If you've ever asked an ITAD vendor "you shred drives, right?" and received a vague "yes" — without any follow-up question about whether those drives are HDDs or SSDs — there's a real chance some of your retired media wasn't actually destroyed in a defensible way. Solid-state drives need fundamentally different handling than magnetic hard drives, and the shred specs that work for a spinning platter can leave intact, recoverable NAND flash chips sitting in the metal-fragment pile.
This piece is for engineers, IT directors, and compliance leads in Connecticut who want to understand why SSD destruction is its own discipline. We'll cover how NAND flash storage actually works, why degaussing does nothing to it, why a 25mm hard-drive shredder may not be adequate for SSDs, what NIST 800-88 says about particle size for solid-state media, the form-factor problem (M.2, NVMe U.2, eMMC, BGA), and where firmware-level secure erase fits in.
How Magnetic HDDs Store Data — And Why Destroying Them Is Easy
A traditional hard disk drive stores data as magnetic polarity transitions on a rotating platter coated with a ferromagnetic material. The read/write head flies nanometers above the surface and detects or modifies the magnetization of tiny domains. The data is physical, continuous, and located in geometric positions that can be addressed by cylinder, head, and sector — or in modern terms, by Logical Block Address mapped to a known physical location.
Because the data is held as magnetic state on a continuous surface, two destruction methods work well:
- Degaussing exposes the platter to a strong magnetic field, scrambling the magnetic domains beyond any ability to read. A drive degaussed at appropriate field strength is effectively wiped at the physical level.
- Shredding reduces the platter to fragments small enough that no realistic adversary can reassemble them or read the surviving fragments. Standard hard-drive shredders produce 20-25mm fragments. For HDDs, that's more than adequate.
A shredded HDD also loses its read/write head, its servo electronics, and its spindle motor — so even if a fragment retained surface magnetization, there's no realistic path back to the data without forensic-laboratory-level investment that, in most threat models, isn't credible.
How SSDs Store Data — And Why That Changes Everything
A solid-state drive holds data in NAND flash memory cells. Each cell is a floating-gate transistor that stores electrons in an isolated gate; the presence or absence (or, in MLC/TLC/QLC drives, the precise quantity) of trapped electrons represents the data. The cells are arrayed in pages (typically 4-16 KB) grouped into blocks (typically 256 KB to several MB). All of this lives on a small number of NAND flash dies — packaged as chips and soldered to the SSD's circuit board.
From a destruction perspective, three properties of SSDs matter enormously and have no equivalent in HDDs.
1. Magnetism is irrelevant
NAND flash holds data as trapped charge in a transistor gate, not as magnetic polarity. Degaussing an SSD does nothing. You can run a flash drive through the most powerful commercial degausser made and pull it out with every bit of data intact. This trips up some IT teams who, in good faith, applied their old HDD destruction process to a pile of SSDs and got a "destroyed" sticker on media that's still readable.
2. Wear-leveling defeats naïve overwriting
NAND flash cells wear out after a finite number of program/erase cycles (somewhere between hundreds and tens of thousands depending on cell technology). To extend drive life, SSD controllers implement wear-leveling — when an application writes to logical block X, the controller doesn't necessarily write to the same physical cell each time. It picks a cell that has fewer wear cycles and updates an internal map (the Flash Translation Layer, or FTL) so that logical X now points to the new cell.
The implication for data destruction is significant: when you ask the operating system to "overwrite" sensitive data with zeros, the OS sends writes to logical block X. The controller writes those zeros to a fresh cell and orphans the previous cell — but doesn't necessarily erase it. The original sensitive data sits in a now-unmapped cell that's not reachable through normal LBA addressing, but is absolutely recoverable by anyone who reads the NAND chips directly with appropriate equipment.
3. Over-provisioning hides cells from the OS
Most SSDs reserve 7-28% of their NAND capacity as over-provisioning — flash that the controller uses for wear-leveling spare capacity, bad-block remapping, and write-amplification buffering. The OS never sees this capacity. It can't be overwritten by any user-mode tool because the OS doesn't have addresses for it. Sensitive data that ever passed through over-provisioned blocks during normal operation may still be sitting in those cells when the drive is retired.
