AI chip · Tesla
Tesla Dojo D1
Tesla's custom training accelerator for FSD neural nets; deployed in Dojo ExaPOD compute tiles at Tesla training clusters. Fabricated by TSMC.
Market position
Tesla's sovereign training silicon. 50 B transistors, mesh-connected in 25-chip tiles, tiles form ExaPODs. Used exclusively to train FSD models on Tesla's video corpus. Musk announced project shutdown Aug 2025 in favour of a rumoured 'Dojo 2' / consolidated AI5 strategy.
Efficiency and power
How much work you get per watt and per dollar, and how much power a full rack draws.
- Perf per watt
- 0.91 FP8 TFLOPS/W362 TFLOPS ÷ 400 W = 0.91
Common questions
How much does Tesla Dojo D1 cost?
Tesla Dojo D1 doesn't have a public launch list price. Vendors like this one usually price through direct sales rather than a public sheet. See list price per TFLOP for chips that do publish.
How much power does one Tesla Dojo D1 draw?
400 W at the chip. A full server draws more once you add CPU, memory, networking and cooling: see the rack power number above. Compare to other chips on perf-per-watt.
Key facts
- Class
- Compute SoC (AI accelerator)
- Designer
- Tesla
- Safety-critical?
- No (data-center inference / training)
- Record as of
- 2026-09-27
Specifications (9 fields, click to expand)
Straight from the vendor datasheet. Dense throughput shown first; sparse (2:4) numbers in parentheses where the vendor publishes them. Full datasheet linked below.
- Process node
- TSMC N7
- Transistors
- 50 B
- Die size
- 645 mm²
- TDP
- 400 W
- BF16 (dense)
- 362 TFLOPS
- FP8 (dense)
- 362 TFLOPS
- Form factor
- Training tile (25 D1 chips per tile)
- Announced
- 2021-08-19
- Released
- 2023-07-01
Source: vendor datasheet
Foundry & process
- Fabbed at
- TSMC (all chips TSMC makes →)
- Process node
- TSMC N7
- Transistors
- 50 B on 645 mm² die
Compare with
Sources
- Tesla AITesla
Adoption rows appear as we confirm each chip-in-robot pairing from a public source. See every chip we track for the full catalog or the Tesla page.