# Lambda

Official pricing: https://lambda.ai/pricing  
Category: gpu-cloud · Isolation: vm

## Pricing regimes (raw)

- **B200 SXM6, on-demand (×1)** (resource): $0/vCPU-h, $0/GiB-h
- **GH200, on-demand (×1)** (resource): $0/vCPU-h, $0/GiB-h
- **H100 SXM, on-demand (×1)** (resource): $0/vCPU-h, $0/GiB-h
- **H100 PCIe, on-demand (×1)** (resource): $0/vCPU-h, $0/GiB-h
- **A100-40GB SXM, on-demand (×1)** (resource): $0/vCPU-h, $0/GiB-h
- **A100-40GB PCIe, on-demand (×1)** (resource): $0/vCPU-h, $0/GiB-h
- **A10, on-demand (×1)** (resource): $0/vCPU-h, $0/GiB-h
- **RTX-A6000, on-demand (×1)** (resource): $0/vCPU-h, $0/GiB-h

## Features

Yes: persistent disk, volumes, ≥24 h sessions, full VM (own kernel), Docker inside, root, SSH, public IPv4, HTTPS preview URLs, raw TCP inbound, open internet, private networking, SOC 2, SSO, GPU, arm64, Python, Node.js, audit logs, EU data residency, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

No: snapshots, memory snapshots, fork/clone, pause/resume, idle auto-stop, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, browser, desktop GUI, computer-use API, browser + desktop control, anti-bot stealth, CAPTCHA solving, residential IPs, preinstalled agents, egress allowlist, static egress IP, self-hosting, BYOC, open source, HIPAA, Windows, macOS, HTTP method/path egress rules, wake on request, live resize, live fork (no pause), memory fork, scoped API keys, MCP server, automatic snapshots, snapshots on demand, agent harness API, hosted agent API (their own agent), secret proxy, secret proxy for any API

Unknown: everything else. Evidence (source + quote) per feature: https://battleships.dev/data/providers/lambda.json → feature_evidence

## Caveats

- CPU/RAM/local SSD included. Billing one-minute increments from launch health-check success to termination, including idle. Per-GPU price changes with node count. No spot list found.
- Any null count, CPU/RAM, billing increment or minimum is unverified, not unlimited/free.
- This is GPU-product research, not a claim that every feature of the entire vendor documentation was audited. Unverified features are null.
- GPU multi-count bundles, variant/region selection and fractional slices require a SKU-aware estimator; unsupported rows remain unpriced in strict cards.
- Modes with gpu=null are intentionally excluded, not free GPU modes. Never substitute 0 for a null rate. Flags alt indicate DIY/ML infrastructure, not the same thing as managed untrusted-code sandboxes.
- Unknown fee, storage or bandwidth means full monthly total may be unavailable. Rates exclude taxes.
- Instances include local SSD (512 GiB to 4 TiB by SKU; card uses the 512 GiB minimum). Persistent filesystems are billed per GiB-month in hourly increments; docs quote $0.20/GiB-month only as an example ("might not reflect current pricing"), so no filesystem rate is encoded. Filesystems have no ingress/egress charges.
- GH200 (on-demand-gh200-1) is an Arm (Grace) host: x86 containers may not run.
- No stop/suspend: instances can only be launched, restarted or terminated; 'sudo shutdown' puts the instance in Alert status and billing continues. Idle GPU time is billed at the full rate.
- Weekly invoicing; refunds only as service credits that expire after 12 months; major credit cards only.
- Filesystem usage cannot be capped; up to 24 filesystems per account; us-south-1 filesystems limited to 10 TB; filesystems must be attached at launch and are region-bound.

## How this provider charges

As of 2026-09-28. On-Demand GPU Cloud instances and 1-Click Clusters.

CPU/RAM/local SSD included. Billing one-minute increments from launch health-check success to termination, including idle. Per-GPU price changes with node count. No spot list found.

