Why SSDs Degrade: TBW, NAND Memory, and Wear Leveling Explained

Modern SSDs have become the standard for data storage - they are fast, silent, energy-efficient, and more reliable than traditional hard drives. However, all flash memory faces a fundamental limitation: it wears out over time. As the total bytes written (TBW) limit is approached, write speed drops, controllers rely more on spare cells, delays increase, and eventually the drive reaches the end of its useful life.

What Causes SSD Degradation?

The reasons behind SSD degradation go beyond just the TBW figure on the box. The lifespan of an SSD is influenced by the type of NAND memory, cell architecture, controller behavior, wear leveling algorithms, garbage collection efficiency, TRIM support, and even how full the drive typically is.

To understand why SSDs slow down over time and what determines their longevity, let's look inside: how NAND memory is structured, what happens to its cells during writes, how controllers fight wear, and why different SSD models age differently.

In this article, we'll thoroughly break down how SSDs degrade, what metrics like TBW and Endurance mean, which algorithms modern drives use to extend their lifespan, and how users can influence device longevity. We'll also touch on the future of non-volatile memory technologies - such as MRAM and RRAM - that promise to overcome many fundamental NAND issues.

How NAND Memory Works: Cells, Pages, and Blocks

To grasp why SSDs lose speed and endurance over time, it's important to understand how their memory cells operate. At the core of every SSD is NAND flash - non-volatile memory that stores data as an electrical charge inside a transistor's floating gate.

Cells

Each NAND cell is a tiny transistor capable of holding a specific amount of charge. The number of charge levels defines the cell type:

SLC (1 bit, 2 states) - fastest and most durable.
MLC (2 bits, 4 states) - balances speed and endurance.
TLC (3 bits, 8 states) - dominates modern consumer SSDs.
QLC (4 bits, 16 states) - highest density, lowest lifespan.

The more states a device must read and write, the more complex and error-prone the process - and the faster the cells wear out.

Pages

Data is written in pages, typically 4-16 KB each. While a page can be written once, it cannot be overwritten directly - it must first be erased.

Blocks

Erasure happens in blocks, which consist of dozens or hundreds of pages (e.g., 256 pages × 16 KB = 4 MB). This is a key NAND feature: to overwrite a single page, the controller must move other data, erase the entire block, and then write new information.

Why is this important?

Accelerated cell wear
Accumulation of "garbage"
Need for wear leveling algorithms
Write delays

The more bits each cell stores (e.g., TLC or QLC), the faster it reaches its rewrite cycle limit.

Why SSDs Degrade: The Physics of Wear and Rewrite Limits

SSD degradation is not a myth or a software bug - it's a direct result of how NAND memory works. Unlike RAM or magnetic disks, each flash memory cell has a strictly limited number of rewrite cycles. After exceeding this limit, a cell loses its ability to hold a charge and data is lost.

The Main Cause: Oxide Layer Wear

Each NAND cell stores charge inside a floating gate, surrounded by a thin dielectric layer. Every write operation pushes electrons through this layer via tunneling. Over time, this causes:

Thinning and damage of the oxide layer
Increasing charge leakage

When leakage becomes too great, the cell can no longer reliably distinguish voltage levels, causing data errors. Initially, the controller compensates by raising voltage or moving data to reserved areas, but eventually the cell is marked as failed.

Bits Per Cell Accelerate Wear

The more states a cell must store, the harder it becomes for the controller to accurately read and write data.

Average rewrite cycle limits:

SLC: 50,000-100,000 cycles
MLC: 3,000-10,000 cycles
TLC: 1,000-3,000 cycles
QLC: 150-500 cycles

This is why TLC and QLC degrade faster: they must hold charge more precisely, and any oxide damage reduces accuracy.

Uneven Block Wear

If an SSD always wrote to the same blocks, they'd fail in a matter of months. In reality, controllers try to distribute writes evenly - this is called wear leveling. However, perfect distribution is difficult: system files, caches, and OS logs constantly use the same space, causing "hot" blocks to fail first.

Block Erasure: The Main SSD "Killer"

Page writes themselves aren't especially harmful, but block erasure is a heavy operation. The more frequently an SSD must perform such cycles, the faster its resource is depleted.

Degradation Reduces Speed

As SSDs age:

The number of healthy blocks decreases
Remaining blocks bear more load
Write speeds drop
The available SLC cache shrinks
The controller must move data more often

At later stages, SSDs may operate noticeably slower than their original specs.

NAND: A Fundamentally Limited Technology

This is why new types of non-volatile memory are being developed to overcome oxide wear issues - more on those later in the article.

