An operator provides 5 million driverless rides in a week

Where things stand (14 Oct 2026)

Waymo is the only operator anywhere near the threshold, and it is still at about a tenth of it.

• Official weekly figures: more than 400,000 rides a week across six US metros (Waymo blog, 2 Feb 2026). Then 500,000 paid rides a week across 10 US cities (Waymo post, March 2026), repeated by Pichai on the Q1 call in April. Since then Waymo has only repeated "more than 500,000" (CNBC 1 Sep 2026; Waymo's Tokyo post 14 Sep; Motley Fool 11 Oct).
• Fleet: more than 4,000 vehicles, of which about 300 are the new Ojai and roughly 950 more Ojai are staged.
• Target: 1 million a week by the end of 2026. FutureSearch puts its Q4-2026 median at 713,000 (p10 537,000, p90 1.0 million) and gives about a 1-in-10 chance of hitting the million.
• Growth has slowed. California weekly paid trips rose from about 122,000 (end 2024) to 315,000 (June 2026), but trips per day fell from April to June 2026. Weekly rides went from roughly 175,000 to 450,000 in 2025 (about 2.6x), but 2026 looks like only about 1.6x.
• Supply is lining up: about 5,100 Ojai imports expected by the end of 2026, Hyundai Ioniq 5 deliveries starting Q4 2026 (reports say up to 50,000 by 2028), $16 billion of equity plus a $5 billion loan (8 Oct 2026), and California approval in August to expand across 18 counties.

Tesla has about 1,000 robotaxis (NHTSA estimate, Oct 2026). It passed 1 million unsupervised miles on 3 Sep 2026, says 300+ Cybercabs serve Austin, and has a 125,000-a-year Cybercab line that batteries currently limit. It does not disclose ride counts. Zoox and Uber's own fleets are far smaller.

Reference class and model

The threshold needs about a 10x rise from Waymo's level. The last 10x (50,000 in May 2024 to 500,000 in March 2026) took 22 months, but growth multiples are shrinking (about 5x, then 2.3x, then roughly 1.7x year over year).

Analyst views of 2030 differ widely:

• Wells Fargo: 465 million Waymo rides in 2030, about 9 million a week on average.
• Morgan Stanley: about 1 billion miles and $2.5 billion revenue in 2030, about 2.5 million a week.
• Goldman: a $19 billion US robotaxi market in 2030.

FutureSearch's median for Other Bets revenue reaching $1 billion a quarter (Feb 2030) implies roughly 4 million a week around then.

I ran a Monte Carlo simulation in the container (/workspace/model.py):

• Waymo: starts at 0.75 million a week at the end of 2026; 2027 growth median 1.9x with yearly decay; a 5% yearly chance of a stall. This gives a median of about 5 million at the end of 2030 and a chance of crossing 5 million of about 8% by end-2028, 31% by 2029, 50% by 2030 and 61% by 2031.
• Reporting friction: a qualifying statement must be on an official channel, weekly, paid, and US-specific. Once London or Tokyo operate, a global total won't count. Waymo has also gone about 10 months without an official weekly figure before (Apr 2025 to Feb 2026), and has sometimes used monthly or annual totals. I modelled a 65% chance that the first report qualifies, 45% that a later one does if not, with a lag of about 4 months.
• Tesla: about a 20% chance of crossing by the end of 2031, times 0.65 for reporting, modestly correlated with Waymo. Others add about 2.5%.

Output: about 0.04 by end-2028, 0.20 by 2029, 0.39 by 2030 and 0.52 by 2031. Plausible parameter changes put the 2031 figure between 0.45 and 0.59. I smoothed it slightly to 0.515.

Paths and counterarguments

• Paths to YES: Waymo's fleet ramp (Ojai plus Ioniq) brings growth back to about 2x a year through 2028–29 and it announces "over 5 million paid rides per week in the US". Or Tesla scales unsupervised Cybercabs across Sun Belt states.
• Paths to NO: the 2026 slowdown persists (rides per vehicle are falling and new cities ramp slowly), the Ojai import route is cut off, a safety or political backlash hits, or demand saturates at premium prices. Or growth is real but reported only as global, monthly or annual totals.
• Strongest case for lower: with growth decaying from 1.6x, Waymo would reach only 2–3 million a week by 2031. Strongest case for higher: only about 1.6x a year is needed, and capital and vehicles are in place.

