๐Ÿ’พ GiB vs GB Storage Discrepancy

Your "1 TB" SSD shows 931 GB in Windows. Nobody lied to you. Storage marketing uses decimal (powers of 10). Operating systems use binary (powers of 2). The gap is 6.9% at GB scale. At the petabyte level it approaches 10%. Quantify the difference before procurement signs the PO.

๐Ÿ“ Drive Configuration

Input the box-label capacity in marketing GB. Output: what your OS actually reports in GiB, plus the exact deficit in absolute and percentage terms. No surprises when the procurement spreadsheet meets the OS disk manager.

GiB vs GB: The 6.87% Storage Discrepancy Every Engineer Should Know

If you have ever plugged in a "1 TB" drive and seen your operating system report only 931 GiB of usable space, you have encountered the decimal vs. binary measurement gap. Drive manufacturers label capacity using the metric (base-10) system where 1 GB = 1,000,000,000 bytes (109). Operating systems, file systems, and memory allocators operate in binary (base-2) where 1 GiB = 1,073,741,824 bytes (230). The ratio between these two definitions is 109 / 230 โ‰ˆ 0.9313, meaning every "gigabyte" on the box translates to approximately 0.9313 GiB of actual usable capacity โ€” a 6.87% deficit that compounds as capacity scales up.

Why 1 TB = 931 GiB

The math is deterministic. A drive labeled "1 TB" contains 1,000,000,000,000 bytes (1012). When the OS divides this by 240 (1 TiB = 1,099,511,627,776 bytes), the result is approximately 0.9095 TiB โ€” or 931.3 GiB. This is not a defect, a formatting overhead, or a hidden partition. It is purely a unit-of-measurement mismatch. The drive contains exactly the number of bytes advertised; the OS is simply using a different unit to count them. For an enterprise storage array with 100 TB of raw capacity, the discrepancy amounts to nearly 7 TB of "missing" space โ€” enough to store an entire additional application dataset.

Impact on Enterprise Procurement

This discrepancy has real financial consequences in enterprise procurement and capacity planning. When an infrastructure team requests "500 TB of usable storage," the vendor responds with a quote for 500 TB of raw (decimal) capacity โ€” which yields only 465 TiB of usable space. The gap of 35 TiB represents roughly $700-$1,400/month in cloud block storage costs, or $8,000-$17,000/year. Multiply this across a fleet of storage arrays, backup targets, and archival tiers, and the aggregate budget error can reach six figures annually. Always specify capacity requirements in binary units (GiB/TiB) when dealing with OS-level provisioning, or explicitly factor in the 6.87% conversion margin when procuring raw storage.

Beyond Marketing: Where Binary Units Are Mandatory

In memory (RAM), the binary convention is universal โ€” a 16 GB DIMM is genuinely 16 GiB (234 bytes). In networking, speeds use decimal (1 Gbps = 109 bits/sec). In storage, the split persists: NAND flash chips are fabricated on binary boundaries (128 GiB, 256 GiB), but consumer products are labeled in decimal. The JEDEC and IEC standards bodies have long recommended using the binary prefixes (kibi-, mebi-, gibi-, tebi-) to eliminate ambiguity, but consumer marketing has resisted the change. As a result, every engineer working at the intersection of hardware procurement and system provisioning must internalize this conversion or risk systematic under-provisioning.