Originally published on Forbes by Thomas Coughlin, April 16, 2026.: The costs of storage to support AI and other enterprise workloads have exploded. Especially for solid state drives, SSDs, used for primary storage to support the high bandwidth memory, HBM, demand of graphical processing units, GPUs. Some companies push SSD storage for all enterprise storage, but with the differences in pricing, you may want to think again about doing so.
Vdura has set up a website that tracks enterprise SSD and hard disk drive, HDD, prices. The company’s Vdura flash volatility index shows that by Q1 2026 30 TB QLC SSD capacity cost 22.6 times as much as equivalent 30 TB HDD storage capacity. Enterprise SSD prices increased by almost 24% in the three weeks between March 4 and March 23, 2026.
The most recent price increases build on a sustained period of SSD price escalation. Pricing for 30TB TLC enterprise SSDs increased by 472% between Q2 2025 and Q1 2026, rising from $3,062 to $17,500. Over the same period, 30TB QLC SSD pricing increased from $2,450 to $15,121, fundamentally reshaping the economics of flash-based storage. Over the same time period 30TB HDD prices increased from $495 to $668, or about 35%.
The figure below from the Vdura page shows that if we assume a 15% increase in HDD capacity pricing and a 20% change in SSD capacity pricing the difference between SSD and HDD storage capacity would be 23.6X. This should be compared to a 4.9X price difference between SSD storage capacity and HDD storage capacity in CQ2 2025.
Using the Storage Economics Optimizer Tool on their website, VDURA analyzed the cost impact of flash pricing changes across common storage architectures for a 25 PB deployment delivering 1,000 GB/s of sustained performance. At Q2 2025 pricing, an all-flash architecture carried a 3-year cost of $9.69M. By Q2 2026, that same configuration increased to $48.17M, a ~397% increase driven primarily by flash media pricing.
Tiered storage, where multiple types of storage are used together, with storage management software, like that from Vdura, to move data between slower but less expensive storage, such as HDDs and faster storage, such as SSDs can provide a great way to optimize storage performance and price.
With this approach data in immediate use is kept on the faster, but more expensive storage and with data not currently used kept on the slower, less expensive storage. When hot data gets cold it moves to the less expensive storage and when colder data is needed again it moves to the more expensive but higher performing storage.
Vdura, formerly known as Panasas, is a provider of digital storage solutions for AI and high-performance computing, HPC, based upon the PanFS, parallel file system. The company offers flash first storage that can also include tiered layers of hard disk drives.
SSD capacity prices are now over 20 times those of HDD capacity prices. This will encourage storage customers to look at using storage tiering that includes SSDs and HDDs with data movement as needed.