# Lambda

Canonical: https://inetgeek.com/gpu/lambda/

GPU cloud from an AI hardware vendor, billed by the minute per GPU with no egress fees.

## Who it suits

Renting exactly the instance shape you need — Lambda publishes 1x, 2x, 4x and 8x plans for the same card, so a single H100 is a real product rather than a fraction of a node.

Consider something else if You want the cheapest single H100: its H100 SXM is $4.29 per GPU-hour single-card, more than Hyperstack ($3.20) and RunPod ($3.99) charge for the same card, though under Crusoe and Nebius at $4.50, and reserved capacity is quoted through sales rather than published.

## iScore

56 / 100 (80% interval 50–62), rank 5 of 13 in gpu — interval overlaps the next rank, so this is an ordering, not a finding.

inetGeek's own reading, computed from sourced facts against category peers by disclosed rules. Not a provider claim. Intervals are 80%; a provider with few documented facts is pulled toward the category middle rather than scored on what little is known.
Read this score with care:
- Price is thin: 3 of 10 criteria are scored, so it sits near the category middle with a wide interval (35–56). Read the interval, not the point.
- Capabilities is thin: 0 of 1 criteria are scored, so it sits near the category middle with a wide interval (94–100). Read the interval, not the point.

Method: https://inetgeek.com/iscore/. Full output with every criterion's basis: https://inetgeek.com/scores.json. Every gpu peer ranked: https://inetgeek.com/gpu/lambda/alternatives/

- Price: 45 (35–56), weight 50%, 3 of 10 criteria scored; scored for peers, not here — undocumented or stated without a number or a yes/limited/no: budget_input_price, flagship_input_price, flagship_output_price, free_card_required, free_offer_type, free_tier, storage_price
- Capacity and limits: 49 (36–62), weight 24%, 2 of 3 criteria scored; scored for peers, not here — undocumented or stated without a number or a yes/limited/no: context_window
- Capabilities: 100 (94–100), weight 3%, 0 of 1 criteria scored; scored for peers, not here — undocumented or stated without a number or a yes/limited/no: prompt_caching
- Trust and compliance: 73 (67–79), weight 13%, 10 of 12 criteria scored; scored for peers, not here — undocumented or stated without a number or a yes/limited/no: support_included, uptime_sla
- Evidence: 89 (89–89), weight 11%, 4 of 4 criteria scored

## What the documentation says

### Pricing

- Entry paid plan: Pay-as-you-go per GPU-hour rather than a monthly plan: the cheapest self-serve instance is a 1x Quadro RTX 6000 at $0.69 per GPU-hour plus sales tax. ([source](https://lambda.ai/pricing), read 2026-10-17)
- Pricing model: On-Demand instances are priced by hourly usage and billed in one-minute increments; 1-Click Clusters per GPU per hour billed weekly; filesystems per GiB used per month. ([source](https://docs.lambda.ai/public-cloud/billing/), read 2026-10-17)
- Egress over allowance: Free. Lambda states transparent pricing with no egress fees. ([source](https://lambda.ai/service/gpu-cloud), read 2026-10-17)

### Infrastructure

- Regions: 14 On-Demand Cloud regions: nine in the USA plus Tokyo, Osaka, India, Germany and Israel. Not every instance type is available in every region. ([source](https://docs.lambda.ai/public-cloud/on-demand/), read 2026-10-17)

### GPU

- H100 SXM, per GPU-hour: $4.29 per GPU-hour for H100 SXM 80GB on a single-GPU instance, falling to $3.99 on an 8-GPU node. The PCIe variant is $3.29. ([source](https://lambda.ai/service/gpu-cloud), read 2026-10-17)
- Cheapest GPU, per hour: $0.69 per GPU-hour for a Quadro RTX 6000 with 24GB; the V100 is $0.79 and the A6000 $1.09. ([source](https://lambda.ai/service/gpu-cloud), read 2026-10-17)
- Largest GPU offered: 180GB per GPU on B200 SXM6, from $6.69 per GPU-hour on an 8-GPU node to $6.99 on a single GPU. ([source](https://lambda.ai/service/gpu-cloud), read 2026-10-17)
- Billing granularity: By the minute. ([source](https://lambda.ai/service/gpu-cloud), read 2026-10-17)
- Reserved pricing: Reserved capacity is quote-only. Committed 1-Click Clusters list $6.16-$5.54 per H100 GPU-hour and $9.86-$8.87 per B200 GPU-hour for 2 weeks-1 year, 16 to 256 GPUs. ([source](https://lambda.ai/pricing), read 2026-10-17)

### Compliance

- SOC 2: SOC 2 Type II attestation, listed under the security and compliance pillar of Lambda's trust framework. ([source](https://lambda.ai/trust), read 2026-10-17)
- ISO 27001: ISO 27001 certified, alongside ISO 27017, 27701 and 22301. ([source](https://lambda.ai/trust), read 2026-10-17)
- HIPAA: Platform Guidelines, binding with the Cloud Terms of Service, prohibit submitting personal health information to the Services. Contractual prohibition on PHI, not a statement about a HIPAA audit or BAA. ([source](https://lambda.ai/legal/terms-of-service), read 2026-10-17)
- GDPR / data residency: Processes European personal data under the GDPR and relies primarily on the European Commission's Standard Contractual Clauses for EEA, UK and Swiss transfers. ([source](https://lambda.ai/legal/privacy-policy), read 2026-10-17)
- PCI DSS: Platform Guidelines, binding with the Cloud Terms of Service, prohibit submitting payment card or financial account information to the Services. Contractual prohibition on cardholder data, not a statement about a PCI DSS assessment. ([source](https://lambda.ai/legal/terms-of-service), read 2026-10-17)
- Encryption at rest: All data on the 1-Click Cluster persistent file storage system is encrypted at rest using industry standard algorithms and parameters. ([source](https://docs.lambda.ai/public-cloud/1-click-clusters/security-posture/), read 2026-10-17)
- Customer-managed keys: Superintelligence/Supercluster page states each cluster uses AES-256 encryption at rest with unique keys under your control. Stated for single-tenant Superclusters, not for public-cloud on-demand instances. ([source](https://lambda.ai/superintelligence), read 2026-10-17)
- Role-based access control: Two account-level roles, Admin and Member; both have full access to resources, while only admins invite users, change billing and manage workspaces. ([source](https://docs.lambda.ai/public-cloud/access-security/), read 2026-10-17)
- Private networking / VPC: 1-Click Cluster compute nodes sit on a dedicated network segment with no inbound firewall connectivity, on a logically isolated Ethernet fabric. ([source](https://docs.lambda.ai/public-cloud/1-click-clusters/security-posture/), read 2026-10-17)
- Dedicated infrastructure: 1-Click Cluster compute nodes run on single-tenant hardware with GPUs, storage, memory and NICs unshared; management head nodes are multi-tenant. ([source](https://docs.lambda.ai/public-cloud/1-click-clusters/security-posture/), read 2026-10-17)

## Compared with

- [Lambda vs CoreWeave](https://inetgeek.com/compare/coreweave-vs-lambda/)
- [Lambda vs RunPod](https://inetgeek.com/compare/runpod-vs-lambda/)
- [Lambda vs Hyperstack](https://inetgeek.com/compare/hyperstack-vs-lambda/)
- [Lambda vs Vast.ai](https://inetgeek.com/compare/vast-ai-vs-lambda/)