TBW, DWPD, and Endurance: What SSD Lifespan Metrics Mean

To estimate SSD longevity, manufacturers use several key metrics. These do not indicate lifespan in years, but rather the rewrite resource, which depends on NAND type, controller quality, and memory management algorithms.

TBW - Total Bytes Written

This is the most widely known resource metric. It shows how many terabytes of data can be written before blocks begin to fail. For example:

300 TBW means the SSD is guaranteed to handle 300 TB of writes.
At 20-30 GB written per day, even 150 TBW lasts over 10 years.

TBW depends on cell type:

TLC - average resource
QLC - lowest
MLC - very high
SLC (mostly used as cache) - virtually "eternal" for consumer use

DWPD - Drive Writes Per Day

DWPD is common in enterprise settings. It indicates how many times the entire SSD capacity can be rewritten per day during the warranty period. For example:

1 DWPD for a 1 TB SSD means 1 TB can be written daily throughout the warranty.
0.3 DWPD is standard for consumer SSDs.
3-10 DWPD is typical of server-grade models.

Endurance - Overall Memory Robustness

This term refers to an SSD's overall ability to endure many rewrite cycles. It includes several factors:

TBW
NAND type and quality
Wear leveling efficiency
Block over-provisioning
Controller and firmware reliability

These parameters help assess how quickly a drive may slow down, how long it will remain stable, how much write load it can withstand, and when it's time to consider replacement. Even after reaching TBW, SSDs don't die immediately - they can keep working as long as spare blocks remain, but error rates rise and the controller must move data more often.

Why Do Different SSDs Have Different TBW?

Lifespan depends not only on NAND type but also:

Drive capacity (1 TB models always have higher TBW than 256 GB ones)
Chip quality
DRAM presence
Controller sophistication
Amount of reserved space

This shows why longevity depends not just on rewrite cycles but also on how the controller distributes load among cells - which is the job of wear leveling, discussed next.

Wear Leveling: How Controllers Distribute Wear and Prolong SSD Life

If SSDs always wrote data to the same blocks, even the best models would fail within months. That's why every modern SSD uses wear leveling - a system that evenly distributes wear across all NAND cells. This is a vital technology that literally extends the drive's life.

Why Is Wear Leveling Necessary?

NAND can only be erased in blocks, and each block has a limited cycle count. If some areas are filled and cleared more often than others (due to temp files, browser caches, or OS logs), they degrade far faster. Wear leveling ensures all blocks are used about equally often.

Two Main Types of Wear Leveling

Dynamic Wear Leveling
The basic approach: only active writes are distributed, so new data goes to less worn blocks. However, "cold" data (e.g., system files) isn't touched - these blocks still age, just more slowly.
Static Wear Leveling
More advanced and effective, this method moves even rarely changed data periodically, freeing up blocks and ensuring all participate in wear cycles. Static wear leveling is essential for maximum SSD longevity, especially with TLC and QLC.

How Does the Controller Decide Where to Write?

The controller tracks:

Rewrite cycles per block
Presence of bad pages
OS workload
Free reserve (over-provisioning)
SLC cache status
Controller and NAND temperature

The goal: use worn blocks less, fresher blocks more, and ensure cold data blocks also see regular use.

Wear Leveling: Not Just Longevity, but Speed

Without even load distribution, the controller would need to move data more often to free blocks, causing:

Write speed drops
Higher latency
Poor SLC cache utilization
Increased garbage collection load

Wear leveling operates constantly and invisibly, maintaining stability even under heavy loads.

Why Wear Leveling Matters Most for TLC and QLC

Multi-level cells store more states, so they:

Are more sensitive to oxide degradation
Have lower endurance
Lose charge retention ability faster

Effective wear leveling is a key differentiator between high-quality SSDs and cheap models.

Garbage Collection and TRIM: How SSDs Free Up Space and Why It Matters

SSDs cannot overwrite data on top of old pages - pages must be erased first, and erasure is only possible at the block level (hundreds of KB to MB). That's why SSDs use memory cleaning and preparation systems: Garbage Collection and TRIM. These mechanisms work together and directly impact drive speed, stability, and lifespan.

Why SSDs Need "Garbage Collection"

When a file is deleted in the OS, the SSD doesn't automatically know - the pages remain marked as occupied. To write new data, the controller must deal with "dirty" pages, even if they're no longer useful to the user. To free up space, the controller must:

Find blocks with invalid pages
Move valid data elsewhere
Erase the entire block
Write new data

This is called Garbage Collection (GC).

How Garbage Collection Works

GC is a background system that:

Selects blocks with lots of "junk" pages
Moves live data to new blocks
Erases old blocks
Prepares them for writing

However, moving data causes additional write cycles, accelerating NAND wear. A quality SSD minimizes this through smart algorithms and sufficient over-provisioning.

