XPeng plans to raise monthly production capacity for its IRON humanoid to more than 1,000 units by the end of 2026, the company said on July 15, ahead of a broader rollout the following year. Shares in the U.S.-listed automaker (NYSE: XPEV) rose about 4% on the disclosure, first reported by the Wall Street Journal.

The number is the point. Plenty of humanoid makers have shown a robot walking; far fewer have committed to a monthly build rate. A four-figure cadence would put XPeng among the small group of companies — most of them Chinese — talking about output in the thousands rather than the dozens, and it reframes IRON as a manufacturing program rather than a demo.

The scarce commodity in humanoids right now is not a viral clip. It is a production line that can repeat. — EW analysis

The store is the first customer

XPeng's deployment plan is unusually concrete for the category. The company intends to place IRON in its own Chinese retail stores as a shopping assistant from the first quarter of 2027, then extend it to international showrooms later that year. That sequencing matters: a carmaker's own sales floor is a controlled, well-lit, low-stakes environment where a robot can greet customers and answer questions without the reliability bar of a factory shift or a private home. It is a captive first market that also doubles as a marketing surface for the cars.

It is the same logic Tesla applies to Optimus — use in-house demand to absorb early, imperfect units — and XPeng is explicit about the parallel. The bet across both companies is that automotive supply chains, tuned for batteries, actuators, sensors and high-volume assembly, transfer to humanoids more cheaply than a robotics startup can build from scratch. Whether that advantage holds depends on the part of the problem cars don't solve: the manipulation software that lets a machine handle an unstructured world.

Key Facts

  • Target: 1,000+ IRON humanoids per month by end of 2026 (per WSJ report, July 15)
  • XPeng shares (NYSE: XPEV) rose about 4% on the news
  • In-store shopping-assistant deployment planned for Q1 2027 in China, international later in 2027
  • Strategy mirrors Tesla's Optimus: reuse EV supply chain and factories for physical AI

Why it matters

Targets are not shipments, and a monthly capacity figure is easier to announce than to hit — early humanoid lines have a history of slipping. But the disclosure sharpens a trend that has defined 2026: the humanoid contest is increasingly a manufacturing contest, and automakers are entering it with an unfair head start on parts and plants. The open question for XPeng is not whether it can build a thousand robots a month, but whether it can give a thousand robots a month something useful and reliable to do.

Frequently Asked

How many humanoids does XPeng plan to build?

It aims for more than 1,000 IRON units a month by the end of 2026, according to a Wall Street Journal report dated July 15, 2026. Shares rose about 4% after the disclosure.

Where will IRON be deployed first?

In XPeng's own Chinese retail stores as a shopping assistant from Q1 2027, then international showrooms later that year.

Why is a carmaker building humanoids?

To reuse EV batteries, actuators, sensors and factory lines for physical AI — the same crossover Tesla is pursuing with Optimus.