◆ Gentoo Logic · Modeling Warehouse

AI/LLM · AI architecture

12 AI/LLM companies — how they build with AI

How AI/LLM companies build with AI — which foundational models they run and how directly, ranked by proximity to the model. Avg Foundation Proximity Score 100/100.

Native 12Direct API 0Cloud-hosted 0Middleware 0

AI21 Labs

AI/LLM · Multi-Cloud · Native (own / self-hosted weights)

Trains Jamba, its own hybrid Mamba-Transformer long-context models.

Aleph Alpha

AI/LLM · On-Prem · Native (own / self-hosted weights)

German sovereign-AI lab training its own Pharia/Luminous models with explainability for gov/regulated EU.

Anthropic

AI/LLM · Multi-Cloud · Native (own / self-hosted weights)

Trains the Claude family across Google Cloud TPUs and AWS Trainium (both partners/investors). Pioneered the open Model Context Protocol (MCP) for tool/data connectivity around its models.

Cohere

AI/LLM · Multi-Cloud · Native (own / self-hosted weights)

Enterprise-focused LLM lab training its own Command/Embed/Rerank models, served across clouds (Google Cloud, Oracle, AWS) and deployable in-VPC for data-sensitive customers. North platform targets RAG.

Contextual AI

AI/LLM · Multi-Cloud · Native (own / self-hosted weights)

RAG-native enterprise platform training its own grounded language models.

Hugging Face

AI/LLM · AWS · Native (own / self-hosted weights)

The model/dataset hub on AWS, storing weights as Git LFS repos over S3 and running Inference Endpoints/Spaces. Stewards open libraries (Transformers, Diffusers) — the open-source center of gravity for ML.

Inflection AI

AI/LLM · Azure · Native (own / self-hosted weights)

Trains its own Inflection models (now enterprise-focused) on Azure supercompute.

Liquid AI

AI/LLM · Multi-Cloud · Native (own / self-hosted weights)

MIT-spinout training its own non-transformer Liquid Foundation Models for efficient/edge inference.

Mistral AI

AI/LLM · Multi-Cloud · Native (own / self-hosted weights)

European lab releasing strong open-weight models (Mistral/Mixtral) alongside commercial ones; serves via its own La Plateforme and partner clouds (Azure, GCP). Le Chat is the consumer surface. Data-sovereignty positioning.

OpenAI

AI/LLM · Azure · Native (own / self-hosted weights)

Trains and serves its own GPT frontier models on Microsoft Azure supercomputers (custom GPU clusters). Product data on Azure stores; the API/Responses platform fronts the models for developers.

Reka AI

AI/LLM · Multi-Cloud · Native (own / self-hosted weights)

Multimodal frontier lab (ex-DeepMind/Meta) training its own Reka model family.

xAI

AI/LLM · On-Prem · Native (own / self-hosted weights)

Trains the Grok frontier family on its own Colossus supercluster; real-time data via X.

← Back to the full orbit map (all 260 companies) · Score your own stack →