Why TRIM Is Critical for Speed and Endurance

The TRIM command lets the OS tell the SSD which pages are no longer valid. When a file is deleted:

On HDD, it's just marked as deleted
On SSD, the OS sends a TRIM command: "these pages can be considered empty"

This allows:

Garbage Collection to prepare free space in advance
The controller to avoid copying unnecessary data
Fewer extra write cycles
Significantly increased SSD lifespan

Without TRIM, SSD performance drops and wear accelerates.

How GC and TRIM Affect Real-World Speed

A well-maintained SSD writes quickly because it works with pre-cleaned pages. But if the drive is:

Over 80-90% full
Subjected to heavy, frequent writes
Lacking TRIM or effective GC

It soon faces a shortage of free blocks, leading to:

Sharp write speed drops
Rapid SLC cache exhaustion
More intensive cleaning operations by the controller

This is why users sometimes notice SSDs "getting slower," especially after long use or when free space is low.

Garbage Collection Still Works Without TRIM

This matters, for example, with:

Old OSes (Windows XP, early macOS)
External SSDs via USB
RAID arrays without TRIM support

But in such cases, GC works "blind" and less efficiently, harming endurance and stability.

Why SSD Speed Drops: Cache, Full Drives, and Controller Degradation

Many users notice their SSDs slow down over time, especially on writes. This isn't always a sign of major failure - speed depends on many factors: NAND condition, free space, controller design, cleaning algorithms, and even temperature. Let's highlight the main reasons.

1. SLC Cache Exhaustion - The Main Culprit Behind Sharp Speed Drops

Most modern SSDs use SLC cache - a part of NAND operating in SLC (1 bit) mode, serving as a fast write buffer. While the cache is free:

Writes are very fast (500-3000 MB/s)
The controller smoothly distributes data to TLC/QLC

But once the cache fills, the SSD must write directly to TLC/QLC, which:

Are 5-10 times slower
Require more time to program charge levels
Are more sensitive to overheating and wear

This especially affects QLC SSDs, where speed can drop to 80-150 MB/s.

2. SSD is Over 80-90% Full

As free space runs out:

The SLC cache (which is partly elastic) shrinks
Garbage Collection load increases
More pages must be copied during block cleaning

In this state, the SSD constantly struggles for space, causing speed to plunge by 20-70%.

3. NAND Memory Wear

Old blocks:

Write more slowly
Produce more ECC errors
Require rewrites
Reduce random operation performance

The controller will try to compensate for errors, but this slows both write and read speeds.

4. Controller Overheating

SSDs heat up during:

Long writes
Laptop use without cooling
Stress tests, large file copies

If controller temperature exceeds 70-80°C, thermal throttling kicks in to cool it, drastically reducing speed. Prolonged overheating accelerates controller and NAND degradation.

5. Missing TRIM or Ineffective GC

Without TRIM, the SSD treats deleted data as present and wastes resources managing it. This leads to:

Longer Garbage Collection cycles
More data movement
Write amplification (more write traffic)

All of which directly reduce speed.

6. Controller Slowdown Due to Wear or Errors

The controller is the "brain" of the SSD, and its degradation affects performance too. Signs include:

Delays opening files
Write speed drops even on an empty SSD
Unstable S.M.A.R.T. values

Over time, the controller must perform more ECC error corrections, slowing the drive.

7. SSD Enters Protection Modes

If errors or power issues are detected, the SSD may:

Throttle speed
Disable part of the cache
Limit NAND channel parallelism
Increase internal delays

This protects data but noticeably reduces performance.

How to Check SSD Health: S.M.A.R.T., Wear, and Speed

Monitoring SSD health helps detect memory, controller, or speed issues early - and shows how close the drive is to exhaustion. Modern SSDs provide many diagnostic metrics for health assessment without physical disassembly or complex tests.

S.M.A.R.T.: The Main Source of SSD Health Info

All SSDs support S.M.A.R.T. monitoring - a set of parameters reported to the OS and diagnostic tools. Key indicators include:

Percentage Used / Media Wearout Indicator
The main wear metric. 0% = new, 100% = exhausted. In practice, drives keep working past 100%, but error risk rises.
Total Host Writes / NAND Writes
Shows how much data has actually been written. These figures let you compare usage to the manufacturer's TBW spec.
Wear Leveling Count
Number of block rewrite cycles. Growth is normal, but a wide spread between blocks may indicate early degradation.
Reallocated Blocks / Bad Blocks
Number of remapped (failed) blocks. If the count grows, NAND is starting to fail.
Uncorrectable Errors / CRC Errors
Read errors ECC couldn't fix. This is a red flag: data may be corrupted.

