What Does the Samsung 990 Pro Firmware Bug Do?
The Samsung 990 Pro launched with firmware version 0B2QJXD7, and that launch firmware is where the health-decline bug lives. Reported drive health falls fast on a drive that has barely been written to.
What Should You Avoid Doing with a Degraded 990 Pro?
Do not apply Samsung's firmware update if your drive is already at 0% health.
The update process writes to the drive's firmware area, and a drive you need data off should be imaged before anything writes to it. The update does not reverse health already recorded, so it provides no benefit to drives already at 0%.
Do not run data recovery software on a 990 Pro that shows 0 bytes or is not detected.
If the drive is not visible as a block device, no software tool can access it. Software recovery only works when the drive presents its capacity to the operating system.
Do not power-cycle the drive repeatedly hoping it will come back.
Each power cycle forces the controller to attempt firmware initialization. If the firmware modules are corrupted, repeated initialization attempts can deepen the corruption. Leave the drive powered off until it reaches a recovery lab.
Do not use Samsung Magician's Secure Erase or Over Provisioning tools.
These features write to the drive. On a drive with firmware-level instability, writing to the NAND risks corrupting the flash translation layer. Magician cannot fix the SMART reporting bug; only Samsung's firmware update addresses the counter logic.
Which Samsung 990 Pro Models Are Affected?
Every 990 Pro variant shipping with firmware 0B2QJXD7 is susceptible.
| Capacity | Standard Model | Heatsink Model |
|---|---|---|
| 1TB | MZ-V9P1T0BW | MZ-V9P1T0CW |
| 2TB | MZ-V9P2T0BW | MZ-V9P2T0CW |
The 990 EVO uses Samsung's Piccolo controller (S4LY022), a different design from the controller in the 990 Pro.
Samsung drives other than the 990 Pro, along with Galaxy phones and SpinPoint hard drives, are covered by Samsung data recovery for Galaxy phones, SSDs and hard drives.
What Firmware Updates Has Samsung Released for the 990 Pro?
Only the first update (1B2QJXD7) addresses the SMART degradation directly. Samsung's own Tool & Software Download page describes the later releases it lists as addressing high temperatures logged on Samsung Magician, intermittent non-recognition and blue screens, and read-operation stability.
| Version | Purpose |
|---|---|
0B2QJXD7 | Launch firmware. Contains the SMART calculation bug. |
1B2QJXD7 | Stops accelerated SMART decline. Does not reset counters. |
3B2QJXD7 | A later separate release. Not the health fix. |
4B2QJXD7 | December 2024. Addresses reports of high temperatures logged on Samsung Magician. |
7B2QJXD7 | September 2025. Addresses the intermittent non-recognition and blue screen issue. |
8B2QJXD7 | December 2025. Improves read-operation stability. |
The 1B2QJXD7 update stops the erroneous decline. It does not restore the health value a drive has already recorded.
How to Read Your Current 990 Pro Firmware Revision
Your drive reports its own firmware revision. The value lives in the NVMe Identify Controller data structure, and you match the string it returns against the table above.
Three categories of tool read that value:
- Vendor utility. Samsung Magician reads the active revision off the drive.
- General SMART reader. CrystalDiskInfo displays the firmware string.
- NVMe command-line tooling. nvme-cli queries the Identify Controller structure directly with its nvme id-ctrl command.
Compare the full eight characters, not a partial match. The six revisions in the table above differ in their leading character alone, so reading the last seven characters & stopping there doesn't tell you which build is on the drive.
What the revision means for you depends on the drive's state, not on the number by itself. A healthy drive still reporting normal health is what the update targets. A drive whose health has already fallen gets that figure frozen where it stands rather than restored; the math is corrected going forward but the recorded health figure stays put.
A drive locked read-only, or one that won't enumerate, should be imaged before anything writes to it. The update is a write operation, so that drive belongs in NVMe SSD data recovery before it goes near a firmware updater.
Pricing
Samsung 990 Pro firmware degradation recovery: published NVMe SSD recovery tiers below.
Free evaluation, firm quote before paid work, no data recovered = no charge.
