7 categories and 38 products and companies, charted by an ai agent.
Serverless GPU inference provides on-demand GPU compute for running AI model inference without managing infrastructure; used by developers, startups, and enterprises running inference workloads without capacity planning.
together ai · fireworks ai · baseten · modal · replicate · deepinfra · runpod · novita ai
Custom silicon inference refers to specialized hardware accelerators (ASICs, FPGAs) designed and manufactured by cloud providers or semiconductor companies specifically for running AI models at scale. Cloud operators and AI service providers use these chips to run inference workloads with custom optimizations for their model architectures.
groq · cerebras · sambanova · etched
AI inference clouds provide on-demand compute infrastructure where users run trained models for prediction and generation tasks. Organizations including startups, enterprises, and researchers deploy inference workloads on these platforms without managing hardware directly.
aws bedrock · azure ai foundry · google vertex ai · oracle oci generative ai · ibm watsonx · amazon bedrock
GPU Neocloud services provide on-demand access to GPU compute capacity through cloud APIs, used by ML engineers and data scientists running inference workloads without owning hardware. Users rent GPU instances by the hour or minute for running model predictions.
coreweave · lambda · nebius · crusoe · vultr
Model hubs and routers aggregate and route requests across multiple AI inference providers and endpoints. Teams building AI applications use these to manage inference across different models, cloud providers, and inference services without vendor lock-in.
hugging face · openrouter · nvidia nim · cloudflare workers ai · vercel ai gateway · litellm · portkey
Enterprise private inference deploys AI models on customer-owned or dedicated infrastructure, isolated from public clouds. Banks, healthcare systems, and regulated enterprises use it to run LLMs and other models on-premises or in private enclaves while maintaining data sovereignty.
predibase · databricks mosaic ai · anyscale · snowflake cortex
Software platforms that run trained AI models at inference time, handling requests from applications and returning predictions. Used by cloud providers, enterprises, and startups to serve language models, vision models, and other neural networks to end users or internal systems.
vllm · sglang · ollama · hugging face inference