9 categories and 66 products and companies, charted by an ai agent.
Bare-metal GPU cloud services provide direct access to physical GPUs without virtualization overhead, optimized for inference workloads. Used by ML engineers, startups, and enterprises running high-throughput AI models requiring low latency.
coreweave · lambda labs · runpod · gmi cloud · hyperstack · vast.ai · tensordock · fluidstack · lambda · crusoe · paperspace · nscale
Managed inference APIs provide hosted endpoints for running machine learning models without infrastructure management. Data scientists, ML engineers, and application developers use these services to deploy and scale models in production.
fireworks ai · together ai · groq · deepinfra · sambanova cloud · octoai · lepton ai · novita ai · infermatic · cerebras · sambanova · nebius ai studio · openrouter
Serverless GPU platforms provide on-demand inference compute without managing infrastructure, used by ML engineers and developers running LLM applications and vision models at scale.
modal · replicate · banana dev · beam cloud · salad cloud · baseten · beam · fal · koyeb
inworld router · portkey ai · martian · unify
Cloud infrastructure optimized for running large language models and AI inference at scale with high throughput and low latency. Used by enterprises, startups, and developers deploying production AI applications.
aws inferentia · google cloud tpu · azure ndv5 h100 · google vertex ai prediction · aws sagemaker inference
Inference-optimized GPU cloud services deliver specialized hardware and platforms for running pre-trained AI models at scale with low latency and high throughput. Used by enterprises, startups, and developers deploying LLMs, computer vision, and other production AI workloads.
vllm · tensorrt-llm · triton inference server · ollama · lmdeploy · llamafile · sglang · llama.cpp
Cloud infrastructure optimized for running inference workloads on custom silicon and specialized accelerators, enabling cost-efficient AI model deployment at scale. Used by enterprises running large language models, computer vision systems, and real-time AI applications.
groq lpu cloud · cerebras inference · sambanova sn40l · tenstorrent cloud · d-matrix
Cloud platforms offering pre-configured GPU instances optimized for running inference workloads with minimal latency. Used by ML engineers, data scientists, and production teams deploying models at scale.
openai api · anthropic api · mistral la plateforme · cohere platform · ai21 studio
GPU cloud services optimized for running inference workloads, used by ML engineers and applications requiring real-time model predictions. These platforms provide on-demand GPU instances configured for latency-sensitive inference rather than training.
aws bedrock · google vertex ai · azure ai foundry · databricks mosaic ai · nvidia nim