Every NAND cell tolerates a finite number of program and erase cycles. Drive controllers spread writes across the array through wear levelling, so the useful measure is not cycles per cell but total host writes the drive can absorb before cells begin to retire faster than spares can cover.
Write amplification is the hidden multiplier
Writing 1 GB from the host rarely causes exactly 1 GB of NAND writes. Garbage collection and block-level erase granularity mean the drive writes more internally, a ratio called write amplification. Sequential large writes keep it near one; small random writes on a nearly full drive can push it much higher.
This is why leaving free space matters. A drive kept above roughly 10 to 15 percent free has more room to consolidate blocks and does less internal rewriting.
Sizing guidance
For general desktop use, annual host writes are commonly in the low tens of terabytes, so a mainstream drive rated in the hundreds of terabytes will outlast the machine. For continuous capture or write-heavy servers, calculate expected daily writes, multiply by the intended service life, and choose a rating with clear headroom above that figure.