Speed Tests: Performance Drops as a Warning Sign

Speed changes are the first symptoms of degradation or shortage of free blocks. When testing, run:

Sequential Write Test
Reveals SLC cache size, direct TLC/QLC write speed, and any throttling due to overheating. If speed drops sharply after cache fills, the SSD is operating at its limit.
Random 4K Operations Test
Reflects controller health, internal algorithms, and NAND latency. Drops in random write/read speeds often signal wear or GC/TRIM issues.

SSD Temperature Monitoring

Many drives are sensitive to overheating. If temperature exceeds 70-80°C:

Throttling activates
The controller lowers clock speeds
NAND wears out faster

Monitoring temperature under load helps identify poor cooling, missing thermal pads, or inadequate airflow.

How to Interpret Results

If you observe:

Growing reallocated block count
SLC cache shrinkage
Noticeable speed drops
Increased ECC error rates

This indicates early SSD degradation. In this case, you should:

Back up your data
Avoid prolonged heavy writes
Monitor temperature
Keep at least 15-20% free space

Regular diagnostics help you gauge how close your drive is to its limits and what actions to take to prolong its life.

How to Extend SSD Lifespan: Practical Tips

While SSDs do wear out, proper use can significantly slow degradation and maintain high speeds for years. Modern drives already use advanced wear leveling, block cleaning, and caching algorithms, but user habits still matter. Here's how to maximize SSD life with minimal effort:

1. Don't Fill Your SSD Completely

Leave at least 15-20% free space. This ensures:

Proper SLC cache operation
Effective wear leveling
Fast garbage collection
Reduced write amplification

If you fill an SSD to 90-95%, speed can drop dramatically, and wear increases sharply.

2. Don't Disable TRIM

TRIM is crucial - it tells your SSD which pages are empty. Without it, the controller:

Wastes resources moving unneeded data
Performs costly block erasures more often
Wears out NAND faster

Make sure TRIM is enabled in your OS.

3. Watch the Temperature

SSDs are vulnerable to overheating. Temperatures above 70-80°C:

Trigger throttling
Slow down the drive
Accelerate cell wear
Damage the controller

This is especially important for NVMe drives in thin laptops. Use thermal pads and ensure good airflow.

4. Avoid Constant Heavy Write Loads

Regular tasks like video editing, virtual machines, and large file copies strain TLC and QLC drives. If your workflow involves intensive writes:

Consider SSDs with DRAM
Choose TLC or MLC models
Use a separate drive for working files

5. Keep SSD Firmware Updated

Firmware updates often include:

Improved wear leveling algorithms
Bug fixes
Cache optimizations
Better thermal management

Manufacturers regularly enhance stability and endurance via updates.

6. Disable Defragmentation

On SSDs, defrag is:

Useless
Generates massive unnecessary writes
Accelerates wear

Modern operating systems usually disable defrag for SSDs by default, but it's worth double-checking.

7. Use Over-Provisioning (If Space Allows)

Over-provisioning is unallocated space the controller uses for:

Wear leveling
Efficient GC
Faster writes

Keeping 5-10% of the drive unpartitioned can noticeably improve endurance.

8. Don't Use SSDs Until Their "Last Breath"

When S.M.A.R.T. shows:

High Wear Leveling Count
Growing reallocated block count
Increasing uncorrectable errors

Back up and consider replacement early - degradation accelerates rapidly afterwards.

9. Choose the Right SSD for Your Needs

Different memory types suit different use cases, and your choice directly affects lifespan:

For gaming, office, and everyday use, TLC SSDs offer good speed and endurance.
For large archival storage that rarely changes, QLC models with high capacity and low cost per GB are suitable.
For professional workloads and heavy writes, pick SSDs with DRAM cache and high TBW - these handle intense loads and maintain stable speeds.
There are also promising types of non-volatile memory that lack NAND's main issues and are seen as the future of reliable storage - learn more in the article MRAM and RRAM: The Future of Memory Beyond DRAM and NAND.

Conclusion

SSDs have long been the primary storage type, offering high speed, instant data access, and low power usage. But they are naturally subject to wear, not because of a particular model's quality, but due to NAND memory's core properties. SSD cells endure a limited number of rewrite cycles, and their degradation gradually reduces speed, increases errors, and places greater strain on the controller.

However, SSD lifespan depends greatly on how efficiently memory management is organized. Technologies like wear leveling, garbage collection, TRIM, SLC caching, and block reservation significantly extend drive life and stability. Users can further boost endurance by managing free space, keeping temperatures low, updating firmware, and choosing drives suited to their tasks.

With proper care, even TLC and QLC SSDs can last for many years without critical issues, and emerging non-volatile memory types will eventually eliminate NAND's degradation limits altogether.

SSDs are fast and reliable - and understanding how they work and age lets you make the most of their capabilities.