The Open & Sovereign AI Ecosystem
Open AI now means more than open weights. Chinese labs (DeepSeek, Qwen, Kimi/Moonshot, MiniMax, Z.ai/Zhipu) alongside Meta, Mistral and others are pushing strong models into the open ecosystem, while routing layers such as OpenRouter, Portkey, Merge Gateway and LiteLLM make multi-model deployment practical for companies that do not want to depend on a single model provider."Open" in the AI era is contested, so we grade it: the model layer is split by how open each release really is, from permissively-licensed open weights to source-available and restricted licences. The wider map also includes proprietary-but-independent enablers — coding tools, model gateways, sandboxes, memory layers and physical-AI platforms like Wayve — that support firm sovereignty by preserving choice, auditability and deployment control. Read this as an ecosystem map of the open and sovereign AI stack, not an OSI-only landscape.
Source-available & restricted models
Agent sandboxes
Physical AI & autonomy
Agent frameworks & platforms
Persistent & autonomous agents
Agent memory & context
Coding assistants & devtools
Search & relevance engines
Data pipelines & transformation
Labeling, feedback & data governance
LLM observability & tracing
Evaluation & testing
Safety, guardrails & red-teaming
Workflow automation
Durable execution & orchestration
API & developer workflow tools
Productivity & collaboration
Internal tools & databases
CRM, ERP, support & identity
Design, CMS & commerce
Open web / browser / search