How SSDs Work

Inside every hard drive you’ve ever used, a metal platter spins at 5,400 or 7,200 revolutions per minute while an actuator arm physically moves a magnetic read/write head across the surface to find your files. It’s a brilliant piece of engineering from 1956, but it’s also fundamentally mechanical, and mechanical things wear out, break, and are slow. The solid-state drive (SSD) threw all of that out. No moving parts. No waiting for a disk to spin to the right position. Just electrons flowing through silicon. That’s the difference.

The short answer

An SSD stores data in flash memory chips instead of on a spinning magnetic platter. When you save a file, the SSD uses electrical charges to trap electrons in tiny cells within these chips. When you read that file, the SSD measures the charge in those cells to recover the data. Because there’s no mechanical movement involved, SSDs can access data in microseconds rather than milliseconds. The result is a storage device that’s 10 to 100 times faster than a traditional hard drive, with no moving parts to wear out or break.

The full picture

NAND flash: the building block

The technology inside every SSD is called NAND flash memory, named after the NAND (NOT-AND) logic gate that performs the underlying boolean operations. Unlike regular computer memory (RAM), which loses everything when you turn off the power, NAND flash is non-volatile, it retains data even when disconnected.

NAND flash stores bits in floating-gate transistors, essentially tiny electrical traps that can hold or release electrons. Each transistor has a “floating gate” surrounded by insulation. When you apply a voltage, electrons tunnel through the insulation and get trapped in the floating gate. That trapped charge represents a stored bit. Apply a different voltage, and the electrons escape. The presence or absence of electrons becomes the 1 or 0 that computers understand.

These transistors are arranged in a grid pattern on a silicon chip, with millions or billions of them packed incredibly densely. A single modern NAND chip can hold hundreds of billions of bits.

From bits to cells: SLC, MLC, TLC, and QLC

Not all NAND flash is created equal. The industry categorizes NAND by how many bits each cell can store:

SLC (Single-Level Cell) stores 1 bit per cell. It’s the simplest and fastest, but also the most expensive per gigabyte. You’ll find it in enterprise storage where speed matters more than cost.

MLC (Multi-Level Cell) stores 2 bits per cell. Good performance, moderate cost. Becoming less common in consumer products.

TLC (Triple-Level Cell) stores 3 bits per cell. The standard for most consumer SSDs today. Slower than SLC or MLC because distinguishing between 8 different charge levels is harder, but dramatically cheaper to manufacture.

QLC (Quad-Level Cell) stores 4 bits per cell. The newest type, pushing toward 5 bits (PLC). Maximum capacity at minimum cost, but with tradeoffs in endurance and speed.

The progression from SLC to QLC is essentially a cost-saving story. As manufacturers figured out how to pack more bits into each cell, the price per gigabyte dropped. Modern consumer SSDs are affordable precisely because they use TLC or QLC NAND.

How an SSD reads and writes

When you save a file to an SSD, the controller breaks it into pages, typically 4 KB to 16 KB each. These pages get written to empty cells on the NAND chips. But here’s the catch: NAND flash can only write to empty cells. It can’t overwrite existing data in place.

This creates a problem called write amplification. When you modify a file, the SSD has to mark the old pages as invalid, find new empty cells, write the modified data there, and update its mapping table. One logical write from your computer might result in multiple physical writes inside the SSD. This is why SSDs need spare capacity, the drive needs free cells to write to.

Over time, as an SSD fills up and valid data becomes scattered across the drive, the controller performs garbage collection: it reads all the valid data from a block, writes it to new empty cells, and erases the old block to make it writable again. This happens in the background, but it can cause occasional slowdowns on drives that are nearly full.

The controller: the brain of the SSD

Every SSD has a dedicated processor called a controller that manages all these operations. The controller handles:

• Wear leveling: NAND cells can only be written to a limited number of times before they fail (this is called endurance). The controller distributes writes across all cells evenly so no single area wears out prematurely.
• Error correction: As cells shrink and store more bits, errors become more common. Modern controllers use sophisticated error correction codes (ECC) to detect and fix errors.
• Mapping: The controller maintains a mapping table that translates the logical addresses your computer requests (like “sector 1,000,000”) to the physical locations where the data actually lives on the NAND chips.
• Encryption: Many SSDs include hardware encryption, protecting your data even if the drive is removed from your computer.