| Regime | When applicable | Billing | Numbers ($/GPU-hour unless slice) | Source |
|---|---|---|---|---|
| on-demand  | B200 SXM6; count [1]; Region/availability must be checked at launch. | per-minute; min 60s | 6.99; list; per_gpu_hour | [source](https://lambda.ai/pricing) |
| on-demand  | GH200 ; count [1]; Region/availability must be checked at launch. | per-minute; min 60s | 2.29; list; per_gpu_hour | [source](https://lambda.ai/pricing) |
| on-demand  | H100 SXM; count [1]; Region/availability must be checked at launch. | per-minute; min 60s | 4.29; list; per_gpu_hour | [source](https://lambda.ai/pricing) |
| on-demand  | H100 PCIe; count [1]; Region/availability must be checked at launch. | per-minute; min 60s | 3.29; list; per_gpu_hour | [source](https://lambda.ai/pricing) |
| on-demand  | A100-40GB SXM; count [1]; Region/availability must be checked at launch. | per-minute; min 60s | 1.99; list; per_gpu_hour | [source](https://lambda.ai/pricing) |
| on-demand  | A100-40GB PCIe; count [1]; Region/availability must be checked at launch. | per-minute; min 60s | 1.99; list; per_gpu_hour | [source](https://lambda.ai/pricing) |
| on-demand  | A10 ; count [1]; Region/availability must be checked at launch. | per-minute; min 60s | 1.29; list; per_gpu_hour | [source](https://lambda.ai/pricing) |
| on-demand  | RTX-A6000 ; count [1]; Region/availability must be checked at launch. | per-minute; min 60s | 1.09; list; per_gpu_hour | [source](https://lambda.ai/pricing) |
| on-demand  | RTX6000 Quadro; count [1]; Region/availability must be checked at launch. | per-minute; min 60s | 0.69; list; per_gpu_hour | [source](https://lambda.ai/pricing) |
| on-demand  | B200 SXM6; count [2]; Region/availability must be checked at launch. | per-minute; min 60s | 6.89; list; per_gpu_hour | [source](https://lambda.ai/pricing) |
| on-demand  | H100 SXM; count [2]; Region/availability must be checked at launch. | per-minute; min 60s | 4.19; list; per_gpu_hour | [source](https://lambda.ai/pricing) |
| on-demand  | A100-40GB PCIe; count [2]; Region/availability must be checked at launch. | per-minute; min 60s | 1.99; list; per_gpu_hour | [source](https://lambda.ai/pricing) |
| on-demand  | RTX-A6000 ; count [2]; Region/availability must be checked at launch. | per-minute; min 60s | 1.09; list; per_gpu_hour | [source](https://lambda.ai/pricing) |
| on-demand  | B200 SXM6; count [4]; Region/availability must be checked at launch. | per-minute; min 60s | 6.79; list; per_gpu_hour | [source](https://lambda.ai/pricing) |
| on-demand  | H100 SXM; count [4]; Region/availability must be checked at launch. | per-minute; min 60s | 4.09; list; per_gpu_hour | [source](https://lambda.ai/pricing) |
| on-demand  | A100-40GB PCIe; count [4]; Region/availability must be checked at launch. | per-minute; min 60s | 1.99; list; per_gpu_hour | [source](https://lambda.ai/pricing) |
| on-demand  | RTX-A6000 ; count [4]; Region/availability must be checked at launch. | per-minute; min 60s | 1.09; list; per_gpu_hour | [source](https://lambda.ai/pricing) |
| on-demand  | B200 SXM6; count [8]; Region/availability must be checked at launch. | per-minute; min 60s | 6.69; list; per_gpu_hour | [source](https://lambda.ai/pricing) |
| on-demand  | H100 SXM; count [8]; Region/availability must be checked at launch. | per-minute; min 60s | 3.99; list; per_gpu_hour | [source](https://lambda.ai/pricing) |
| on-demand  | A100-80GB SXM; count [8]; Region/availability must be checked at launch. | per-minute; min 60s | 2.79; list; per_gpu_hour | [source](https://lambda.ai/pricing) |
| on-demand  | A100-40GB SXM; count [8]; Region/availability must be checked at launch. | per-minute; min 60s | 1.99; list; per_gpu_hour | [source](https://lambda.ai/pricing) |
| on-demand  | V100 16GB; count [8]; Region/availability must be checked at launch. | per-minute; min 60s | 0.79; list; per_gpu_hour | [source](https://lambda.ai/pricing) |
| reserved  | H100 ; count [16]; Region/availability must be checked at launch. | weekly-increments; min unknowns | 6.16; list; per_gpu_hour | [source](https://lambda.ai/pricing) |
| reserved  | H100 ; count [64]; Region/availability must be checked at launch. | weekly-increments; min unknowns | 5.85; list; per_gpu_hour | [source](https://lambda.ai/pricing) |
| reserved  | H100 ; count [256]; Region/availability must be checked at launch. | weekly-increments; min unknowns | 5.54; list; per_gpu_hour | [source](https://lambda.ai/pricing) |
| reserved  | B200 ; count [16]; Region/availability must be checked at launch. | weekly-increments; min unknowns | 9.86; list; per_gpu_hour | [source](https://lambda.ai/pricing) |
| reserved  | B200 ; count [64]; Region/availability must be checked at launch. | weekly-increments; min unknowns | 9.36; list; per_gpu_hour | [source](https://lambda.ai/pricing) |
| reserved  | B200 ; count [256]; Region/availability must be checked at launch. | weekly-increments; min unknowns | 8.87; list; per_gpu_hour | [source](https://lambda.ai/pricing) |

## Gotchas
- CPU/RAM/local SSD included. Billing one-minute increments from launch health-check success to termination, including idle. Per-GPU price changes with node count. No spot list found.
- Any null count, CPU/RAM, billing increment or minimum is unverified, not unlimited/free.
- This is GPU-product research, not a claim that every feature of the entire vendor documentation was audited. Unverified features are null.
- GPU multi-count bundles, variant/region selection and fractional slices require a SKU-aware estimator; unsupported rows remain unpriced in strict cards.

## Worked example (required protocol workload)
4 vCPU / 8 GiB, 50 concurrent × 8 hours/day × 22 days = **8,800 instance-hours/month**. 30% CPU utilization does not reduce allocated GPU uptime. 50 GiB snapshots and 100 GiB egress are additional. The requested CPU-only workload has no GPU type/count, so its full GPU-provider total is **null**, not a fictitious CPU equivalent.

Explicit GPU sensitivity example: add **1 × B200** to each of those 50 workers at the B200 quoted shape. Compute component = 8,800 × $6.99 = **$61,512.0000**. CPU/RAM are included in that SKU (not independently resizable).
This is compute-only, before platform fees, storage, egress, setup, minimum rounding, cold-start/idle tails and capacity quotas. 50 concurrent GPUs are not promised by a unit rate.

## Evidence scope
Sources read are listed in the provider/features JSON. Raw pricing pages, relevant docs and API payloads are retained under `../raw/`. All unknown feature toggles remain null.