in my standard disclaimer: my performance testing methods do not align with those prescribed by the SD specification — it’s entirely possible that it would have done better had it been tested under proper conditions.
I’ll note that Dr. Gough did a review of this card a couple years ago. How did his results compare to mine?
- Interestingly, the laser etching on the back of the card is completely different. This could mean that Patriot sourced these cards from multiple manufacturers, and mine came from a different manufacturer than his did.
- His card was far less skimpy than mine — his came at 0.14% skimp (as opposed to 5.48% for mine).
- His card had manufacturer ID
9f. I don’t know exactly who this manufacturer ID is assigned to, but I suspect it might belong to Silicon Power. - Sequential read and random write performance was pretty comparable. However, my sample got better sequential write performance, whereas Dr. Gough’s sample got better random read performance.
On endurance:
- Sample #1 has survived 4,939 read/write cycles so far and has not yet experienced any errors.
- Samples #2-#5 are still in the package, waiting to be tested.
November 8, 2025