Why Software Overwriting Often Fails on SSDs
This is where many destruction programs quietly fall apart. A team applies a NIST 800-88-style overwrite tool, the tool reports "100% complete, drive sanitized," everyone signs off. But the tool only had access to LBA-addressable space. Data in over-provisioned cells, in cells orphaned by wear-leveling, or in cells with reallocated bad-block status is untouched and untouchable from user mode.
The defensible software-only approach to SSD sanitization is the drive's own firmware-level secure erase — specifically the ATA Sanitize Device command or the NVMe Format NVM command with the appropriate Secure Erase Settings, both of which instruct the SSD controller itself to erase all NAND cells (LBA-addressable and over-provisioned) and, on drives that implement it, to cryptographically erase the internal media encryption key. Per NIST Special Publication 800-88 Rev. 1, these are the recommended Purge-category sanitization methods for SSDs.
The tradeoffs:
- Vendor implementation varies. NIST 800-88 explicitly acknowledges that the effectiveness of secure-erase commands depends on correct implementation by the drive manufacturer. Older or budget-tier SSDs have shipped with broken or partial implementations.
- Failed or degraded drives may not respond to secure-erase commands at all. A drive that won't boot, won't enumerate, or returns errors during the sanitize sequence can't be purged in software — destruction becomes the only option.
- Verification is hard. Unlike an HDD overwrite where you can read back the LBA space and confirm zeros, an SSD secure erase happens inside the controller. You're trusting the firmware reported the truth.
- The audit trail. Cryptographic erase is fast and effective on self-encrypting drives, but proving to an auditor that every drive in a fleet had a successful crypto erase requires logging discipline that many ITAD vendors don't bother with.
For high-sensitivity data — regulated health records, financial records, research IP, classified-adjacent material — most compliance teams skip the secure-erase question and go straight to physical destruction. The economics are different from HDDs (SSDs depreciate faster and have less commodity scrap value), so the value-recovery argument for wiping is weaker.
The 25mm Shred Problem
Here's where SSD destruction differs most sharply from HDD destruction in practice. A standard industrial hard drive shredder produces fragments around 20-25mm. For a magnetic platter, that's overkill — no meaningful data recovery from a 20mm fragment is realistic.
For an SSD, that fragment size is potentially inadequate. NAND flash chips on a typical 2.5" SATA SSD are roughly 12-16mm on a side; M.2 NVMe SSDs use even smaller packages. A 25mm shred can easily leave intact NAND chips in the fragment stream. An attacker who recovered intact chips from the scrap pile could, with the right equipment, read the data directly off them — bypassing the SSD controller, the encryption layer (if it wasn't cryptographically erased), and any of the protections the drive provided in life.
This isn't a theoretical concern. NIST 800-88 Annex A specifies that for physical destruction of solid-state media, the recommended particle size is much smaller than for magnetic media. Industry practice has converged on shredders that produce particles of 2mm or smaller for SSDs, which guarantees that every NAND die is broken into multiple non-contiguous fragments. Some specifications go finer.
The practical implication: an SSD destruction program needs the right equipment. If a vendor processes your SSDs through the same shredder they use for HDDs and hands you a Certificate of Destruction, ask what particle size was achieved. If the answer is "20mm" or "1 inch" or "industry standard," that may be fine for HDDs and inadequate for the SSDs in the same batch.
Form Factors That Complicate Things
"SSD" is shorthand for at least a dozen different physical form factors, each with its own destruction implications.
2.5" SATA SSDs
The closest cousin to a traditional HDD by appearance. Same outer dimensions, same SATA interface. Inside, a circuit board with NAND chips, a controller, and DRAM cache (or HMB-based controllers without DRAM). These respond to ATA Sanitize Device commands; for physical destruction, the case must be cut and the board reduced to small enough fragments that no NAND chip survives intact.
M.2 SSDs (SATA and NVMe)
The gum-stick form factor used in laptops and small-form-factor desktops since roughly 2015. M.2 drives are essentially a bare circuit board with no enclosure. The drives themselves are tiny (22mm wide, 30-110mm long), and the NAND packages on them are smaller still. M.2 NVMe drives respond to NVMe Format with Secure Erase; for physical destruction, the small drive size means that a coarse shredder may pass the drive through without sufficient size reduction. Purpose-built SSD shredders are designed to size-reduce these tiny drives appropriately.
