960 GB 2.5-inch U.2 NVMe SSD, PCIe 4.0, ordering codes MZQL2960HCJR-00A07 / -00W07 / -00B7C
| Capacity | 960 GB |
|---|---|
| Interface | NVMe |
| Form Factor | U.2 7mm |
| PCIe Version | PCIe 4.0 |
| NAND Type | TLC |
| Endurance (DWPD) | 1 |
| TBW | 1,752 TB |
| Sequential Read | 6,500 MB/s |
| Sequential Write | 1,500 MB/s |
| Random Read IOPS | 580K IOPS |
| Random Write IOPS | 70K IOPS |
| Power | 9.5 W |
| DRAM Cache | Yes |
| Power Loss Protection | Yes |
| Encryption | AES-XTS 256 |
| Condition | New · Used |
Samsung PM9A3 960 GB, 2.5-inch U.2 at 7 mm, catalogue code MZQL2960HCJR, orderable as -00A07, -00W07 and -00B7C. 6,500 MB/s sequential read and 1,500 MB/s write on PCIe 4.0 x4, 580K random read and 70K random write IOPS, 1,752 TBW at 1 DWPD. TLC V-NAND, capacitor-backed PLP, AES 256-bit full disk encryption, 9.5 W reading and 8.0 W writing. New and used both stocked.
Nine hundred and sixty gigabytes is the entry point of the PM9A3 U.2 range in Samsung's own data sheet, and it is the capacity in this family that has aged the least. Every drive generation since 2015 has improved sequential read by roughly doubling the link, and at this capacity that doubling lands on a volume small enough for a boot or cache role, where an eleven minute clone instead of a thirty minute one changes how maintenance windows are planned. The rest of this page is about where that improvement stops.
| Product family | Samsung PM9A3 (PCIe 4.0 data centre NVMe SSD, 1 DWPD) |
|---|---|
| PN | MZQL2960HCJR (catalogue code) |
| Ordering codes | MZQL2960HCJR-00A07 / MZQL2960HCJR-00W07 / MZQL2960HCJR-00B7C |
| Capacity | 960 GB (unformatted; 1 GB = 1 billion bytes, formatted capacity is less) |
| Form factor | 2.5-inch U.2, 7 mm (100.20 × 69.85 × 7.0 mm) |
| Interface | PCIe 4.0 x4, NVMe 1.4, single port |
| NAND | Samsung V-NAND TLC |
| Sequential read | 6,500 MB/s (up to, 4 KB transfer size, queue depth 32, four workers) |
| Sequential write | 1,500 MB/s |
| Random 4K read | 580K IOPS (queue depth 32) |
| Random 4K write | 70K IOPS (queue depth 32) |
| Endurance (TBW) | 1,752 TB |
| Endurance (DWPD) | 1 drive write per day for five years |
| Power | 9.5 W active read, 8.0 W active write, 3.5 W idle with DIPM off |
| Power-loss protection | Yes, capacitor-backed, with end-to-end data path protection |
| Encryption | AES 256-bit full disk encryption, TCG Opal 2.0, IEEE 1667 encrypted drive |
| MTBF | 2.0 million hours |
| Environment | 0 to 70 °C operating, −40 to 85 °C non-operating, 5 to 95% non-condensing |
| Features | TRIM, garbage collection, S.M.A.R.T. |
| Model | Capacity | Interface | Seq read / write | Random 4 KB | TBW | Power |
|---|---|---|---|---|---|---|
| PM953 | 960 GB | PCIe 3.0 x4 | 1,000 / 870 MB/s | 240K / 19K | 1,400 TB | 9 W |
| PM963 | 960 GB | PCIe 3.0 x4 | 1,800 / 930 MB/s | 350K / 30K | 1,366 TB | 7.5 W |
| PM983 | 960 GB | PCIe 3.0 x4 | 3,200 / 1,100 MB/s | 400K / 40K | 1,366 TB | 8.7 W |
| PM9D3a | 960 GB | PCIe 5.0 x4 | 12,000 / 1,600 MB/s | 1,000K / 70K | 1,752 TB | 11 W |
| PM9A3 | 960 GB | PCIe 4.0 x4 | 6,500 / 1,500 MB/s | 580K / 70K | 1,752 TB | 9.5 W |
Every drive in that table is stocked in the same 2.5-inch U.2 carrier and every one of them is still in service somewhere. The sequential read column is the clearest measure of what four generations bought: 1,000 MB/s on the Gen3 PM953, 3,200 on the Gen3 PM983, 6,500 here on Gen4, 12,000 on the Gen5 part. Nothing in the flash changed enough to explain a twelve times improvement; the link did, one lane generation at a time.
