CoreWeave Cloud (HGX H100 / H200 / B200 / A100 nodes)

per GPU-hour (NOT per node). UNIT LABEL CORRECTED 2026-10-04 (parent agent): this previously read 'per 8-GPU node per HOUR', which contradicts price_usd=6.155 -- 49.24 / 8 = 6.155, so the stored number was always per-GPU and only the label was wrong. Reading the node rate as per-GPU makes every derived CoreWeave cost-per-token figure 8x too low. Found by the repo-wide audit (finding A2). All figures below are unchanged and verbatim: HGX H100 $49.24 on-demand, $19.71 spot (North America); HGX H200 $50.44 / $20.93; HGX B200 $68.80 / $34.11; HGX A100 80GB $21.60 / $9.65; L40S $18.00 / $7.88; L40 $10.00 / $6.27; GH200 single-GPU $6.50. Europe: H100 $49.24 / $19.51, H200 $50.44 / $20.64, A100 $21.60 / $9.51, B200 $68.80 / $34.87. GB200 NVL72 quoted $42.00/h for the whole 72-GPU Superchip. Fetched 2026-10-03. PRICE CLASS: this is a PUBLISHED LIST PRICE - an asking rate a new customer is quoted, not a realised, discounted or negotiated transaction rate. A real bill is at or below this and typically well below it once commitments, credits and committed-use discounts are applied PERISHABLE-FIGURE STAMP (added 2026-10-04, uniform across data/supply/ so it is greppable). PRICE CLASS: PUBLISHED LIST PRICE - an asking rate a new customer is quoted, NOT a realised, discounted or negotiated transaction rate. A real bill is at or below this and typically well below it once commitments, credits and committed-use discounts apply. OBSERVED: 2026-10-03 RE-VERIFY BY: 2027-01-04 - the repo's quarterly re-verification bar. After that date this figure is presumed drifted and must be re-checked against its source before it is quoted again. A price without its observation date is noise; this stamp exists so a reader can tell a fresh figure from a stale one without reading prose. UNIT: per GPU-hour (NOT per node) READING RULE: UNIT TRAP: the $42.00/h figure quoted here is for the WHOLE 72-GPU Superchip, not per GPU. Dividing by 72 is required and is NOT what price_usd carries.

lead_time_weeks is null because the quantity is structurally not applicable: on-demand cloud capacity starts on request, so there is no order-to-delivery lead time. The constraint is the price per GPU-hour and the availability snapshot, not an interval. The one caveat worth keeping: scarce tiers gate capacity behind a reservation or Capacity Block commitment (AWS caps the booking horizon at 8 weeks, source spl-aws-capacity-block-booking-horizon), and that is a commitment window, not a lead time.

EAR 742.6(a)(6)(iii)(A) + EAR 744.23/744.6 (Entity, Unverified, MEU list screening of the end user) + the advanced-computing IaaS obligations of 91 FR 1684 (consignee KYC, and management/limitation of IaaS access to the hardware)

LAYER 1, EXPORT LICENCE: under the Jan 2026 BIS rule 91 FR 1684 (effective 2026-01-15) the license review policy for advanced-computing commodities with TPP < 21,000 AND total DRAM bandwidth < 6,500 GB/s - the NVIDIA H200 and AMD MI325X are the rule's named examples - moved from PRESUMPTION OF DENIAL to CASE-BY-CASE REVIEW, but only for exports FROM THE UNITED STATES to end users located in China or Macau, and only against six exporter certifications filed via SNAP-R before shipment: (1) items below the thresholds with the count of units shipped in the US; (2) sufficient US supply, judged 'taking into account normal lead times'; (3) no diversion of global foundry capacity away from US end users; (4) aggregate China+Macau shipments <= 50% of total US end-use shipments of that product; (5) no end-user or use prohibition; (6) the consignee runs rigorous KYC. Every shipment must then pass a qualified third-party testing lab - US-headquartered, not controlled by a D:5/Macau entity, no ownership stake in any party, qualification revocable by BIS at any time - verifying TPP, total DRAM bandwidth, interconnect bandwidth and copackaged DRAM capacity. Reexports and in-country transfers of the same chips to China/Macau REMAIN PRESUMPTION OF DENIAL. LAYER 2, END-USER AND PARENT SCREENING: per BIS guidance of 2026-05-31 a licence is required where the consignee is HEADQUARTERED in Country Group D:5 (15 CFR part 740 supplement no. 1) or Macau OR whose ULTIMATE PARENT COMPANY is so headquartered - even when the consignee itself sits outside D:5/Macau. The requirement sits at EAR 742.6(a)(6)(iii)(A) for destinations outside the US, with an exception available only under the license exception at 740.2(a)(9)(ii). The destination country alone therefore never answers the question; screen the parent. LAYER 3, REMAINING LIMITS: exports to D:5/Macau-headquartered OR -parented entities stay PRESUMPTION OF DENIAL wherever the buyer sits, and the AI Diffusion Rule that would have set a country-tier structure on who could lawfully buy what was RESCINDED on 2025-05-13 and is not being enforced, so any recorded status rests on the instruments named here rather than on tiering.

