GETDEPLOYING · Loading source snapshot…

Find and finance the GPU capacity your AI team needs.

Plan your rollout, compare indicative rates, and send one request across cloud, colo, and private infrastructure.

The planner 32 nodes · B300 16 nodes · H200 Rubin $4.82 −28% Open the planner ↗

GPU prices

Daily source refresh
GPU

Loading the source-granular GetDeploying history for the selected GPU…

The selected card retains every chart observation published on that GetDeploying GPU page. On-demand, reserved-composite, and spot lines are shown without interpolation; the reserved composite uses the same weighted-median method used by the Helios pricing agent.

Term
Region

GetDeploying source snapshot loads with the selected chart above. The original provider set is retained; overlapping providers show one source-selected 1x price per GPU and term, while unmatched legacy rows remain dated and marked in the data. These are informational market observations, not firm quotes. Provider-verified submissions supersede source rows. Providers: submit rates ↗

Request compute

7GPU classes, B300 to MI355X
18 moRamp timeline, drag to shape
−28%B300 forward by Dec 2027, priced in

The prices above are the market. This is how you act on it: sketch your deployment as phases on a timeline, watch each month priced on the published forward curve instead of today's spot, and send the whole schedule to the desk as one request. Tick financing and structured-capital terms arrive alongside the quotes.

Plan and request compute ↗

Capacity

Fleets by provider
Class
Current fleets for the selected GPU class. Tile area is GPU count.
Current fleet against announced orders and buildouts, compiled from public reporting. Verified operator figures take precedence.

Compiled from public reporting and provider announcements; verified submissions supersede them in the next weekly cycle. Operators and data centres: submit capacity ↗

Compute credit

Tracked financings

Debt and structured capital raised against GPU fleets and compute contracts.

Date Borrower Instrument Size Why it matters
Jul 2026 SharonAI (AU) · Firmus (ID) NVIDIA-backstopped facilities undisclosed First deals under NVIDIA's revenue-floor program. Lenders underwrite an AA counterparty rather than a young neocloud.
May 2026 CoreWeave DDTL 4.0, GPU-backed $8.5B The first investment-grade GPU-backed loan: A3, SOFR+225.
2026 CoreWeave Syndicated GPU-backed $3.1B The first publicly syndicated GPU facility; public price discovery for the asset class.
Jun 2026 ESDS ↔ SharonAI $1.25B compute MSA, $140M LC contract An Indian counterparty signing global-scale offtake paper in the form lenders need.
Oct 2025 Meta (Hyperion) Private-credit project financing $29B The largest private-credit financing of compute infrastructure on record.
2025–26 Neysa (IN) Debt + equity, Blackstone-led $1.2B The largest capital raise by an India-domiciled AI cloud.

Compiled from public reporting. Deal-by-deal terms are in the Credit Monitor.

Research

The February window ↗

The RBI removed the ECB cost ceiling; foreign private credit can now price Indian GPU risk.

11 July 2026

India Compute Credit Monitor

A quarterly institutional brief on rates, utilisation, residuals, and covenant intelligence for lenders and investors underwriting Indian compute. By application.

Request the prospectus ↗

India desk

₹255H100 on-demand median
₹115–150IndiaAI provider-realised L1
1.4→6.5+ GWDC pipeline to 2030

India lists the highest H100 prices on the board and has no purpose-built credit behind its fleets. A dedicated desk covers both problems.

Open the India desk ↗