Tom’s Hardware has outlined a revised methodology for evaluating enterprise and data center solid-state drives, explaining on November 17, 2026 how its new test suite is intended to stress high-capacity PCIe 5.0 drives under conditions that resemble sustained AI workloads.
Enterprise SSDs are judged by different standards than consumer drives. Desktop users mostly care about short bursts of activity and fast load times, while server operators need consistent performance over hours or days of continuous reads and writes, predictable latency, and endurance measured in drive writes per day. A drive that looks fast in a brief benchmark can slow considerably once its cache fills or it reaches a steady state.
The publication’s updated approach is designed to capture that behavior, according to its explanation, by pushing drives beyond short synthetic runs and toward the long-duration loads seen in data ingest, model training and inference servers. The methodology accompanies its review of the Phison Pascari X201P, one of the first drives tested under the new regime.
The shift mirrors changes in the storage market itself. Demand from AI infrastructure has pushed vendors to release larger and faster drives, and buyers increasingly want independent testing that reflects how the hardware performs in production rather than in marketing specifications.
Created by Ayen Stabel.
Stabel is AI and can make mistakes.
Sources:
https://www.tomshardware.com/pc-components/ssds/how-we-test-enterprise-ssds-our-new-test-suite-pushes-enterprise-and-data-center-ssds-to-their-limits