The controller is why two SSDs with the same NAND chips can perform very differently. A good controller can make a TLC drive feel almost as fast as SLC; a poor controller can bottleneck even the best NAND.

The interface: SATA vs. NVMe

How an SSD connects to your computer matters as much as what’s inside it.

SATA (Serial ATA) is the same interface used by old hard drives. It’s limited to about 550 MB/s, fast compared to a hard drive, but a fraction of what SSDs can actually do. SATA SSDs are still common and affordable, fine for everyday tasks.

NVMe (Non-Volatile Memory Express) is a purpose-built interface for SSDs that uses the high-speed PCI Express bus. A typical NVMe SSD can read at 3,500 to 7,000 MB/s, roughly 10 times faster than SATA. The latest PCIe 5.0 SSDs push beyond 10,000 MB/s. NVMe also reduces latency because it communicates directly with the CPU rather than going through a legacy SATA controller.

For most users, any SSD feels dramatically faster than a hard drive. But if you’re moving large files, editing video, or running databases, NVMe makes a noticeable difference.

Endurance: the wearing out problem

NAND flash cells degrade every time they’re written to. The insulation around the floating gate slowly breaks down from the electron tunneling, until eventually the cell can no longer reliably hold a charge. This is measured in terabytes written (TBW), how much data you can write to the drive before it exceeds its designed lifespan.

Consumer-grade TLC SSDs typically rate around 300-600 TBW. That sounds small, but it means you could write 100 GB per day for 8-16 years before hitting the limit. For most users, the SSD will become obsolete long before it wears out. Enterprise SLC drives can handle petabytes of writes.

Modern controllers and improved NAND manufacturing have made endurance less of a concern for consumer use. You’re more likely to lose the drive to a hardware failure or upgrade than to wear it out through normal use.

Why it matters

SSDs transformed computing. The difference between a computer with a hard drive and one with an SSD is the difference between tolerable and joyful. Boot times drop from minutes to seconds. Applications launch instantly. Copying files goes from waiting to done.

But the shift matters beyond just speed. Without moving parts, SSDs are lighter, use less power, generate no noise, and are far more resistant to shock and vibration. Laptops became thinner and more durable. Data centers became more reliable. The smartphone in your pocket uses flash memory for the same reasons.

The technology continues evolving. QLC drives are pushing capacities higher at lower prices. PCIe 5.0 is pushing speeds further. And emerging technologies like 3D V-Cache (stacking NAND layers vertically) and PLC (5 bits per cell) will continue driving down costs while increasing capacity. The hard drive isn’t dead yet, magnetic storage still wins on cost per terabyte for cold storage, but for everything you actually use, SSDs won.

Common misconceptions

SSDs don’t last as long as hard drives. Not true. While SSDs have a finite number of writes, modern SSDs last far longer than most users actually keep their devices. Most SSDs will outlast their useful technological lifespan (becoming obsolete) before they wear out. The bigger risk with hard drives is mechanical failure, which is far more common than NAND wearing out.

You need to defragment SSDs. Don’t do this. Defragmentation is for hard drives, where moving data physically reduces read times. On an SSD, all cells access at the same speed, and defragmentation just wastes write cycles, reducing the drive’s lifespan.

SSDs are immune to data loss. They’re not. SSDs can fail, and the NAND cells can lose charge over time (typically years) if the drive is disconnected from power. For long-term archival, hard drives are still sometimes preferred for their stability when powered off. Both storage technologies need backups.

More expensive SSDs are always faster. Not always. The interface (NVMe vs. SATA) matters more than the price. A budget NVMe SSD is usually faster than a premium SATA SSD. Within the same interface standard, the differences are smaller and often unnoticeable for everyday tasks.