← The AI Systems Map
L0 · LabsFrontier labUnited States

Meta AI (FAIR)

Open-weight leader; maker of Llama and Superintelligence Labs

Meta's AI org (FAIR plus the 2025 Meta Superintelligence Labs) drove the open-weight movement with the Llama family, though Llama 4's mixed reception in 2025 prompted a strategy reset, a ~$14B investment into Scale AI, and aggressive talent hiring. It runs one of the largest private GPU fleets (hundreds of thousands of Nvidia H100/H200/Blackwell GPUs) across its own datacenters (Prometheus/Hyperion buildouts). Meta funds AI from its ad business rather than external cloud spend.

Flagship model

Llama (open weights)

Compute

Own datacenters, Nvidia GPU fleet

2025 move

Superintelligence Labs + Scale AI stake

Model

Open-weight distribution

How it fits the stack

Meta AI (FAIR) with what it depends on (above) and what it feeds (below). The figure renders as a crawlable diagram and upgrades to an interactive 3D graph as it scrolls into view.

depends onsourcesusesdepends oncompetes withcompetes withcompetes withpartners withMeta AI (FAIR)LabsAI Talent Market /Researcher CompchokepointCommon CrawlchokepointDeepSpeed / MegatronNvidia Data-Center GPU(Blackwell/Rubin)chokepointAlibaba Qwen (Tongyi)Mistral AIOpenAIScale AI
Meta AI (FAIR)Depends on ↑Related

Meta AI (FAIR) in the AI stack. Meta AI (FAIR) with its immediate upstream dependencies (top) and downstream dependents (bottom) in the AI value chain. Hover a node in 3D, or read the full relationships below.

Graph data (text) — 9 entities, 8 relationships
  • Meta AI (FAIR)depends onAI Talent Market / Researcher Comp
  • Meta AI (FAIR)sourcesCommon Crawl
  • Meta AI (FAIR)usesDeepSpeed / Megatron
  • Meta AI (FAIR)depends onNvidia Data-Center GPU (Blackwell/Rubin)
  • Meta AI (FAIR)competes withAlibaba Qwen (Tongyi)
  • Meta AI (FAIR)competes withMistral AI
  • Meta AI (FAIR)competes withOpenAI
  • Meta AI (FAIR)partners withScale AI