Low complexity
Simple Copy
Your NVMe drive works, you just need the data moved off it
Functional drive; data transfer to new media
Rush available: +$100
$200
3-5 business days
Low complexity
File System Recovery
Your NVMe drive isn't showing up, but it's not physically damaged
File system corruption. Visible to recovery software but not to OS
Starting price; final depends on complexity
From $250
2-4 weeks
Medium complexity
Circuit Board Repair
Your NVMe drive won't power on or has shorted components
PCB issues: failed voltage regulators, dead PMICs, shorted capacitors
May require a donor drive (additional cost)
$900–$1,200
3-6 weeks
Medium complexity
Most Common
Firmware Recovery
Your NVMe drive is detected but shows the wrong name, wrong size, or no data
Firmware corruption: ROM, modules, or system files corrupted
Price depends on extent of bad areas in NAND
$900–$1,200
3-6 weeks
High complexity
PCB / NAND Swap
Your NVMe drive's circuit board is severely damaged and requires NAND chip transplant to a donor PCB
NAND swap onto donor PCB. Precision microsoldering and BGA rework required
50% deposit required; donor drive cost additional
50% deposit required
$1,200–$2,500
4-8 weeks
Hardware Repair vs. Software Locks
Our "no data, no fee" policy applies to hardware recovery. We do not bill for unsuccessful physical repairs. If we replace a hard drive read/write head assembly or repair a liquid-damaged logic board to a bootable state, the hardware repair is complete and standard rates apply. If data remains inaccessible due to user-configured software locks, a forgotten passcode, or a remote wipe command, the physical repair is still billable. We cannot bypass user encryption or activation locks.
No data, no fee. Free evaluation and firm quote before any paid work. Full guarantee details. NAND swap requires a 50% deposit because donor parts are consumed in the attempt.
- Rush fee
- +$100 rush fee to move to the front of the queue
- Donor drives
- A donor drive is a matching SSD used for its circuit board. Typical donor cost: $40–$100 for common models, $150–$300 for discontinued or rare controllers.
- Target drive
- The destination drive we copy recovered data onto. You can supply your own or we provide one at cost plus a small markup. All prices are plus applicable tax.
Drives still in read-only mode that enumerate and respond to NVMe commands typically fall in the simple-copy and file-system tiers. Drives that have entered a firmware panic or BSY state requiring board-level diagnostics and controller stabilization fall in the circuit-board-repair and firmware-recovery tiers. Compare to industry-wide data recovery pricing.
Data Recovery Standards & Verification
Our Austin lab operates on a transparency-first model. We use industry-standard recovery tools, including PC-3000 and DeepSpar, combined with strict environmental controls to maintain drive integrity. This approach allows us to serve clients nationwide with consistent technical standards.
Localized Clean Zone
Open-drive work is performed in a 0.02 micron ULPA-filtered laminar clean bench.
Transparent History
Serving clients nationwide via mail-in service since 2008. Our lead engineer holds PC-3000 and HEX Akademia certifications for hard drive firmware repair and mechanical recovery.
Media Coverage
Our repair work has been covered by The Wall Street Journal and Business Insider, with CBC News reporting on our pricing transparency. Louis Rossmann has testified in Right to Repair hearings in multiple states and founded the Repair Preservation Group.
Aligned Incentives
Our "No Data, No Charge" policy means we assume the risk of the recovery attempt, not the client.
Technical Oversight
Louis Rossmann
Our engineers review all lab protocols to maintain technical accuracy and honest service. Since 2008, his focus has been on clear technical communication and accurate diagnostics rather than sales-driven explanations.
We believe in showing the bench rather than just describing it. Open-drive work runs on a 0.02 micron ULPA-filtered laminar clean bench, and we filmed it.
See the particle counter test at the benchHow Is Data Recovered from a Samsung 990 Pro?
Samsung's in-house ARM-based NVMe controller in the 990 Pro does not appear on ACELab's PC-3000 SSD supported list, so recovery leans on board-level work, and Samsung's proprietary NAND encoding rules out chip-off shortcuts.
- Board-Level Diagnostics:
Before attempting firmware-level access, we verify the drive's power delivery is stable. The 990 Pro's M.2 2280 PCB has voltage regulators that supply the controller and NAND. A failed PMIC or shorted capacitor can prevent initialization independently of the firmware bug. We probe the 3.3V rail, controller core voltage, and NAND I/O lines to confirm the hardware is functional before addressing firmware-level issues.
- Controller Stabilization:
On 990 Pro drives, PC-3000 SSD has no technological-mode loader or FTL-rebuild support, so stabilization is board-level work on the drive's original PCB to get it into a state where it responds to read commands.
- Data Imaging:
Once the controller is stabilized and the drive responds to NVMe read commands, we image the entire drive sector-by-sector using PC-3000's Data Extractor. The imaging runs without a controller-specific utility, because none exists. Drives in BSY state require more careful handling, with imaging paused if the controller shows signs of instability.
