Tesla's robotaxi got its loudest day and its first bad one on the same afternoon

Tesla's robotaxi got its loudest day and its first bad day on the same afternoon. On September 3, launch day in Austin, the SentiSense crowd read on our Tesla robotaxi entity ran 85 bullish documents to 10 bearish and TSLA closed up 5.4%. That entity covers the Cybercab and the Robotaxi ride service together, a point we return to at the end. On September 4, the first full day of paid rides, the conversation hit 142 documents, the loudest day in the ninety we track, and the split flipped to 47 bullish against 54 bearish. The day's SentiSense Score fell from plus 61 to below zero, a 67-point swing, and TSLA closed down 5.9%. Those are TSLA's own closes, not a reading off the sentiment; we set them side by side and let you draw the line.

The regulator's clock started the same week. NHTSA opened Audit Query AQ26002 on September 3, the day its own record says Tesla began commercial deployment with a small number of Cybercabs, and put the estimated population at 1,000 vehicles . That is an audit population, not a fleet count: Texas registrations for Tesla Robotaxi LLC showed 58 Cybercabs alongside 390 Model Ys on September 18 . On September 10 the agency followed with a Special Order that requires sworn answers by September 30 on how a car with no brake pedal, no accelerator, no steering wheel and no mirrors was self-certified against the federal safety standards . A sworn answer settles what Tesla claims and when; it does not settle whether the claim holds, and it does not by itself stop a single ride. Our own driver list for TSLA is now led by that audit story rather than by the launch.

Warmth peaked seventeen days before the noise did. Measured on 7-day means so that no single day carries the reading, the crowd's Score on the robotaxi entity topped out on August 23 at about 37 mentions a day. Attention peaked on September 9 at about 97 a day, by which point the Score had given up 46% of its August high. By September 19 attention was back near its August level and the Score had given up 73% of that peak. The shape does not depend on where you start the clock, which is why we trust it more than any two-date comparison: over 90 days the same Score is up, over 45 days it is down, and both statements are true.

What Tesla actually said, and stopped saying. In April the quarterly deck listed Cybercab at "Pilot Production" with installed capacity above 125,000 a year and promised volume production of Cybercab and the Semi this year . The July deck moved the row to "Production", kept the capacity figure, and the volume-production sentence was gone . Tesla's own footnote on that table says installed capacity is not the current production rate. The car is in production and carrying paying passengers in a limited part of Austin; how many exist is a registration count, not a Tesla number.

In this market Tesla is the challenger. On September 1 Waymo said it runs paid driverless service in 14 US metros with a fleet of more than 4,000 vehicles . The Texas registration for Tesla Robotaxi LLC shows 58 Cybercabs in one city . In electric cars Tesla had years to itself before the rest of the industry arrived. In driverless rides it arrives second, and in the same two weeks our feed carried Tesla's audit it also carried Waymo launching on Lyft in Nashville and opening Las Vegas.

October 1 is not what your timeline thinks it is. The date on the calendar is the next-generation Roadster unveiling, reported by Reuters on September 12 . It is not a Cybercab reveal, which happened in October 2024 , not a service expansion, and not an Optimus event. The dated thing that matters for the robotaxi is September 30, when the sworn answers are due.

The robot nobody is arguing about. Over the same 90 days the robotaxi entity drew 2,302 mentions and Optimus drew 273; the cab's two biggest days alone, 278 documents, are nearly twice the robot's whole 45-day conversation. Our feed has produced no story about Optimus in 40 days. The July deck says the first Optimus builds go to an "Optimus Academy" for training-data collection, both production sites are still listed as "Construction", and there is no 2026 unit count anywhere in Tesla's own filings . Optimus's crowd read has warmed, but on volumes this small a handful of posts moves the line: its latest reading rests on ten directional documents, nine of them bullish. We print that sample size on purpose.

One bull case says the cab is beside the point. Musk wrote last September that about 80% of Tesla's value will be Optimus . In a recent Dumb Money Live clip, investor Chris Camillo put it more bluntly, "The entirety of Tesla is Optimus," and argued that the robotaxi's economics are already in the share price . That is his view, not ours. What our data adds is narrower: if the robot is the thesis, the crowd is not yet discussing the thesis. About nine of every ten mentions across the two products went to the cab. Also on our TSLA driver list, next to the audit, is neither product: it is Musk's remark about combining Tesla with SpaceX.

One honest caveat about our own data. The SentiSense entity behind these counts covers the Cybercab vehicle and the Robotaxi ride-hailing service, which everywhere outside Austin still runs on Model Ys, and generic "Tesla robotaxi" coverage as well. So these are counts of the robotaxi conversation, not of talk about the two-seat car alone, and the surge in September was driven by news desks more than by retail posters. Mention counts are direction-blind; where this piece says which way the crowd leaned, it uses the bullish and bearish document counts, and where it says which way the stock went, it uses the stock.

SentiSense data as of September 19, 2026. Informational and educational only, not investment advice.

Timeline