NVIDIA
Dominant AI accelerator supplier
AI Infrastructure
The hardware and software stack powering the intelligence revolution.
29 companies · 6 segments
Silicon — Chips & Accelerators
AMD
MI300/MI350 GPU challenger
Cerebras
Wafer-scale AI chips (WSE-3)
Groq
LPU inference accelerators
Etched
Transformer-only ASIC (Sohu)
Tenstorrent
RISC-V-based AI processors
Networking — Interconnect & Fabric
Astera Labs
Rack-scale AI connectivity silicon
Enfabrica
Accelerated compute fabric chips
Broadcom
Tomahawk switches & custom AI ASICs
Cloud & Compute — Neoclouds
CoreWeave
Largest AI-native neocloud
Lambda
GPU cloud for training & inference
Crusoe
Vertically integrated AI cloud + DCs
Nebius
European AI cloud (ex-Yandex)
Together AI
Open-model inference & training cloud
Data & Storage
Pinecone
Managed vector DB for RAG
Weaviate
Open-source vector database
VAST Data
All-flash data platform for AI
Scale AI
Labeled training data & evals
Models — Frontier Labs
OpenAI
GPT models, ChatGPT
Anthropic
Claude; safety-first frontier lab
xAI
Grok; Colossus supercomputer
Mistral
European open-weight frontier lab
Cohere
Enterprise AI ($7B valuation, 2025)
DeepSeek
Chinese open-weight reasoning models
Orchestration & Serving
Anyscale
Managed Ray distributed compute
BentoML
Inference platform for any model
Modal
Serverless GPU compute
Fireworks AI
Fast inference for open models
Databricks
Lakehouse + Mosaic AI training