The write column tells a shorter story, and the last two rows are the reason this page exists. Between Gen4 and Gen5 at 960 GB, sequential write moves from 1,500 to 1,600 MB/s and random write does not move at all, staying at 70K on both. Read meanwhile roughly doubles. That asymmetry is not a testing artefact.
Write bandwidth is set by how many dies the controller can program at once, and 960 GB of TLC is about fifteen dies. Fifteen dies cannot saturate four Gen5 lanes no matter what the controller is rated for, so the extra link speed goes entirely into reads, which are served from flash that is already on the die and do not need to be spread across it. On the same die count, Gen4 and Gen5 write the same number.
That is worth stating plainly for anyone choosing between the two: on a Gen4 host, a Gen5 drive at this capacity buys you a faster read path you cannot use, because the host link caps it, and no write advantage at all. If the job is reading, go up a generation. If the job is writing, this capacity is the wrong place to spend the difference — buy capacity instead, where die count does scale with the money.
Power is the one column where Gen4 is quietly ahead: 9.5 W against 11 W at the same capacity, and 228 W against 264 W for twenty-four bays. In a chassis where every watt is already allocated, that difference is one more drive's worth of headroom.
One drive write a day for five years on a 960 GB volume is 960 GB of host writes a day, which averages 11 MB/s. At the rated 1,500 MB/s write, a full fill of the drive takes eleven minutes. So the endurance rating is not a constraint on any workload that writes in bursts and then reads: it permits one complete rewrite every single day for the warranty period, and the rewrite takes under a fifth of an hour.
The constraint appears only under sustained random writes. At 70K IOPS of 4 KB writes the drive absorbs 287 MB/s, and at that rate 1,752 TB is exhausted in about seventy-one days. Nobody runs a volume at its random write ceiling for five years, but a database log or a write-ahead file that does so for a month will consume several percent of the rating without anyone noticing. That is the workload to keep off this drive, and it is a much narrower exclusion than the one that applies to QLC parts at four times this capacity.
Samsung lists one dimension for the whole PM9A3 U.2 range, 100.20 by 69.85 by 7.0 mm, so the height does not change with capacity and the drive fits every 2.5-inch bay including the 7 mm-only caddies that some 1U chassis use. It will sit in a 15 mm bay as well, because those bays accept the thinner drive; the reverse is not true, which is why a 15 mm part is a purchase decision rather than a default.
The mechanical freedom is exactly why this capacity should not be used for a capacity tier. Twenty-four bays of 960 GB hold 23 TB and draw 228 W, which is 9.9 W per terabyte; the same bays filled with the 15.36 TB part in this family hold sixteen times the data and draw 264 W against 228. Use this drive where the bay is already paid for and the data is small, and do not use it to build a tier.
| Code | Type | Capacity | Condition | Quote |
|---|---|---|---|---|
| MZQL2960HCJR | Catalogue code | 960 GB | New / used | Enquire |
| MZQL2960HCJR-00A07 | Ordering code | 960 GB | New / used | Enquire |
| MZQL2960HCJR-00W07 | Ordering code | 960 GB | New / used | Enquire |
| MZQL2960HCJR-00B7C | Ordering code | 960 GB | New / used | Enquire |
The suffix records the distribution block and the build, not the capacity or the performance, and all three strings describe one drive. -00W07 is the code Samsung prints in its own data sheet part list; -00A07 and -00B7C are the same drive shipped through different channels and appear on more purchase orders than the data sheet code does. Quote whichever one the customer's system holds, and check that the packing list matches it.
The prefix carries the meaning that matters at goods-in. MZQL is the 2.5-inch U.2 carrier in this family; MZ1L at the same capacity is the M.2 22110 version and MZQLV is the earlier PM953 shell. The density field reads 960 here, and 1T9, 3T8, 7T6 or 15T further up the range.
Cover runs for one year on sealed drives and tested pulls alike, and it follows the drive rather than the chassis it came out of. Used units are tested, wiped and graded before packing, with the SMART record kept against the serial and available on request. Warehouse stock ships in three to seven working days and tracking is issued at dispatch. Two pieces is the minimum on a line, capacities in this family combine on one order, and a quotation is returned the same working day once part number, condition and quantity are confirmed.
Tell us who you are — we reply within one working day with a firm quote, stock status and lead time. Volume discounts available.
15.36 TB E1.L 1U long ruler NVMe SSD, PCIe 4.0, ordering code MZEL215THBLA-00AMV
7.68 TB E1.S 1T NVMe SSD, PCIe 4.0, ordering codes MZCL27T6HBLA-00AMV / MZTL27T6HBLA-00A07
3.84 TB E1.S 1T NVMe SSD, PCIe 4.0, ordering codes MZCL23T8HCLS-00AMV / MZTL23T8HCLS-00A07
1.92 TB E1.S 1T (9.5 mm) NVMe SSD, PCIe 4.0, ordering code MZCL21T9HCJR-00AMV