From $6.50 usage-based🇺🇸
GPU cloud for training and serving large AI models, with a managed Kubernetes service, Slurm integration and managed inference.
Lambda AI features and pricing (October 2026)
GPU cloud offering single instances, 1-Click Clusters and Superclusters on NVIDIA hardware for AI training and inference.
Usage-based
from $0.79 usage-based
Pricing details
Prices per GPU hour on the pricing page, from $0.79 for NVIDIA V100 and from $3.99 for H100, plus applicable taxes; clusters with terms from two weeks, as of September 2026.
What is Lambda?
Lambda rents NVIDIA GPU capacity for AI training and inference, from single cloud instances for experiments to multi-thousand GPU clusters for AI labs and enterprises.
Lambda describes itself as an AI-only infrastructure company. Its product line covers self-service instances, 1-Click Clusters with NVIDIA HGX B200 and H100 systems for distributed workloads, and Superclusters with NVIDIA GB300 NVL72 that run single-tenant with hardware isolation. The company was founded in San Francisco in 2012 and lists hyperscalers, regulated enterprises and frontier labs as customer groups. Its trust portal documents SOC 2 Type II and several ISO certificates.
Lambda AI features and use cases
- Tests models on single GPUsOn-demand instances give data scientists a GPU machine for prototyping, fine-tuning or running an open model. They pay per GPU hour and can shut the machine down when the experiment ends.
- Trains models on clusters1-Click Clusters provide 16 to more than 2,000 GPUs connected for distributed training, booked for terms from two weeks. Teams get cluster capacity without building their own data center.
- Isolates sensitive workloadsSuperclusters run single-tenant with a shared-nothing architecture. Companies with strict security requirements keep their training data and weights separate from other customers.
What sets Lambda apart
- Publishes per-GPU hourly prices for instances and clusters, including older GPUs such as V100 for low-cost experiments.
- Trust portal lists SOC 2 Type II plus ISO/IEC 27001, 27017, 27701 and ISO 22301.
Where Lambda is strongest
United States Available internationally
Lambda was founded in San Francisco in 2012 and names an office in San Jose in its privacy policy.
Questions to ask before you buy Lambda
Lambda's pricing page does not name data center regions, so ask where instances and storage run before moving personal or confidential data. Cluster bookings have fixed terms, which affects cost if a project ends early. The trust portal lists subprocessors, but it does not show a data processing agreement publicly, so request one directly.
Lambda reviews
No reviews yet. Be the first to share how your team uses it.
Every review shows its verification level. We read each one before it goes live. How we check reviews
Questions about Lambda
What does a Lambda GPU cost?
The pricing page lists on-demand prices per GPU hour, starting at $0.79 for NVIDIA V100; H100 instances start at $3.99 per GPU hour. Taxes are added.
Can I rent a whole cluster?
Yes. 1-Click Clusters offer 16 to more than 2,000 GPUs for terms from two weeks, and Superclusters are sold as single-tenant installations.
Which certifications does Lambda have?
The trust portal lists SOC 2 Type II, ISO/IEC 27001:2022, ISO/IEC 27017:2015, ISO/IEC 27701:2019 and ISO 22301:2019.
Lambda update historyno updates yet
No changes recorded for Lambda yet.
Changes the vendor made and confirmed user reports, each checked against the vendor's own pages. We review every tool monthly.
Alternatives to Lambda
Tools in the same category (GPU rental) with the most overlapping markets.
From $0.02 usage-based🇺🇸
Inference API for more than 100 open and commercial models billed per token, plus GPU instances and dedicated clusters.
Usage-based🇺🇸
GPU cloud with low-cost hourly Pods, serverless endpoints that scale to zero and clusters for developers and small teams.
From $1.50 usage-based🇺🇸
US AI cloud on its own data centers, with NVIDIA and AMD GPUs, managed Kubernetes and Slurm, and managed inference.