UNITS CORRECTION 2026-10-04 (parent agent; arithmetic 49.24/8=6.155 confirms it): price_usd is PER GPU-HOUR, not per node. The prior label said 'per 8-GPU node per HOUR'. The stored number was always right, so the defect was invisible to review and to validate.py -- it is exactly the failure price_basis exists to prevent, and CoreWeave is a reference rental point in docs/08-cost-per-token.md. The most USEFUL price structure in the slice, because CoreWeave publishes all three columns - on-demand, spot, and a third 'Inference Single GPU' rate - which lets you read the market rather than guess at it. On on-demand H100 ($6.155/GPU-hr) CoreWeave sits between AWS p5 ($6.88) and the neoclouds. On SPOT H100 ($19.71/8 = $2.46/GPU-hr) it is the cheapest published H100 hour anywhere in this slice, ~64% under AWS on-demand and ~29% under even Runpod's in-stock Pod rate. The third column is the one to understand before quoting: 'Pricing based on GPUs is exclusively applicable to the CoreWeave inference platform customers' with the footnote 'contact your account executive' - so the $6.16 'Inference Single GPU' H100 number is gated to their inference platform, NOT a bare-GPU rate you can walk in and buy, and it happens to equal the on-demand per-GPU figure anyway. Also note regional asymmetry: spot is CHEAPER in Europe than North America for H100 ($19.51 vs $19.71), H200 ($20.64 vs $20.93) and A100 ($9.51 vs $9.65) - if you are EU-based, Europe is the cheaper spot market, the reverse of the usual intuition. Region-restricted deployments carry a stated 1.5x premium. The GB200 NVL72 line is the outlier worth noting: $42.00/h for 72 Blackwell GPUs is $0.58 per GPU-hour, which is not comparable to the 8-GPU node pricing and suggests rack-scale pricing is a different economic model entirely - and it is quoted for Europe too, which is unusual for a rack-scale SKU. Spot is published per region and changes continuously. EXPORT CONTROL (2026-10-04, mechanism-level; sources spl-bis-2026-05-d5-guidance, spl-bis-2026-01-advanced-computing-case-by-case): renting capacity does NOT remove the export-control question, it relocates it. Providing remote access to controlled computing capability can constitute a transfer under the EAR, so the provider must screen the end user; and separately, under the 91 FR 1684 licence regime the consignee must run KYC procedures to prevent unauthorized remote access and must manage and limit IaaS access to the hardware - obligations pushed onto the customer, who must be told to implement them. For a customer headquartered or parented in Country Group D:5 or Macau the 742.6(a)(6)(iii)(A) end-user/parent test can bite even when the capacity is consumed in a permitted region. Recorded true because the underlying hardware is controlled and the provider carries the screening duty, not because access to a US region is itself restricted. LEAD TIME: null, and correct as null. For a listed on-demand instance type the provisioning delay is effectively zero - capacity is allocated per request - so a week count would be meaningless; the real gate for scarce shapes is either a quota increase or a capacity reservation, not a lead time. Where a published commitment horizon exists it IS recorded: AWS documents that a Capacity Block may be reserved with a start time up to EIGHT WEEKS in the future (source spl-aws-capacity-block-booking-horizon), which is a floor on how far ahead scarce capacity must be committed, not a wait. On-demand and Spot tiers publish no lead time at all; their constraint is availability, and the repo records that as availability=in-stock or a spot/reality note rather than as a fabricated week figure. Probe log in source spl-negative-no-oem-publishes-lead-time.