Tesla Dojo
Tesla's in-house training supercomputer silicon for autonomy (wound down 2025)
Tesla designed the Dojo D1 tile-based training chip and supercomputer to train its Full Self-Driving vision models, an ambitious wafer-scale-style architecture built with TSMC. In August 2025 Tesla disbanded the Dojo team and pivoted toward its AI5/AI6 inference chips plus large Nvidia and AMD GPU purchases, effectively ending Dojo as a standalone training program. It remains notable as a case study in the difficulty and cost of building custom training silicon.
Chip
Dojo D1 tile (training supercomputer)
Status
Team disbanded Aug 2025
Pivot
AI5/AI6 inference + merchant GPUs
How it fits the stack
Tesla Dojo 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.
Tesla Dojo in the AI stack. Tesla Dojo 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) — 3 entities, 2 relationships
- Tesla Dojo —depends on→ TSMC (Taiwan Semiconductor Manufacturing Company)
- Tesla Dojo —competes with→ Nvidia
Context — capital, rivals, policy · · 1