GPU.ai pricing methodology - Per-second billing, zero markup, secure vs community tiers

Fetched 2026-10-04. GPU.ai is a cloud aggregator that reads live pricing from 12+ GPU clouds and normalizes to per-GPU-hour rates. Methodology: multi-GPU node prices are divided by GPU count before setting a floor price. Zero markup - raw provider rates. Billed per second. Two tiers: SECURE (vetted datacenter capacity) and COMMUNITY (independent marketplace suppliers). Stock counts are launchable inventory, not theoretical capacity. Market range column shows the span of public on-demand list prices across major clouds for the closest equivalent SKU (never spot or committed-use discounts). Data refreshes hourly and is available via a public no-auth JSON API. The 90-day price trend chart shows daily floor prices. Key limitation: GPU.ai is itself a provider, so its 'secure' rates may include its own markup over the underlying cloud rates, despite the 'zero markup' claim. The 'community' rates are from third-party marketplace suppliers and carry higher risk. The market range column is the most useful for cross-provider comparison.