The AI landscape is vast, and it's easy to get lost in.
New models, new tools, new names land constantly. Some of it is noise. Some of it changes what we'd recommend to you. We keep track of which is which.
Strip away the noise and it's six layers.
Open any layer.
The companies we use, deploy, and track for clients. What each one makes, and where it sits. These are the biggest names in each layer, not every name.
There's no official map of AI. Chipmakers draw it as a hardware stack, investors draw market maps, and vendors draw it around whatever they sell. Six layers is how we explain it to clients: each layer is a different kind of decision, from which model to build on to where your data lives.
Nothing here is sponsored. Last reviewed July 2026.
Builds general-purpose models for chat, coding, and images, for consumers and via API.
Builds the Claude model family, with products for chat, autonomous work, coding, and API access.
Builds multimodal models and AI tools across consumer apps, developer APIs, and research.
Builds the Grok model family for chat, image and video generation, voice, and developer APIs.
Builds Meta AI and the Muse frontier models; keeps open tools like PyTorch and SAM, plus the earlier open Llama family.
Builds enterprise models and a secure agent platform, with a retrieval and multilingual focus.
European lab that releases open-weight models alongside hosted chat, coding, and developer platforms.
Chinese lab releasing open-weight language and reasoning models, with a free chat app and API.
Alibaba's open-weight model family spanning text, multimodal, and coding, with a chat platform and API.
Chinese lab releasing the open-weight Kimi models, with a chat app, agents, and a developer API.
Chinese lab releasing the open-weight GLM model family for chat, coding, and reasoning, with a chat app and API.
Abu Dhabi's Technology Innovation Institute, releasing the open-weight Falcon models, including Arabic-language models.
Copilot assistants across Microsoft 365, GitHub, and Windows, plus its own MAI models, built on Azure.
AI answer engine with a web browser, a desktop agent, and a developer search API.
AI code editor with autonomous agents, inline completion, and automated code review.
Connected workspace with built-in AI agents that draft, search, and run work across your docs and tools.
Open-source autonomous agent operated through chat, running tasks on your own machine.
Maker of Devin, the autonomous software engineer, spanning cloud agents and the Devin Desktop editor.
Autonomous agent that plans and completes multi-step research and web tasks on its own.
Browser platform that turns prompts into deployed apps, with hosting built in.
AI builder that generates and hosts full web apps from plain-language prompts.
Text-to-image and image-to-video generation, used through its web app and Discord.
AI video generation, in-context video editing, and world models for film and production work.
TikTok's parent company; builds the Seedance video and Seedream image models and the Doubao assistant.
Image-generation models offered both as a hosted API and as open weights for self-hosting.
Voice synthesis, conversational voice agents, music, and audio tools, via apps and an API.
AI avatar video, multilingual video translation, and an API for programmatic video.
AI image and video generation with camera-motion control, plus lip-sync and marketing tools.
Text- and image-to-video generation with synced audio, built by the Chinese company Kuaishou.
Chinese lab behind the Hailuo video platform, alongside open-weight language models and a developer API.
Video generation and creative AI agents; its Ray models turn text and images into video.
AI music generation. Full songs, vocals included, from a text prompt, via web and mobile apps.
Visual workflow automation for building AI agents and integrations, self-hostable or cloud-hosted.
No-code automation across thousands of apps, with AI agents and an MCP bridge to the major assistants.
Visual automation platform for connecting apps and building AI agents into everyday business workflows.
Open-source frameworks for building LLM agents, plus a hosted platform for tracing, evals, and deployment.
Open-source framework and hosted platform for document parsing and retrieval over complex data.
Open-source framework for orchestrating multi-agent systems, with a hosted management platform.
Type-safe Python agent framework from the Pydantic team, paired with the Logfire observability platform.
Supplies the chips, software platforms, and open models that AI systems are built and run on.
Cloud platform with managed model APIs, its own Nova models, ML infrastructure, and custom AI chips.
Repository and platform for open models, datasets, and demos, with hosted inference.
One API in front of hundreds of models from every major lab, with routing, fallbacks, and unified billing.
Inference provider running open-weight models on its own LPU hardware, in the cloud or on-prem.
Builds wafer-scale AI chips and a fast inference cloud, an alternative to GPU-based compute.
Cloud platform for running, fine-tuning, and hosting open-weight models on managed GPUs.
Cloud platform for running open-weight models via API, with an open-source packaging tool.
Web infrastructure for apps and agents, with a TypeScript AI SDK, a multi-provider model gateway, and AI dev tools.
Runs open-weight models locally via a CLI and local API, with an optional hosted service.
Desktop app for discovering and running local models, with a headless server, CLI, and SDKs.
You don't need to know all of this. You need someone who does.
That's the job. We track the landscape so you can make one good decision at a time. Book a call and tell us what you're weighing.