Compute on-demand: H200 starting at $0.47/HR to $1.99/HR, B300 at $4.99/HR, B200 at $3.49/hr, H100 at $2.43/HR (Spot $0.94/HR). Reserved clusters available with quotes from 50+ datacenters.
Key features
GroqMetal: dedicated bare-metal infrastructure tuned for speed and reliability
GroqCore: production-ready inference stack with no infrastructure expertise required
GroqAssured: enterprise-grade governance, auditability, and control
256 LPUs per rack with 40 PB/s SRAM bandwidth
1,000 tokens per second per user throughput
128 GB on-chip SRAM per rack
315 PFLOPS of FP8 inference compute
13 globally distributed data centers across four continents
2,500+ RL environments on the Hub
100+ open-source model evaluations
Managed training workflows with full visibility
GLM-5.3 hosted inference
OpenAI-compatible API
SLURM and Kubernetes orchestration
Infiniband networking for distributed training
Grafana monitoring dashboards
Secure code sandboxes for RL
What makes it different
Fast inference at scale without sacrificing affordability
LPU pioneered technology integrated with NVIDIA GPUs
Fully integrated platform combining infrastructure, inference, and control
Integrated stack combining training, inference, compute and evaluations in one platform
Backed by leading investors including NVIDIA, Founders Fund, and AI researchers
Large-scale open-source community with 2,500+ environments and models
Applied research team providing hands-on support for training
Competitive pricing with spot market and reserved cluster options
Prime Intellect