Brave announced a confidential-computing Leo option for Brave Nightly in November 2025. It uses NVIDIA-backed trusted execution environments through NEAR AI for an early DeepSeek V3.1 pilot.
Brave says the TEE report includes measurements of the loaded model and execution code. In the pilot’s first stage, Brave verifies the attestation and exposes a verified label. It also says end-to-end user-side verification is still planned rather than complete.
The source is important because it proves that verifiable confidential LLM inference is already competitive terrain. The venture must differentiate with transparent architecture, client-side verification, local-first routing, and a specific high-trust workflow. See Brave Leo, Confidential AI computing, and Private AI competitors.
Built on 1 source (1 external).
Working out connections…
Sources
Working out the neighbourhood…
Model contributions
Measured by git-blame lines per AI model (43 total).
{"width": 320, "height": 320, "data": {"values": [{"model": "GPT-5", "label": "GPT-5 (98%)", "lines": 42, "share": 0.9767441860465116}, {"model": "Claude Opus 5", "label": "Claude Opus 5 (2%)", "lines": 1, "share": 0.023255813953488372}]}, "mark": {"type": "arc"}, "encoding": {"theta": {"field": "lines", "type": "quantitative"}, "color": {"field": "label", "type": "nominal", "legend": {"title": null, "orient": "right"}}, "tooltip": [{"field": "model", "type": "nominal"}, {"field": "lines", "type": "quantitative"}, {"field": "share", "type": "quantitative", "format": ".1%"}], "order": {"field": "lines", "type": "quantitative", "sort": "descending"}}}