A comprehensive interactive dataset tracks memory and storage prices from 1960 through 2026, building on John C. McCallum's classic collection. The tool displays price per gigabyte over time across DRAM, NAND flash, and HBM categories, with users able to hover for details, toggle series, zoom via slider, and export visualizations. Each memory type gets its own trend line with historical context.

DRAM pricing spans from Pre-DDR through DDR5 generations, using McCallum's data through 2024 and monthly Amazon retail prices via Keepa thereafter. NAND flash coverage focuses on consumer NVMe SSDs since 2016, excluding enterprise drives and filtering obvious pricing errors. The dataset fills gaps with sparse pre-NVMe estimates where no continuous record exists, providing the most complete retail-price picture available.

HBM costs come from industry analyst estimates rather than spot markets, tracking generations from HBM2e through the projected HBM4 launching Q3 2026. Quarterly breakdowns show spending across Nvidia, AMD, Google TPU, and Amazon Trainium accelerators, broken down by component including HBM, logic die, and packaging. The analysis uses production-volume-weighted averages across the four largest accelerator designers.

Methodology relies on cheapest retail listings in nominal USD, meaning prices may lag contract rates and often reflect end-of-life clearance items rather than cutting-edge technology. Updates occur monthly for DRAM and NAND data, quarterly for HBM estimates. Compiled by David Shim at Stanford's DAM project, the dataset serves researchers and engineers tracking hardware cost trends.