Why Chip-Off Does Not Work on the 990 Pro
Samsung uses proprietary NAND encoding and hardware-level self-encrypting drive (SED) technology on the 990 Pro. The controller encrypts all data written to NAND using keys stored in the controller's secure area. Removing the NAND chips and reading them on external hardware produces encrypted, scrambled data that cannot be decoded without the controller's encryption keys and Samsung's proprietary descrambling algorithm. Controller-level access is the only viable recovery path.
Why Can't Data Recovery Software Fix a Firmware-Degraded Samsung 990 Pro?
A read command from Disk Drill or R-Studio does not reach the NAND directly. It travels a fixed path down the host software stack, crosses one hardware boundary, then hands off to the controller. Everything a consumer tool can do lives on the host side of that boundary. Everything a firmware-degraded 990 Pro breaks lives on the controller side.
- 1. Application syscall
- An application, or a recovery tool, issues a standard write() or read() syscall. The operating system kernel's block layer receives it.
- 2. The NVMe driver builds the command
- The block layer hands the request to the NVMe driver, which assembles a 64-byte NVMe command describing the logical blocks to read.
- 3. Submission Queue in host DRAM
- The driver places that command into a Submission Queue, a ring buffer that lives in the host system's DRAM, not on the SSD.
- 4. The doorbell ring: the software-to-firmware boundary
- The host executes a PCIe memory-write to ring a doorbell register on the SSD controller. This is the boundary. The host has placed the command in shared memory & poked the controller; from here, execution is the controller's responsibility.
- 5. Controller & FTL take over
- The controller uses DMA to fetch the command from host memory. Its Flash Translation Layer (FTL) maps the logical block address to a physical NAND location, the step that makes raw flash look like a linear block device.
Consumer tools, Disk Drill, R-Studio, PhotoRec, CrystalDiskInfo, all send standard NVMe or ATA commands down this exact pipeline, & every one of them needs a controller that enumerates & answers. A 990 Pro stuck in firmware panic never services the command. One locked to read-only refuses any write. One with a corrupt FTL cannot translate the address to a NAND page. In each case the command dies at the doorbell, at the controller interface, & the software never touches the NAND underneath.
This is also why the answer is not a better piece of software. The 990 Pro controller is Samsung in-house silicon & is absent from ACELab's PC-3000 SSD supported-controller list, so no automated utility rebuilds its FTL from the host side. Reaching the data means reaching the controller below the doorbell: component-level voltage-rail probing & microsoldering on the drive's original board to bring the controller to a state where it answers reads. The original controller has to stay in the loop, because it holds the hardware encryption keys that make the NAND readable at all. Rossmann recovers these drives at the board level; what is unavailable is PC-3000 SSD automated utility support, not the recovery itself.
Working below the command-stack boundary, at the board level rather than through host software, is the same discipline behind our broader SSD data recovery service. It is the reason a lab reaches a firmware-degraded 990 Pro that no downloadable tool can see.
How Does Read-Retry Recovery Pull Data Off a Degraded 990 Pro?
The bug lives in the firmware's SMART math, not in worn cells. On launch firmware 0B2QJXD7 the drive's own health record declines far faster than the writes justify, while the V-NAND cells underneath are healthy with write life remaining.
We write-block the drive first, so no command can alter the firmware area or the flash translation layer while we work, then stabilize the original board until the controller answers read commands.
Once the controller answers read commands, we image the drive block-by-block through PC-3000's Data Extractor. Where a NAND page returns marginal or uncorrectable data, we apply read-retry: re-reading the same page with adjusted read-voltage thresholds so the controller's error correction can resolve cells sitting near a threshold boundary.
The PC-3000 SSD has no technological-mode loader for the 990 Pro, so this stabilization is board-level work on the original drive.
This is not a firmware reflash. A reflash writes a new firmware image to the drive's firmware area, which is the operation Samsung's update performs, and that is a write we will not put on a drive before its data is off.
Read-retry recovery reads what is already on the NAND without rewriting firmware. It is also not chip-off: the SED key binding described above leaves desoldered NAND as ciphertext, so the live original controller is the only path to intelligible data.
Frequently Asked Questions
Why is my Samsung 990 Pro health dropping so fast?
The original firmware version 0B2QJXD7, shipped with the 990 Pro at launch, drives an erroneous accelerated health decline that has no relationship to actual NAND wear. Samsung released firmware 1B2QJXD7 to halt the accelerated decline, but it does not reverse the health already recorded on an affected drive.
Will Samsung's firmware update fix my 990 Pro?
