Why Do SSDs Wear Out When RAM Practically Never Does?

An SSD stores each bit by pushing electrons through a layer of glass-like insulation with 15 to 20 volts, and every trip through that layer leaves a little permanent damage behind, so a flash memory cell of the mid-2010s was rated for only about 1,000 to 3,000 rewrites. RAM never forces charge through its insulation: it fills and drains a tiny electric bucket through a switch at low voltage, which is why it can be topped up about 490 million times a year without wearing, and also why it forgets everything the moment the power stops. RAM does still age (Google’s servers showed rising error rates after 10 to 18 months), but not from being written to.

Sixty-four milliseconds. That is how long a cell of RAM is trusted to hold one bit before the computer must recharge it. Do the sum and it turns absurd. One second divided by 0.064 is about 15.6 top-ups a second. Over a year, that comes to roughly 490 million, for each of the billions of cells on the module. All of that, just to remember what it already knew.

Now walk over to the SSD in the same machine. A flash cell of the mid-2010s was rated for around 1,000 rewrites. Not per year: in its whole life.

So the chip recharged half a billion times a year ships with no wear rating at all. The chip that tolerates a thousand comes with a warranty that counts every terabyte you write. Your intuition says that is backwards. It has the wrong picture of what a “write” is. The answer lies in what each chip does to an electron to make it stay put. It also lies in a layer of glass a few millionths of a millimeter thick.

How Does A Computer Chip Store A 1 Or A 0?

Press a key and somewhere in your machine a bit flips. A bit is the smallest unit of information, a 1 or a 0. Inside a chip, a bit is not a number written down anywhere. It is a small pocket of electric charge, or the lack of one.

Everything else follows from one question: where do you keep the charge, and how do you get it in and out?

The first is a bucket with a tap. You put a tiny capacitor, a part that holds charge, behind a transistor, an electronic switch. Open the switch, let charge pour in or drain out, close the switch. That is RAM.

The second is an island sealed inside glass. You park the electrons on a scrap of conductor wrapped on every side in an insulator, with no tap at all. This is flash, the memory inside every SSD, phone and USB stick.

Both chips hold a bit as charge. The difference is the door: RAM has one, and flash bricked it up. The forgetting, the wear and the speed gap all come from that single choice.

How Does RAM Hold A Bit, And Why Does It Forget?

The RAM on your memory sticks is DRAM, short for dynamic random-access memory. Each cell is one transistor and one capacitor, which engineers shorten to 1T1C. In MIT’s course notes, the voltage across that capacitor is the stored bit.

To write a 1, the chip opens the transistor and the wire behind it fills the capacitor. To write a 0, it opens the transistor and drains it. The charge moves through the transistor’s channel at the chip’s normal low voltage. Any ordinary current takes that same path. Nothing crosses an insulator, so nothing gets stressed.

The price is that the bucket leaks. A 2017 review by Cai and colleagues puts DRAM’s memory span at “milliseconds to seconds.” To cope, the memory controller walks the module and recharges every cell in place. The standards, the review notes, call for every cell to be refreshed at least once every 64 milliseconds.

Walk the arithmetic:

1. 64 milliseconds is 0.064 seconds.
2. 1 ÷ 0.064 = 15.6 refreshes per second.
3. A year has 31,536,000 seconds, so 15.6 × 31,536,000 ≈ 490 million refreshes per cell per year.

Cut the power and the refreshing stops. The buckets drain in under a second and your unsaved document is gone. This is why RAM forgets when the power goes off, and why a computer needs a permanent store as well. A bucket with a hole in it is a strange way to build a memory. It is also why the memory never wears out.

How Does Flash Memory Keep Data With The Power Off?

A flash cell is a transistor with one extra piece: a floating gate. The review describes it as sealed above and below by layers of oxide. The oxide is silicon dioxide, the same compound as window glass. The bottom layer is called the tunnel oxide. A Stanford thesis on flash memory gives its thickness as about 6 to 7 nanometers. A nanometer is a millionth of a millimeter.

