Foxconn, the world's biggest contract electronics manufacturer, said on Friday that a new collaboration with OpenAI is in the works, even as it pushed back against growing talk of an AI bubble.
Chairman Young Liu told reporters at the company's annual technology showcase in Taipei that he had met OpenAI Chief Executive Sam Altman earlier in the day, and that there would be an announcement once details were confirmed. Liu did not elaborate on what the collaboration would cover, and no terms were disclosed.
Any deal would build on last year's agreement between the two companies to work together on designing and engineering data centre racks, components and other AI hardware. That earlier partnership already put Foxconn, formally known as Hon Hai Precision Industry, closer to the centre of OpenAI's infrastructure plans.
The timing matters because Foxconn's business has tilted sharply toward AI. It is Nvidia's biggest AI server maker, and the company said in August that its cloud and networking products segment, which includes AI servers, accounted for 51 percent of second-quarter revenue, exceeding 50 percent for the first time. A business once defined by consumer electronics assembly is now carried, in revenue terms, by AI infrastructure.
Liu also addressed the market mood directly. AI-sector stocks have been hit hard by concerns that relentless heavy investment is not sustainable amid climbing borrowing costs. Liu said the industry was not in a bubble, and that Foxconn expects to maintain double-digit growth next year. Coming from the manufacturer that actually builds much of the hardware, that is a notable counterpoint to the spending anxiety that has weighed on technology shares this month.
The company's United States footprint featured in the remarks as well. Foxconn has six factory buildings in Houston, Texas, where it manufactures AI servers for Nvidia, and Liu said it sees a need for further expansion there.
Separately, Foxconn is putting robots to work in those Texas facilities. Brand Cheng, chairman of the unit Foxconn Industrial Internet, told Reuters that more than 10 humanoid robots had been deployed at its Texas facilities for internal use, and that the company plans to extend deployment beyond Texas. Presentation materials seen by Reuters showed the unit introduced two wheeled models in October: one that lifts things vertically, and one that can bend at the waist for assembly as well as loading and unloading tasks. The unit is developing AI models for humanoid robots based on Nvidia's GR00T foundation model.
For now, the OpenAI collaboration remains a headline without details. What is already clear is the direction: the world's largest contract manufacturer is binding itself more tightly to the leading AI labs and chip designers, expanding the American factories that build AI servers, and beginning to staff those factories with robots of its own. Whether the next announcement is about data centre hardware, manufacturing, or something else, Foxconn is positioning itself as infrastructure for the AI era rather than a bystander to it.