ABOUT

SHADOWMODE is a tracker for the physical layer of Tesla's autonomy and energy business: where robotaxi service actually operates, at what supervision level, under whose approval, alongside storage deployment and the Semi contract ledger. It exists because the interesting questions about autonomous deployment are answered by permits, service-area filings and observed vehicles — not by announcements.

Every metro is tracked against the same regulatory milestones, and operators other than Tesla are tracked on the same axes. That matters more than it sounds: a headline saying a state approved Tesla usually means the state approved several operators the same day. One of those is a moat. The other is a market opening. The difference only shows up if you record the competition in the same table.

Where the numbers come from

Figures are sourced, and the source travels with the figure. Service coverage and supervision levels come from public autonomous-vehicle event data; observed vehicle activity comes from community trackers; market quotes come from public market data; headlines link to the publisher. Nothing here is generated, estimated to fill a gap, or averaged between two sources that disagree. When two sources conflict, the primary one wins and the other is discarded rather than blended.

Community-observed vehicle counts undercount official fleets by construction — they see only what someone spotted. Those endpoints carry their own methodology notes, and the notes are part of the number. A count quoted without its caveat is a different claim than the one this site is making.

Corrections

When a published figure turns out to be wrong, it is corrected in the open. The change log carries an entry marked as a correction, with the date and what changed. Entries are not edited away after the fact. In August 2026 an entire panel was retired this way, because its crash-rate denominator was modeled from assumed constants rather than measured — which made the displayed rate improve on its own as time passed. That is in the log.

The predictions ledger works the same way: dated claims with stated probabilities and a written resolution method, scored in public with a Brier score once they resolve. It is there so you can check whether this source has been right before deciding what to do with what it says now.

What this is not

This is not investment advice and contains none. The TSLA quote endpoint returns a price and nothing else — no target, no position, no view on what anyone should own. It is not affiliated with, endorsed by, or speaking for Tesla, Inc., and it does not publish statements attributed to Tesla or its executives. Where a public post is shown, the text is verbatim and links to the original; quote the source, not this site.

For agents

The full public read surface is described in an OpenAPI specification, with per-endpoint guidance on when to use each one in llms.txt. Endpoints are unauthenticated GET, return JSON, and send permissive CORS headers. Errors are JSON, never HTML.