{"slug": "a-graded-registry-of-gpu-cloud-prices-and-fabric-disclosure", "title": "A graded registry of GPU cloud prices and fabric disclosure", "summary": "Compute Assay launched ASSAY-1, a free, versioned specification, registry, and weekly report that defines the GPU-hour as a graded deliverable good, after finding a 3.8× price spread across H100 capacity tracked in its registry. The registry covers 133 listings, with 28 live-API prices from AWS, Azure, Oracle, and Vast refreshed 2026-09-10 and 105 from a weekly survey most recently dated 2026-09-14, and it excludes quote-gated teaser rates such as TensorDock's $2.25 and Voltage Park's $1.99 from headline figures. Compute Assay argues cash-settled GPU futures now listing on four venues price a unit that is not fungible, since substitutability depends on network fabric and workload class rather than chip count alone.", "body_md": "The delivery layer for compute\n\n# The exchanges built a price for compute.\n\nNobody defined the product.\n\nCash-settled GPU futures are listing on four venues. All of them price the GPU-hour. None can deliver one — because a GPU-hour is not a fungible unit, and whether one cluster can substitute for another is a networking question the financial layer cannot answer. Compute Assay is the specification, registry, and weekly report that define the deliverable.\n\nSubstitutability is a *relation* between two fleets with respect to a workload class — not a property of a fleet. Eight H100s on oversubscribed Ethernet and eight in a rail-optimized non-blocking InfiniBand plane are identical goods for batch inference, different goods for large-model serving, and not comparable for frontier training. The indexes count chips. This describes deliverable capacity. — [ASSAY-1 §0.1](https://computeassay.com/spec.html)\n\n## One chip. One week. 3.8× price spread.\n\nThis is not a ripoff chart — it is an undefined-product chart. The green dots publish enough about their network fabric to be underwritten; the red ones publish nothing. The spread is what an unwritten specification costs.\n\nExcluded from the chart and spread: TensorDock's \"starting at\" $2.25; Voltage Park's \"starting at\" $1.99. Floor and teaser rates — where the bookable price is quote-gated or varies by host — are listed on the [H100-SXM page](https://computeassay.com/gpus/h100-sxm.html) with that label; they never anchor a headline number here.\n\n[ASSAY-1 — the specificationThe record format, headline metrics (CSD·NBD·DCB·PFI), workload classes, substitutability grades, and verification harness that define compute as a deliverable good. Versioned like an RFC. Free forever.](spec.html)\n\n[The Delivered Compute ReportWeekly. What GPU capacity actually exists, what it costs, and whether it can do the job the price implies. Every row sourced.](report.html)\n\n[Compute has no kilowatt-hourThe frame: why compute is pre-kWh, and the three-step path from a unit of capacity to a unit of delivered intelligence.](frame.html)\n\n[The Disclosure LeaderboardProviders ranked by disclosed training-class capacity and fabric transparency — what they state in writing, not measured performance. Re-ranked weekly from public sources.](https://computeassay.com/leaderboard.html)\n\n[Run-cost calculatorWhat your training run costs this week, with a disclosed training-class band — from live tracked prices.](https://computeassay.com/calculator.html)\n\n[Open dataThe full graded registry as JSON/CSV, free under CC BY 4.0. Every row sourced.](https://computeassay.com/data.html)\n\n### Get the Delivered Compute Report\n\nOne email a week: headline prices, deliverability grades, one teardown, and the gap between the index and what you can actually take delivery of.\n\n[Subscribe free](https://computeassay.substack.com/subscribe)\n\nLive-API prices (28 listings: AWS, Azure, Oracle, Vast) refreshed **2026-09-10**; the other 105 from the weekly survey, most recently **2026-09-14**. Every row carries its own *as of* date — staleness is disclosed, never hidden.\n\n## What an assay is\n\nIn every physical commodity market, an *assay* is the test that determines grade and purity — the reason a buyer in one port can trust a cargo loaded in another. Compute has never had one. Eight H100s on shared Ethernet and eight H100s in an NVLink domain with non-blocking InfiniBand sell under the same name at prices that differ by multiples, and the difference is written nowhere a contract can cite. ASSAY-1 is that test, for compute: published free, graded conservatively from what sellers state in writing, and revised in the open.", "url": "https://wpnews.pro/news/a-graded-registry-of-gpu-cloud-prices-and-fabric-disclosure", "canonical_source": "https://computeassay.com/", "published_at": "2026-09-21 19:58:07+00:00", "updated_at": "2026-09-21 20:24:34.733298+00:00", "lang": "en", "topics": ["ai-infrastructure", "ai-chips", "structured-data"], "entities": ["Compute Assay", "ASSAY-1", "H100", "AWS", "Azure", "Oracle", "Vast", "TensorDock"], "alternates": {"html": "https://wpnews.pro/news/a-graded-registry-of-gpu-cloud-prices-and-fabric-disclosure", "markdown": "https://wpnews.pro/news/a-graded-registry-of-gpu-cloud-prices-and-fabric-disclosure.md", "text": "https://wpnews.pro/news/a-graded-registry-of-gpu-cloud-prices-and-fabric-disclosure.txt", "jsonld": "https://wpnews.pro/news/a-graded-registry-of-gpu-cloud-prices-and-fabric-disclosure.jsonld"}}