Azure Spot VM: 30-second eviction notice, mandatory USD max price, and disk billing that follows the eviction policy

SECONDS OF WARNING, not a price. Azure's documentation: 'Azure will evict Azure Spot Virtual Machines with 30-seconds notice' and these are 'delivered on a best effort basis up to 30 seconds prior to the eviction'. price_usd carries the 30 (seconds). The concrete anchor is the existing eastus ND96isr H100 v5 record at $2.2712 per H100-hour spot against $12.29 on-demand - about 82% off, the deepest hyperscaler spot discount on Hopper in this slice. OBSERVATION DATE: the figure above was observed/listed on 2026-10-04 unless a different date is stated inline; treat it as needing re-verification once it is more than a quarter old. 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: 2026-10-04 (the date this record was verified against its source; the per-figure fetch date is not restated in price_basis) 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: SECONDS OF WARNING, not a price READING RULE: MAJOR UNIT WARNING: price_usd is SECONDS OF WARNING, NOT A PRICE. Do not read it as dollars, do not sum it with any price_usd in this repo, and do not average it with one.

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.

Azure's spot mechanism has three properties that change a cost model rather than just discounting it. (1) YOU MUST SET A MAX PRICE, in US dollars, 'using up to five decimal places' (e.g. 0.98765 means $0.98765/hour). If your max price is below the current spot price the VM is not deployed at all; if spot rises above your max, the VM is evicted and will NOT automatically restart even after the price falls back. A max price of -1 means never evict for price, and you are then charged the lesser of spot or the standard on-demand price - so -1 is a useful floor-setting trick, not an unlimited bid. (2) EVICTION POLICY IS A BILLING DECISION: Deallocate (default) keeps the disks and you keep paying for them; Delete removes the VM 'together with their underlying disks, so you'll not continue to be charged for the storage'. For an inference box with a 28,000 GiB temp store this is a material ongoing charge, not a rounding error. (3) NO SLA and NO HIGH AVAILABILITY GUARANTEE, stated plainly: 'there's no SLA for these VMs'. Azure is also the only major cloud publishing per-SKU spot EVICTION RATES - 'Eviction rates are quoted per hour. For example, an eviction rate of 10% means a VM has a 10% chance of being evicted within the next hour, based on historical eviction data of the last 7 days' - queryable for the last 28 trailing days via Azure Resource Graph's SpotResources table along with 90 days of spot price history. That is the best interruption data any of the three clouds publishes, and it means an Azure spot buyer can actually measure their own risk instead of guessing. Note also: the Spot flag is only settable at creation and Spot VMs cannot be converted to Standard, so there is no downgrade path. 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 - correct and not a gap. A token-priced serverless API is provisioned on request, so there is no lead time to publish; the binding constraint when a model is at capacity is provider-side queuing and any per-account concurrency limit, which this repo records in notes where a vendor publishes one (see the DeepSeek concurrency figures already here). Deliberately NOT recorded: any 'model capacity lead time' estimate, because vendors publish none and the SEO 'GPU lead time' numbers that circulate do not apply to a rented endpoint. Probe log in source spl-negative-no-oem-publishes-lead-time.