Does one Cybercab pay for itself?
21 towns so far, from 1,846 people to 2.7 million, each run through the same 90‑day season with one car: twelve fixed seeds, the first option at every decision, an 18‑hour shift, charging overnight.
All 21 towns hand back the $10,000 you put in. How long is the only real question: about 1.4 months in Ely, NV, 13.3 in Buffalo, NY.
Above roughly 6 to 8 rides a day — measured, not guessed — one car earns more than it costs to run. That is a much lower bar than getting your money back, which is why the interesting question is not whether a town works but how long it takes, and what happens when you take the reason away.
Dover, Tennessee is the case in point. 1,918 people, and it pays — but 84% of its rides are visitors to Fort Donelson National Battlefield. Its residents alone generate about 5 rides a day, which loses money. You can close the battlefield in the game and watch the town fail.
The ladder
Smallest town first. You put $10,000 into the project, and day one leaves it worth $5,366 once the car, the permits and the home charger are paid for — both figures read from the model itself, never typed in. “Money back in” is how long, at the rate its first season set, before the project is worth the $10,000 again. It is a straight-line projection, so treat it as a direction rather than a date. Getting there inside three months would be a remarkable business, so the chip reads fast, steady or slow — none of them is a failure. Every figure here is already after the 30% the network takes off the top of each fare — our estimate, since Tesla has published no owner-operator terms, and the largest single lever in the model. Under each town is the research its row rests on.
| Town | People | Money back in | Worth at day 90 | Worst season of 12 | Ran out of cash |
|---|---|---|---|---|---|
| Malta, MTslow | 1,846 | 11.6 months | $6,568 | $289 | 1 of 12 |
Original Bought to test nowhere to go at all: 169 miles to the nearest town of 10,000.80 numbers behind this row: 26 published, 32 worked out from published figures, 22 our own estimates.9 citations fetched and checked: 8 held, 1 failed and was demoted to an estimate.3 gaps the research could not fill.Researched 9 September 2026. | |||||
| Dover, TNfast | 1,918 | 2.2 months | $11,756 | $1,001 | 1 of 12 |
Original The small band of the first batch. It came out viable ONLY because of Fort Donelson National Battlefield: strip the battlefield and the town fails.79 numbers behind this row: 23 published, 33 worked out from published figures, 23 our own estimates.10 citations fetched and checked: 6 held, 4 failed and were demoted to an estimate.4 gaps the research could not fill.Researched 9 September 2026. | |||||
| Alturas, CAfast | 2,634 | 2.9 months | $10,225 | $1,294 | 1 of 12 |
Requested Not a market pick. It was the run that proved research could go through the Grok CLI subscription instead of the paid API, at a cost of $0.00, and it was published because the work was already done. Its nearest commercial airport is 113 miles away.80 numbers behind this row: 28 published, 30 worked out from published figures, 22 our own estimates.13 citations fetched and checked: 11 held, 2 failed and were demoted to an estimate.3 gaps the research could not fill.Researched 9 September 2026. | |||||
| Presidio, TXslow | 2,909 | 9.1 months | $6,890 | $112 | 1 of 12 |
Original Bought to test poverty, heat, a border crossing, and NO airport within 120 miles.74 numbers behind this row: 26 published, 27 worked out from published figures, 21 our own estimates.13 citations fetched and checked: 12 held, 1 failed and was demoted to an estimate.5 gaps the research could not fill.Researched 9 September 2026. | |||||
| Winner, SDsteady | 2,969 | 3.6 months | $9,275 | $571 | 1 of 12 |
Original Bought to test plains agriculture and a seasonal hunting trade.77 numbers behind this row: 23 published, 32 worked out from published figures, 22 our own estimates.12 citations fetched and checked: 7 held, 5 failed and were demoted to an estimate.4 gaps the research could not fill.Researched 9 September 2026. | |||||
| Clifton, AZslow | 3,752 | 7.4 months | $7,236 | $274 | 1 of 12 |
Original Bought to test one big employer, a copper mine, instead of tourism.80 numbers behind this row: 28 published, 29 worked out from published figures, 23 our own estimates.17 citations fetched and checked: 10 held, 7 failed and were demoted to an estimate.4 gaps the research could not fill.Researched 9 September 2026. | |||||
| Scott City, KSsteady | 3,855 | 3.3 months | $9,555 | $1,193 | 1 of 12 |
Original Bought to test the thinnest county in the batch: irrigated farmland, 4,815 people in the county, no rescue.77 numbers behind this row: 24 published, 31 worked out from published figures, 22 our own estimates.15 citations fetched and checked: 8 held, 7 failed and were demoted to an estimate.4 gaps the research could not fill.Researched 9 September 2026. | |||||
| Ely, NVfast | 3,946 | 1.4 months | $15,117 | $1,833 | 1 of 12 |