U.2 and U.3 NVMe SSDs
Enterprise-class drives, common in newer servers. 2.5" form factor externally, NVMe internally. Often used in mixed-vintage data centers — coexisting with traditional SAS HDDs in the same chassis. Treat as enterprise NVMe; physical destruction requires SSD-grade particle size.
Enterprise EDSFF (E1.S, E1.L, E3)
The ruler and pull-tab form factors increasingly common in current-generation hyperscale-style servers. Higher capacity per drive, denser arrays. From a destruction perspective, same NAND-flash concerns as smaller form factors — needs SSD-appropriate shredding.
eMMC and embedded NAND
This is the form factor most often missed in an ITAD program. Many tablets, point-of-sale terminals, thin clients, smart displays, network appliances, and IoT devices use eMMC or BGA-soldered NAND directly on the device's main board. There's no removable drive to send to destruction. The entire device — or at least the main board — has to be physically destroyed, because the storage chip is permanently soldered in place. If your ITAD program treats these devices as "no drive, nothing sensitive to worry about," sensitive data may be leaving your building inside intact tablets.
Mobile devices and embedded flash
Phones, tablets, and many connected devices contain UFS or eMMC flash storage chips soldered to the logic board. Modern smartphones store data encrypted with a hardware-bound key, and a properly performed factory reset cryptographically erases the key, rendering the data unrecoverable. But for devices that weren't encrypted, or that suffered key-management failures, or where the threat model is high enough to require defense in depth, physical destruction of the logic board is the conservative answer.
What a Defensible SSD Destruction Program Looks Like
If you're auditing or building an SSD destruction program in Connecticut, the elements to insist on:
- Media-aware sorting. The destruction vendor identifies HDD vs SSD per drive — visually, by part number, or by interrogating the drive's controller — and routes each to the appropriate destruction process.
- SSD-appropriate particle size. Shredders rated for SSD destruction producing particles in the 2mm range or smaller, consistent with NIST 800-88 guidance for solid-state media. Ask. Documents like a Certificate of Destruction should reference the destruction method and particle spec.
- Coverage of embedded NAND. Tablets, point-of-sale, thin clients, mobile devices — whatever in your fleet has embedded flash needs a destruction path. Don't let these slip through as "no drive, no data."
- Documented chain of custody. Serial numbers tracked from your facility to the shredder, with the destruction event timestamped, witnessed (if on-site), and recorded.
- R2v3-certified downstream processing. The shredded material — including the NAND chip fragments — goes through certified downstream recyclers, not informal export channels.
Where This Fits with Hard Drive Shredding
Most real ITAD projects involve a mix of HDDs and SSDs in the same batch. A server retirement might involve SAS HDDs in the storage bays and NVMe SSDs in boot slots. A laptop fleet refresh might span machines with SATA SSDs from 2018 and M.2 NVMe drives from 2024. The right destruction program handles both correctly — HDDs through standard hard drive shredding, SSDs through SSD-appropriate destruction at finer particle size — under a single project with unified chain-of-custody documentation. We cover the underlying SSD destruction service in detail on the services side, and the broader data destruction overview for context. If you're still weighing destruction versus software-based methods, see our companion piece on hard drive shredding versus data wiping.
The Short Version
SSDs are not hard drives. The storage physics, the controller architecture, the form factors, and the destruction-particle requirements are all different. A program that treats them identically is a program with a hole in it. If you're not sure how your current vendor handles the SSDs in your fleet, ask three questions: how do you distinguish HDDs from SSDs in intake, what particle size do your SSD shredders produce, and how do you handle devices with embedded NAND. The answers will tell you everything you need to know.
High Tide Management has been handling Connecticut data destruction for more than 25 years. We're R2v3 certified, we run separate destruction paths for HDD and SSD media, and we document chain of custody at the serial-number level. Contact us or call (203) 687-9370 to talk through your specific SSD destruction needs.