PERFORMANCE EVIDENCEComparable benchmark basis
Benchmark performance is not applied until a model family or explicit comparison group is selected.
Read the evidence methodology →No benchmark group is active. Market-price economics still work, but Kovara will not invent performance normalization.
LEADING CURRENT SCENARIOCoreWeave
NVIDIA H100 SXM · Europe, Multiple / Unspecified
Partial economicsTracked GPU price$2.44USD / GPU-hour
Storage$0.00
Interconnect$0.00
Modeled totalIncomplete730 billed hours/GPU
Effective utilized GPU-hourIncomplete65% utilization
Workload cost / 1M tokensIncompleteToken volume missing
Performance $ / 1M tokensIncompletePerformance unavailable
Evidence score80/100Completeness + freshness
SCENARIO COMPARISONCurrent economic paths
Ranking objective: Balanced economics. Evidence score is a modeling completeness / freshness indicator, not a probability.
3 viable01

CoreWeaveNVIDIA H100 SXM · Europe, Multiple / UnspecifiedOFR-0587
GPU price$2.44
TotalIncomplete
EffectiveIncomplete
Perf $/1M tokIncomplete
Evidence80/100
Inspect →02

ScalewayNVIDIA H100 PCIe 80GB · Paris / availability-zone dependent, FranceOFR-0126
GPU price—
TotalIncomplete
EffectiveIncomplete
Perf $/1M tokIncomplete
Evidence62/100
Inspect →03

ScalewayNVIDIA H100 SXM · Paris / availability-zone dependent, FranceOFR-0128
GPU price—
TotalIncomplete
EffectiveIncomplete
Perf $/1M tokIncomplete
Evidence62/100
Inspect →SENSITIVITYWhat changes the economics?
Recalculates the selected scenario under simple price and utilization shocks. It does not predict future prices.
CoreWeavePrice −10%Incomplete$2.19/GPU-h · 65% utilization · effective Incomplete/used GPU-h
Current scenarioIncomplete$2.44/GPU-h · 65% utilization · effective Incomplete/used GPU-h
Price +10%Incomplete$2.68/GPU-h · 65% utilization · effective Incomplete/used GPU-h
Utilization −15 ptsIncomplete$2.44/GPU-h · 50% utilization · effective Incomplete/used GPU-h
Utilization +10 ptsIncomplete$2.44/GPU-h · 75% utilization · effective Incomplete/used GPU-h
EVIDENCE LAYERWhat is measured, modeled and missing
The economics model does not collapse these categories into a single claim.
OFR-0587TRACKED- Comparable USD / GPU-hour price is tracked on the current public Offering.
- Accelerator VRAM is tracked at 80 GB per GPU.
- Compute cost is calculated from tracked hourly price × requested GPU quantity × billed hours.
- Known cost subtotal: $56,969.20. This excludes unpriced components and is not a complete total.
- Network bandwidth metadata is tracked at 200 Gbps.
- Local storage capacity is tracked at 67,553.9944 GB; storage price is not inferred.
ASSUMPTIONS- Billed planning month is 730 hours unless the workload uses an explicit run duration.
- Utilization assumption is 65.0%.
- No egress volume was requested, so egress cost is excluded from the scenario.
- This scenario models compute commitments, requested egress, storage, setup, interconnect and entered overhead; taxes and unlisted charges require contract review.
MISSING / NOT INFERRED- Minimum term is undisclosed; no complete contractual total is claimed.
- Setup fee is undisclosed or invalid; it is not assumed to be zero.
- No applicable public comparable performance benchmark or explicit throughput assumption is available.