The Compute Indices — Researchhttps://thecomputeindices.com/research.html2026-09-07T00:00:00ZMark RuschTCI publishes reproducible reference prices for AI and HPC compute — methodology-driven, vendor-neutral, and rebuildable from public sources by anyone.A GPU-hour is a part number, not a unithttps://thecomputeindices.com/research/a-gpu-hour-is-a-part-number.html2026-09-07T00:00:00ZThe index records an interconnect for all 7,798 observations and has observed one for none of them, 805 rows assert an SXM part and a PCIe bus at the same time, and no source in the panel has ever disclosed a fabric, a power envelope or a thermal limit.Mark RuschCompute prices barely move. Where you buy moves everything.https://thecomputeindices.com/research/what-moves-compute-prices.html2026-09-06T00:00:00ZFive of nine European sellers never changed their H100 price once in fifty days, the same-hour spread across sellers was 8.56x, and the hyperscaler rate cards are region-flat to four decimal places, which is the argument for regional benchmarks and not the argument most people make for them.Mark RuschWhat makes a compute benchmark hedgeablehttps://thecomputeindices.com/research/hedgeable-benchmark-design.html2026-08-16T00:00:00ZDesign rules from oil, power, freight and repo, and which of them compute can actually satisfyMark RuschWhen an index measures its own samplinghttps://thecomputeindices.com/research/composition-vs-price.html2026-08-16T00:00:00ZHow a 10% move in a compute benchmark turned out to be entirely composition noise, and what the unit definition had to do with itMark Rusch