I disregarded a thin Manifold market putting 16% on 5 million a week in 2027. Near-term quarters are near zero: 5 million a week needs about 40,000 vehicles, against roughly 5,000 operating driverless today across both companies.

Sources: waymo.com/blog (2 Feb 2026, 8 Oct 2026); blog.google Q2-2026 remarks; TechCrunch 27 Mar and 24 Sep 2026; futuresearch.ai/waymo-financial-forecast; roadtoautonomy.com (23 Aug 2026); OWID/CPUC passenger-km series (ourworldindata.org/grapher/passenger-kilometers-traveled-self-driving-taxis); goldmansachs.com robotaxi note (Apr 2026); Business Insider Tesla Q2 (22 Jul 2026); Teslarati; NeoTeo (Oct 2026); businessmodelanalyst.com Waymo SWOT (Oct 2026).

Question. By the last day of each quarter, has any operator reported providing at least 5 million paid driverless rides in a single week in the United States?

Definitions.

• Operator: a business operating a passenger road-transport service, including through contracted fleet managers or booking platforms. Any operator qualifies, not just a covered developer. Rides across the same operator's cities, brands, and controlled subsidiaries may be combined; rides of independent operators may not be combined merely because they share a booking platform or vehicle supplier. A parent company's report about its operating subsidiary counts.
• Paid ride: a completed passenger trip for which the passenger or a third-party customer pays a positive fare, including discounted fares and trips paid through a subscription. Free promotional trips, empty repositioning, deliveries, tests without paying passengers, and cancellations do not count. A booking carrying several passengers is one ride, not one ride per passenger; separately booked pooled journeys may each count once.
• Driverless: the automated driving system performs the driving throughout the passenger trip, with no onboard human assigned to drive or supervise the driving as a safety fallback and no remote human directly driving the vehicle. Remote assistance that provides information or advice without directly controlling the vehicle does not disqualify a trip. Supervised driver-assistance trips do not count, even if marketed as self-driving.
• United States: the 50 states and District of Columbia; only trips wholly within this area count.
• Single week: any one consecutive seven-day reporting period, not necessarily a calendar week. An operator's explicit statement that it is currently providing at least 5 million qualifying rides “a week” or “per week” counts as a report of achieved weekly volume even if it does not name the dates. A target, forecast, capacity estimate, or annualized run rate does not count.

Resolves YES if an eligible operator or its parent publishes a report of at least 5,000,000 qualifying rides for one week, or an explicit achieved weekly-volume statement meeting the definition above. The reported figure must establish that the threshold is met by that operator's paid, driverless US rides; a global or mixed-service total counts only if the qualifying US portion is separately reported or has an explicit lower bound of at least 5,000,000. Separate city figures may be summed only for the same week, with no double counting. An unqualified reported figure of 5 million counts; “about,” “nearly,” or a range spanning below 5 million does not establish the threshold.

The Observation period convention applies. For an undated current weekly-volume statement, the operator must present it as current at publication, rather than merely repeat a previously reported historical milestone.

Resolves NO if no qualifying report has been published by the quarter-end. Multiweek averages, cumulative totals, daily figures extrapolated to a week, and targets do not substitute for a reported single-week total or explicit current achieved weekly volume. Reports mixing free and paid rides or supervised and driverless rides do not count unless the qualifying portion alone meets the threshold.

Resolving source. Operator- or parent-authored statements on official domains, official press releases, investor-relations earnings releases or presentations, and regulatory filings such as SEC Forms 10-K, 10-Q, or 8-K and operator-authored reports filed with transport regulators. Examples include Waymo at waymo.com/blog, Alphabet at abc.xyz/investor, Tesla at ir.tesla.com, Zoox at zoox.com, and Amazon at ir.aboutamazon.com; equivalent official sources from other operators qualify. This is a narrower source rule than Public confirmation (A2): independent media reports, executive interviews, employee social posts, and third-party estimates alone do not count. Apply Resolution date (A3) and the Resolution window convention; a dated qualifying publication or archived dated update is required, and once YES it remains YES for later horizons.