It stops the bleeding but does not reverse the damage. Firmware 1B2QJXD7 halts the erroneous health decline, so the health percentage stops dropping abnormally. If your drive was already at 40% health when you applied the update, it stays at 40%. A degraded drive holding data you need should be imaged before anything writes to it, firmware updates included.
Is my Samsung 990 Pro actually wearing out or is it just a counter error?
The NAND cells are not physically worn. Drives hit by the launch-firmware bug report a health collapse that the write volume does not justify. The wear leveling algorithm and NAND flash are functioning normally; the firmware reports the wrong health.
My 990 Pro disappeared from BIOS. Is this the firmware bug?
It can be. Samsung has published later firmware for the 990 Pro on its own Tool & Software Download page: 7B2QJXD7, September 2025, to address the intermittent non-recognition and blue screen issue, and 8B2QJXD7, December 2025, to improve read-operation stability. If your drive no longer appears in BIOS, consumer firmware tools cannot help. The drive needs professional recovery.
How much does Samsung 990 Pro data recovery cost?
Recovery price depends on the drive's failure state, matching our published NVMe SSD pricing tiers. Drives still partially functional and imageable directly fall in the simple-copy and file-system tiers ($200 and $250). Drives requiring controller-level repair or board-level diagnostics to stabilize a non-enumerating controller fall in the circuit-board-repair and firmware-recovery tiers ($900–$1,200 and $900–$1,200). Free evaluation, firm quote before work begins, no charge if recovery is not successful.
Does updating the firmware recover data already lost on a 990 Pro?
No. The firmware update halts the erroneous decline, but it does not reverse the health already recorded. The update corrects only the math going forward. A critically degraded 990 Pro should be imaged before anything writes to it, an update included.
What does read-retry recovery mean for a degraded NVMe SSD?
Read-retry is re-reading a NAND page that returned marginal or uncorrectable data, using adjusted read-voltage thresholds so the controller's error correction can resolve cells sitting near a voltage boundary. On a degraded 990 Pro, the lab pairs it with board-level work to bring the controller to a stable ready state, then images the drive block-by-block. It reads the data already on the NAND rather than rewriting firmware, so it is distinct from a firmware reflash and from chip-off, which the controller's key binding blocks.
Why can't data recovery software fix a firmware-degraded 990 Pro?
Recovery software issues standard NVMe read commands that travel from the application through the OS block layer to the NVMe driver, into a Submission Queue in host DRAM, and finally ring a doorbell register on the SSD controller. That doorbell ring is the boundary. Past it, execution belongs to the controller and its Flash Translation Layer. When a 990 Pro is in firmware panic, locked to read-only, or carrying a corrupt FTL, the controller either never services the command or refuses the write, so every command the software sends dies at the controller interface and never reaches the NAND. Software recovery needs a controller that enumerates and answers; board-level intervention reaches the controller below that boundary.
Is the Samsung 990 Pro controller supported by standard PC-3000 SSD recovery?
No. The 990 Pro controller is Samsung in-house silicon and is absent from ACELab's PC-3000 SSD supported-controller list, so automated firmware-level and utility-driven FTL rebuilding is not available for it. Recovery is board-level: component-level voltage-rail probing and microsoldering on the drive's original board to stabilize the controller. The original controller has to stay in the loop because it holds the hardware encryption keys; desoldering the NAND (chip-off) yields ciphertext without it. Rossmann recovers these drives at the board level. What is unavailable is PC-3000 SSD automated utility support, not the recovery itself.
Sources
- 1.Samsung Semiconductor, "Tool & Software Download", semiconductor.samsung.com/consumer-storage/support/tools/. Lists 990 PRO firmware 4B2QJXD7 (December 2024), 7B2QJXD7 (September 2025) and 8B2QJXD7 (December 2025) with the issue each release addresses.
- 2.Puget Systems, William George, "Samsung 990 Pro Critical Firmware Update", 14 March 2023, pugetsystems.com/blog/2023/03/14/samsung-990-pro-critical-firmware-update/. "The original firmware on these drives was version 0B2QJXD7, and the fixed version is 1B2QJXD7."
Since 2008
Established
As Featured In
Related services
Related Recovery Services
Galaxy phones, Samsung SSDs, and SpinPoint hard drives
Full SSD recovery service overview
Controller reference for the 990 Pro
Drive vanished from BIOS
Firmware panic and capacity loss
FTL encryption blocks software tools
Samsung 990 Pro showing 0% health?
Free evaluation. Firm quote. No data, no fee. Ship from anywhere in the U.S.