Original Bought to test whether a MINOR magnet, a highway and casinos, is enough.78 numbers behind this row: 24 published, 32 worked out from published figures, 22 our own estimates.11 citations fetched and checked: 8 held, 3 failed and were demoted to an estimate.4 gaps the research could not fill.Researched 9 September 2026. | |||||
| Worland, WYsteady | 4,730 | 4.2 months | $8,639 | $317 | 1 of 12 |
Requested Bought to test the top of the small band: cold and industrial, and 62.9 miles from the airport it shares with Cody, which is past the 35-mile line at which a field stops counting as a town airport.80 numbers behind this row: 27 published, 27 worked out from published figures, 26 our own estimates.19 citations fetched and checked: 9 held, 10 failed and were demoted to an estimate.5 gaps the research could not fill.Researched 11 September 2026. | |||||
| Cody, WYfast | 10,377 | 1.4 months | $15,071 | $7,288 | 0 of 12 |
Requested Asked for by Dave, who has spent time in and around Cody and wanted to see what the research said about it. It sits 1.7 miles from the airport that Worland is 62.9 miles from, so the two together isolate what airport proximity is worth.80 numbers behind this row: 30 published, 29 worked out from published figures, 21 our own estimates.16 citations fetched and checked: 12 held, 4 failed and were demoted to an estimate.2 gaps the research could not fill.Researched 11 September 2026. | |||||
| Quincy, ILfast | 38,918 | 1.6 months | $14,205 | $6,437 | 0 of 12 |
Original One of the two worlds the game shipped with, before any research existed. A TOWN: one service area, 2.25 miles of radius. It has since been re-grounded on a researched pack.79 numbers behind this row: 28 published, 34 worked out from published figures, 17 our own estimates.14 citations fetched and checked: 10 held, 4 failed and were demoted to an estimate.2 gaps the research could not fill.Researched 9 September 2026. | |||||
| Stillwater, OKfast | 50,046 | 2.0 months | $12,286 | $7,087 | 0 of 12 |
Requested The first town a reader asked for and got. Suggested through the form, confirmed by email 226 seconds later, researched the same night. The note named the Oklahoma State University enrolment and the county seat, and the Census agreed with the population to within 46 people.80 numbers behind this row: 30 published, 34 worked out from published figures, 16 our own estimates.15 citations fetched and checked: 12 held, 3 failed and were demoted to an estimate.1 gap the research could not fill.Researched 9 September 2026. | |||||
| Sarasota, FLfast | 58,458 | 1.5 months | $14,363 | $9,002 | 0 of 12 |
Requested Asked for by Dave. It carries the largest airport in the whole set by an order of magnitude, 2,112,417 boardings four miles out, against a metro of 479,958 -- and it still only just beats Ely, Nevada, population 3,946. One car can take so many fares a day and no more, so past that point extra demand buys nothing.80 numbers behind this row: 27 published, 30 worked out from published figures, 23 our own estimates.13 citations fetched and checked: 9 held, 4 failed and were demoted to an estimate.5 gaps the research could not fill.Researched 12 September 2026. | |||||
| San Buenaventura (Ventura), CAfast | 109,914 | 2.9 months | $10,209 | $1,399 | 1 of 12 |
Original Planned as a mid-size coastal market. It is also the Census name trap: typed as Ventura it matches NOTHING, and only the legal name San Buenaventura (Ventura) returns a population, land area, airport or climate.80 numbers behind this row: 33 published, 30 worked out from published figures, 17 our own estimates.17 citations fetched and checked: 16 held, 1 failed and was demoted to an estimate.3 gaps the research could not fill.Researched 9 September 2026. | |||||
| Columbia, MOsteady | 130,851 | 3.5 months | $9,362 | $4,284 | 0 of 12 |
Requested Asked for by Dave. A university town of 130,851 with 1,781 rides a day, and it lands twelfth of eighteen -- slower to pay back than Alturas, California, population 2,634. The clearest case on the list that market size is not the thing that decides this.80 numbers behind this row: 29 published, 34 worked out from published figures, 17 our own estimates.13 citations fetched and checked: 11 held, 2 failed and were demoted to an estimate.1 gap the research could not fill.Researched 12 September 2026. | |||||
| Buffalo, NYslow | 274,613 | 13.3 months | $6,413 | $662 | 2 of 12 |
Requested Asked for by Dave, to reach a state the ladder had never touched. It is the first real transit city in the set: 22.6% of Buffalo households have no car, three times anything else here, against a cold 48.8F year. Five utilities file a service territory for Erie County and the federal file could not say which one the city is on, so the research pass settled it: National Grid.77 numbers behind this row: 27 published, 32 worked out from published figures, 18 our own estimates.17 citations fetched and checked: 11 held, 6 failed and were demoted to an estimate.2 gaps the research could not fill.Researched 13 September 2026. | |||||
| Wichita, KSsteady | 400,987 | 4.9 months | $8,194 | $1,791 | 1 of 12 |
