Back to Newsroom

Op-Ed: Pony AI should separate ride growth from cash progress

Pony AI needs a clearer bridge from passenger fares and vehicle deliveries to cash. Here is the disclosure that would make its expansion easier to judge.

Op-Ed by Nicholas Thomas. The views in this article are the author’s analysis and opinion.

Pony AI reported 165.4% growth in robotaxi revenue in its November 18 results. My view: the next useful disclosure is a bridge from passenger fares and vehicle deliveries to cash collected. Expansion becomes easier to judge when readers can distinguish selling more vehicles from earning more on vehicles already deployed.

Two kinds of growth need two explanations

The robotaxi bucket includes fares and joint-deployment vehicle deliveries. Pony says fare revenue rose nearly 400%, but gives no absolute fare/delivery split. Its company-wide service/product split does not isolate those components within robotaxis.

A delivery and a passenger trip can both generate revenue, but they answer different questions. Deliveries can show that partners want to expand. Repeat fares can help show whether an installed fleet is being used. Neither, by itself, tells us how much cash remains after supporting that activity.

I would like to see fare revenue, delivery revenue and cash collections presented together, with the scope of each measure spelled out. That would let readers follow a deployment from the initial sale into ongoing passenger demand, without treating two ways of classifying revenue as interchangeable.

The cash improvement came from lower capital spending

US$ millions; outflowsQ3 2025Q3 2026
Operating cash use56.79560.532
Capital expenditure22.77817.707
Free cash outflow, non-GAAP79.57378.239

Lower capital expenditure saved $5.071 million; additional operating cash use absorbed $3.737 million. The net improvement was $1.334 million. That arithmetic establishes a smaller combined outflow; it does not establish better operating cash conversion.

Pony attributes the greater operating cash use mainly to working capital and lower capital spending to investment pacing. The release does not provide a complete quarterly working-capital bridge.

Working capital is the cash tied up between paying for business activity and collecting from customers. Timing matters. A growing operation may need more of it before the associated receipts arrive. The useful question is whether that gap closes as deployments mature, rather than whether one quarter’s total looks better.

The strongest case for patience

Joint deployment could allow partners to fund more of the fleet. If that happens, lower capital spending could represent a durable improvement in the business model. Equally, cash invested ahead of expansion can be productive. A quarter of higher operating cash use does not settle either question.

That is why I would avoid treating the spending reduction as proof of deferred investment—or proof that the model has already become less capital intensive. Both interpretations need evidence about who funds vehicles, when customers pay and what ongoing costs Pony retains.

What would change my view

A consistent breakdown showing stronger fares from mature deployments, timely collections and lower capital needs alongside improving operating cash flow would make the expansion story more persuasive. It would also narrow the need for this disclosure request.

My test is straightforward: can readers follow the same business activity from revenue to cash without filling in the missing steps themselves? That is the standard I would apply to the next set of results.

Source and calculation notes: Pony AI’s unaudited release, quarter ended September 30, 2026; checked November 18 at 7:09 a.m. EST. Table amounts convert reported US$ thousands to millions. Calculations use the financial tables, not rounded headline figures. This analysis covers the release, not the earnings call.

More from the Newsroom