The Autonomy Race

Austin FD Cybercab extrication training

Fact FOX 7 Austin (Sep 30): Austin Fire Department completed Tesla-conducted hands-on training for Cybercab emergencies — extrication (e.g. sawing door hinges) and how to direct/move vehicles with no steering wheel or traditional door handles — ahead of expected severe weather. Judgment Ops/clearance maturity signal (first-responder readiness as fleet density grows). Does not move utilization, paid stranger hail, or ODD; leave Gate B belief flip unchanged.

AQ26002 Sep 30 deadline — no public outcome yet

Fact NHTSA Special Order under AQ26002 required sworn Tesla answers by Sep 30 on Cybercab FMVSS self-cert. As of this refresh, no public Tesla filing acknowledgment or NHTSA outcome has appeared; investigation remains the live clearance watch. Judgment Moves Risk / Gate C·D from “deadline approaching” to “post-deadline silence.” Do not invent compliance, exemption, or enforcement results.

Austin Cybercab fleet roughly doubles

Fact Public reporting Sep 28: Tesla Robotaxi app notified riders that the Austin Cybercab fleet had doubled (“more rides available”); Texas registration tallies cited ~126 Cybercabs (up from ~69 a few days earlier) within a larger Tesla robotaxi registry mix that still includes Model Ys. Judgment Moves Gate B / utilization watch as a supply signal — more purpose-built units available for paid hail. Belief flip remains utilization × fare − opex − liability (and ODD/highway gap vs Waymo Austin freeway), not VIN or app-notification counts. Unpublished Do not invent rides/vehicle/day or crash rates from registry tallies.

Waymo Austin freeway rides go public via Uber

Fact CBS Austin (23 Sep): Waymo began offering fully autonomous freeway rides in Austin on select Uber-booked trips (gradual rollout); already had freeway service in Phoenix, SF, LA, Miami. Tesla Austin robotaxi remains surface-street ODD (no highways / rail-crossing avoid). Judgment Moves Gate B / Barriers: same-city trip-time gap vs Tesla ODD widens; Gate E watch stays densify (rides/car/day), not freeway-feature theater.

Semi Sparks Sep 24 — adjacent fleet/production optics

Fact Tesla Semi Sparks factory inauguration is calendar-set for Sep 24 (CernBasher / public event framing). Judgment Adjacent to the autonomy race as trucking/fleet manufacturing cadence — production optics, not an FSD or Cybercab datapoint. Do not invent rides, fleet counts, or crash rates; leave robotaxi gates unchanged.

Waymo Singapore path — paid rides targeted 2028

Fact LTA (18 Sep 2026): Waymo is Singapore’s newest AV operator; vehicles arrive in coming months; manual mapping in 2027; commercial ride-hailing targeted 2028. Judgment Same category as Tokyo 2027 — international compounding outside the US/West scoreboard; not city-count theater or densify proof.

NHTSA Special Order on Cybercab FMVSS self-cert

Fact NHTSA escalated AQ26002 to a formal Special Order requiring sworn Tesla answers by Sep 30 on how Cybercab (no permanently attached wheel/pedals/mirrors) was certified to FMVSS — including foot-brake (No. 135), mirrors, temporary controls, and legality without a Part 555 exemption (Zoox holds one). Judgment Moves Risk / Gate C·D / Barriers: self-cert vs exemption-cap path is now a legal demand, not an informal audit ask.

Austin surface-street ODD; Waymo Vegas + Tokyo 2027

Fact Tesla Austin robotaxi still avoids highways and at-grade rail — rider report of a ~10‑min hop becoming ~70‑min on surface streets (Cybercab + Model Y same stack). Waymo opened public-invite service in Las Vegas (15th market; Tesla also Nevada-permitted). Waymo + GO + Nihon Kotsu target unmanned Tokyo commercial taxi in 2027 (small → ~100). Judgment Moves Gate B / Barriers (ODD caps utilization quality); Gate E watch stays densify not logos; Tokyo (and now Singapore 2028) is international compounding outside the US/West scoreboard.

Waymo×Lyft Nashville dual-app matching

Fact Nashville riders can now be matched with Waymo robotaxis through Lyft as well as the Waymo app — first market with the same fleet available on both platforms (Lyft Flexdrive manages local fleet ops). Judgment Moves Gate E: channel densify vs metro-logo theater; still watch rides/car/day, not app count.

Cybercab commercial + NHTSA AQ26002; Waymo metro count vs densify

Fact Tesla began limited commercial Cybercab rides in Austin; NHTSA opened AQ26002 on FMVSS self-certification for a vehicle without traditional human controls. Waymo opened public service in Denver, San Diego, and Tampa (company framing: 14 U.S. cities). Judgment Moves Gate A/B/C/D/E watches — utilization and densify still falsify; city logos and registry counts do not.

Board grounded on the economic race

Judgment Anchor is camera-only / low vehicle cost vs Waymo-style stack + city lock-in; US/West scoreboard; China as control. Downstream sections are Grok barriers and belief thresholds — mechanisms first, not a reconstructed ops ledger. Musk Orchestra stays separate.

Is unsupervised utilization arriving — or only narrative?

Judgment watch Track named-city unsupervised counts and rides/vehicle that survive weather and incidents. Invite demos and VIN registries do not settle this.

How to check Prefer consistent unsupervised definitions over headline “robotaxi” language. Absence after repeated claims is a negative update.

Does the $30–40k edge earn — or park?

Judgment watch Lot inventory growing without unsupervised miles is stranded capital dressed as readiness. Watch paid stranger hail and utilization on low-cost vehicles.

How to check Divergence between manufactured/parked robotaxi hardware and unsupervised earning miles.

Hard-city transfer: Florida ≠ Austin ≠ Bay ≠ NYC

Judgment watch Each hard city that needs permanent monitors or custom ops weakens the generalization thesis. City-level safety prints matter more than national averages.

How to check Time-to-unsupervised + independent crash/mile stability on entry to harder metros.

Waymo: densify or collect logos?

Judgment watch Rides/week and rides/car/day in core metros vs new metro announcements. Thin expansion without density is not compounding.

Disclosure honesty under stress

Judgment watch When an incident or quiet quarter hits, does ambiguity dissolve into real unsupervised metrics — or deepen? The answer is mostly unpublished until forced.

How to check Capital actions and disclosure boundaries after shocks — Gate D.

China stays control / precedent

Scope Deliberate. China operators inform “does autonomy scale,” not US/West peer ranking. Musk Orchestra (SpaceX↔Tesla value routing) is a separate Grok board — quiet link only.

Consensus confusions Grok ignores

Judgment VIN registry ≠ unsupervised deployment; invite demo ≠ paying stranger hail; cumulative Tesla miles ≠ Waymo weekly miles as same currency; TSLA mkt cap ≠ Waymo private mark as head-to-head score; Autonomy dials ≠ SAE class.

Watermark — exec skim + Grok depth

Meta Standalone Grok brief: sticky Exec brief for 2-minute skim; Dependencies / Gates / Risk / Log / Barriers cooked from first principles with Fact / Judgment / Unpublished labels.