Original The medium band of the first batch.79 numbers behind this row: 24 published, 33 worked out from published figures, 22 our own estimates.13 citations fetched and checked: 7 held, 6 failed and were demoted to an estimate.1 gap the research could not fill.Researched 9 September 2026. | |||||
| Colorado Springs, COslow | 494,743 | 7.0 months | $7,361 | $223 | 2 of 12 |
Requested Asked for by Dave, for a state the ladder had never touched. At 494,743 people it is the largest market added since Chicago, with 1.2 million airport boardings five miles out. Its research held 13 of 14 citations, the best ratio of any town on the list.80 numbers behind this row: 31 published, 35 worked out from published figures, 14 our own estimates.14 citations fetched and checked: 13 held, 1 failed and was demoted to an estimate.2 gaps the research could not fill.Researched 13 September 2026. | |||||
| Las Vegas, NVfast | 679,817 | 2.7 months | $10,590 | $4,837 | 0 of 12 |
Original The other world the game shipped with. A METRO: four districts, and the morning decision is which one each car works.78 numbers behind this row: 30 published, 30 worked out from published figures, 18 our own estimates.20 citations fetched and checked: 13 held, 7 failed and were demoted to an estimate.2 gaps the research could not fill.Researched 9 September 2026. | |||||
| Nashville, TNfast | 721,074 | 2.9 months | $10,205 | $4,697 | 0 of 12 |
Requested Asked for by Dave. The biggest airport on the list by a distance -- 12,058,688 boardings, nearly five times Buffalo -- and the highest demand of any town here at 38,187 rides a day. Read its row with the model's own limit in mind: the research calls Nashville a multi-district metro, and the simulation has no multi-district geometry for a pack, so it is run as a single service area. Its research is also the thinnest on the list, holding 6 of 13 citations, which is why it carries more estimates than any other town.77 numbers behind this row: 22 published, 32 worked out from published figures, 23 our own estimates.13 citations fetched and checked: 6 held, 7 failed and were demoted to an estimate.2 gaps the research could not fill.Researched 13 September 2026. | |||||
| Chicago, ILsteady | 2,731,585 | 3.3 months | $9,577 | $4,466 | 0 of 12 |
Original The large band of the first batch, to put a 2.7 million metro against a town of 1,900 on the same ladder.81 numbers behind this row: 29 published, 34 worked out from published figures, 18 our own estimates.17 citations fetched and checked: 11 held, 6 failed and were demoted to an estimate.No gaps recorded.Researched 9 September 2026. | |||||
A citation fails when the page we fetched did not say what the research claimed. The figure is kept, demoted to an estimate with the failure written into its math, and counted here rather than hidden. A gap is a figure the research looked for and could not find or could not build, written into the pack rather than papered over. These are researched estimates, not measurements — nobody is running a privately owned Cybercab anywhere today. How a town gets researched, and where we check ourselves against a published figure: the methodology.
Your town is not on this list?
Ask for it. Researching a town means pricing its fares, its power, its permits and its demand, and citing every number — real work, so we build the ones people ask for. 21 towns so far.
Ask for your town →You give an email. We send one link to check it is real, and the only other time we use it is to tell you when your town is built. Not a newsletter; never shown on the site, never sold.
The number that would move every row
And the model's value sits below every published measurement of it
How much of a working hour a car actually spends carrying somebody decides everything on this page, and it swings a season by roughly −44% to +27%. The model ships 0.45. Every published measurement we could find is higher, ranging 0.62 to 0.872.
| Where it was measured | Share of on-app time with a rider, or driving to one |
|---|---|
| Twin Cities metro (seven counties), MN | 0.872 |
| Greater Minnesota -- every county outside the seven-county Twin Cities metro | 0.799 |
| New York City | 0.643 |
| Seattle | 0.62 |
| six US metropolitan regions | — |
So this model is probably pessimistic, and every row above may be understated. We have not changed it, because changing it would move every number on this page and we would rather show you the gap than quietly close it.
Why the order is not the order you would guess
Nothing here tracks population. Wichita, at 400,987 people, does worse than Dover at 1,918. Roughly two‑thirds of the difference between towns is costs and weather — electricity, insurance, how many days it rains — and those are the best‑sourced numbers we hold, straight from federal files. The demand research, which is the expensive and least certain part, largely cancels out: bigger markets have proportionally more cars competing in them.
Model version 1.5.0, captured 3.3 months; research read from the live city index on 14 September 2026. A result is only comparable to another result from the same model version. Change how competition or empty miles work and every number here moves without a single research figure changing.