Sealed like that, the electrons stay put with the power off. Where DRAM leaks in milliseconds, the review says flash holds its charge for “several days to years.” The consumer standard, JEDEC’s JESD218, asks a client SSD to keep its data for one year unpowered at 30 °C (86 °F). That holds even after the drive has used up all its rated writes.

But how do you get electrons onto an island with no door? You force them through the wall. The Stanford thesis puts the price at 15 to 20 volts on the control gate. At that voltage, electrons cross the glass by quantum tunneling, a trick in which a particle turns up beyond a barrier it lacks the energy to climb. Engineers call this route Fowler–Nordheim tunneling. The review names it as the mechanism behind every program and erase.

Erasing pulls the electrons back out the same way, and the review notes it works only on a whole block of cells at once. Flash cannot change one byte in place; it erases a block, then writes it again. One erase plus one write is a program/erase cycle, or P/E cycle, the unit flash wear is counted in.

RAM opens a switch at low voltage. Flash drives electrons through a glass wall at 15 to 20 volts on every single write.

What Wears Out Inside An SSD?

Every trip through the glass costs something. The review is blunt: with repeated cycles, “electrons become trapped in the tunnel oxide.” The damage adds up over the life of a block, and the review has a name for it: wearout. The Stanford thesis says the same from the materials side: the high voltage creates defects, or traps, in the oxide.

In plain words, the glass gets peppered with stuck electrons and tiny flaws. Stuck charge changes how much charge the cell appears to hold, so the reader mistakes one level for another. The flaws give stored electrons a path to seep out. The thesis calls that leak stress-induced leakage current.

The SSD fights back with error correction. Its controller stores extra check bits beside the data. On every read, it uses them to repair a limited number of wrong bits. The review’s definition of worn out is simple. The block can no longer hold data safely for as long as the maker guarantees. So a worn-out cell is not a dead cell. It is a cell that can no longer be trusted with your file for the promised year.

What about reading? Reads do not count as P/E cycles, since nothing gets pushed onto the floating gate, but they are not free either. The review notes that a read leans about 6 volts on the neighboring cells, against about 15 volts for a write. A page read millions of times can drift until the controller rewrites it.

Most modern 3D flash uses a charge-trap cell. It stores electrons in an insulator instead of on a conductor, and the review rates it “less susceptible” to cycling damage. It still writes by tunneling through an oxide, so less susceptible is not the same as immune.

Why Do Cheaper, Bigger SSDs Wear Out Faster?

The short answer is bits per cell. The first flash stored one bit per cell (SLC) by splitting the cell’s charge range into two windows. MLC splits the same range into four windows for two bits. TLC uses eight windows for three bits, and QLC stores four bits, which needs sixteen. Each step packs more data into the same silicon. The review adds that each step also makes the bits harder to store and read without error.

The margin for damage shrinks with the windows. In an SLC cell, a lot of trapped charge must build up before a 1 starts to look like a 0. A QLC cell has to tell sixteen shades of “some electrons” apart.

The figures from the mid-2010s, when cells were also shrinking, show the slide. The review reports about 150,000 P/E cycles for older SLC drives and about 10,000 for 50-59 nanometer MLC. For 15-19 nanometer flash, it gives “~3,000 and ~1,000 P/E cycles per block” for MLC and TLC. Today’s drives are built in 3D, but the trade-off has not gone away.

Bigger cuts the other way. More capacity means more cells to spread the writes across. Look at Samsung’s 990 PRO datasheet. The 1 TB, 2 TB and 4 TB models are rated at 600, 1,200 and 2,400 terabytes written, under one five-year warranty. The 4 TB drive gets four times the write budget with no tougher cells. It has four times as many of them.

Can An SSD Last 20 Years? What Is The Average Lifespan Of An SSD?

An SSD survives its own chemistry with two tricks. The first is wear leveling. The controller steers each new write toward the least-used blocks, so the whole drive wears at an even rate. The second is overprovisioning. Makers give the controller more physical blocks than the advertised capacity. That spare room rotates in as blocks tire.

