Pony AI Inc. American Depositary Shares

Q2 2026 total revenue was USD 36.2 million, up 69% from USD 21.5 million a year ago. Robotaxi revenue reached USD 12.1 million, up 691%, and robotruck revenue was USD 13.3 million, up 40%; intelligent solutions revenue was USD 10.8 million, up 4%. Gross profit was USD 6.4 million, with gross margin of 17.5% versus 16.1% last year. GAAP operating loss was USD 65.7 million, non-GAAP operating loss was USD 56.7 million, and net loss was USD 45.4 million, narrower than the prior-year quarter. Cash, cash equivalents, short-term investments, restricted cash and long-term wealth management instruments were USD 1.39 billion as of June 30, 2026, versus USD 1.44 billion as of March 31, 2026. Management said the robotaxi fleet expanded to 2,000 vehicles, with 3,500 vehicles targeted by year-end, and that the company expects to surpass 20 cities by year-end. It also said robotaxi revenue is on track to exceed 3.5x last year’s level, while robotruck growth is expected to persist and strengthen in the second half of the year.

James Peng framed the quarter as evidence that Pony’s “dual engine” strategy is working, citing broad expansion in China and overseas, plus faster commercialization through the joint deployment model. He emphasized Tier 1 city leadership, safer and more proven large-scale driverless operations, and international validation through Uber, Bolt, Stellantis, Verne, and ComfortDelGro. His tone was upbeat and confident, especially on the company’s ability to scale while keeping capital intensity low.

Leo Wang highlighted the hard numbers behind the quarter: revenue of USD 36.2 million, gross profit of USD 6.4 million, gross margin of 17.5%, operating loss of USD 65.7 million, and net loss of USD 45.4 million. He pointed to operating leverage, noting total GAAP operating expenses of USD 72.1 million and non-GAAP operating expenses of USD 63 million, both growing far slower than revenue. He also said cash and equivalents plus investments totaled USD 1.39 billion, net cash used in operating activities was USD 44 million, and capital expenditures were USD 32.2 million this quarter and USD 44.3 million for the first half, with spending tied to fleet, autonomous kits, and data centers. He stressed that the joint deployment model helps keep capital discipline by allowing partners to fund fleet expansion.

Analysts focused on why Uber chose Pony for Europe, how the joint deployment model works, whether PonyWorld 2.0 creates a durable moat, and how domestic expansion will proceed in the second half. Management said Uber values Pony’s reliability at scale and attractive cost structure, and that Pony’s real-world performance in China and Zagreb proved its capability in complex markets. On the joint deployment model, management described an asset-light structure where Pony supplies the vehicle and AI driver, while mobility platforms and local operators contribute demand and fleet management, creating sharing-based or licensing revenue. On PonyWorld 2.0, management argued that its AI-driven loop engineering reduces manual R&D work, speeds city launches, and should not be easily replicated by an open-source world model.