The write budget is what the warranty counts, and it lets you do a rough sum:

Years = TBW ÷ (gigabytes written per day × 365)

TBW is terabytes written, the drive’s rated total. Take the 1 TB 990 PRO at 600 TBW and a laptop writing an illustrative 40 GB a day:

1. 600 TB is 600,000 GB.
2. 600,000 ÷ 40 = 15,000 days.
3. 15,000 ÷ 365 ≈ 41 years.

For that user, the five-year warranty runs out decades before the write budget does.

Twenty years is another matter, and no field study runs that long. Google watched its own drives for six years. The results, presented at USENIX FAST 2016, cover many millions of drive days. Its headline is calming: raw error rates grow “at a much slower rate with wear-out” than expected. Its second finding is the catch. Even after allowing for write cycles, error rates rose with the months a drive had spent in service. Drives age on the calendar, not just on the write counter. And the JEDEC retention rule is one year unpowered, so an SSD in a drawer is not an archive.

The same study settles the SSD-versus-hard-drive question. Flash drives were replaced far less often than hard disks but threw more uncorrectable errors. So an SSD is less likely to die on you and more likely to lose a file. Back up either one, and do not defragment an SSD: it moves data for no benefit and spends P/E cycles doing it.

Why Aren’t SSDs As Fast As RAM?

The same glass answers this one. The MIT notes put a DRAM read at “several tens of nanoseconds.” A read from NAND flash, the kind in an SSD, is “on the order of 10 microseconds.” A microsecond is a thousand nanoseconds, so the SSD read is a few hundred times slower.

Writes are slower still, for the reason you now expect. The MIT notes give it in one line. A write needs high voltage to push electrons across the insulating barrier. A block must also be erased before it is rewritten. That means one small change can trigger far more copying and erasing than the change itself.

Two of the most-asked questions about SSDs, then, share one answer. The sealed island that lets flash remember without power also makes it slow to write, and writing is what wears it.

Does RAM Ever Wear Out?

The title says “practically never,” and that word is doing the work. RAM does not wear from being written, but it does age.

The evidence comes from Google’s own servers. A 2009 field study by Schroeder, Pinheiro and Weber covered millions of DIMM-days. It found that “over 8% of DIMMs” in the fleet saw at least one correctable error per year. These were not the one-off soft errors that earlier work had focused on. The authors found them “dominated by hard errors,” meaning permanent physical faults. And aging “sets in after only 10–18 months in the field.” The authors sound almost surprised. DIMMs have no moving parts, yet they show “a surprisingly strong and early effect of age.”

None of that comes from the 490 million refreshes. A refresh is a low-voltage top-up through an open switch, and it crosses no insulator. The aging is the slow, ordinary decay every chip suffers. The same study found temperature had “a surprisingly small effect” in the field.

So, Why Do SSDs Wear Out When RAM Doesn’t?

Because of the door. Both chips store a bit as a pocket of charge. RAM keeps its charge behind a switch, so a write is a low-voltage pour through an open channel. The chip can be topped up half a billion times a year for as long as its other parts hold out. The cost is a bucket that leaks in milliseconds and a memory that goes blank with the power.

Flash sealed the door. That seal is why an SSD remembers for a year unpowered. It is also why every write punches electrons through 6 or 7 nanometers of glass at 15 to 20 volts. Each punch leaves a few electrons stuck in the glass and a few new flaws. Pack three or four bits into a cell and the margins get so tight that the damage shows sooner. In time, error correction fails to keep a block’s readouts straight for the promised year, and the block is retired.

The picture is not tidy. Google’s data says flash drives also age on the calendar, independent of writes. RAM modules develop permanent faults of their own after a year or two. Neither chip is immortal, but only one of them is damaged by the act of remembering.

For most people the arithmetic is kind. A 1 TB drive rated for 600 terabytes written, fed 40 GB a day, carries about 41 years of write budget against a 5-year warranty. Keep a backup anyway, for the failures that have nothing to do with wear.