Foxconn is preparing to deepen its relationship with OpenAI as the world's largest contract electronics manufacturer looks to expand its role in the rapidly growing AI infrastructure market.
Chairman Young Liu confirmed that a fresh collaboration was in the works after meeting OpenAI CEO Sam Altman on Friday at the company's annual technology event in Taipei. Liu said further details would be announced once the terms were finalised, without disclosing the scope of the proposed arrangement.
The development builds on an existing partnership between the companies, announced in November 2025, to design and develop next-generation AI data-centre hardware and strengthen US manufacturing capabilities.
Existing partnership provides foundation
Under their earlier agreement, OpenAI agreed to share its infrastructure requirements to help Foxconn design multiple generations of AI data-centre racks. The collaboration also covered the manufacturing of critical equipment, including networking, cabling, cooling and power systems, at Foxconn's US facilities.
Importantly, the initial agreement did not include purchase commitments or financial obligations. OpenAI was given early access to evaluate the equipment and an option to purchase the systems.
The Nvidia supplier has since expanded its AI server manufacturing operations in the US. The company operates six factory buildings in Houston, Texas, where it produces AI servers for Nvidia, and Liu said further expansion would be needed. Its cloud and networking business, which includes AI servers, accounted for 51% of Foxconn's second-quarter revenue, crossing the 50% mark for the first time.
Foxconn pushes back against AI bubble concerns
The potential expansion of its OpenAI relationship comes as investors increasingly question whether the enormous investments in AI infrastructure will generate sufficient returns. Even top industry executives have acknowledged market risk as spending continues to increase amid rising borrowing costs.
Nvidia Chief Executive Jensen Huang recently said demand for neocloud companies is "off the charts," but a shortage of working capital is limiting growth. Google CEO Sundar Pichai had previously cautioned that "elements of irrationality" exist within current investment valuations.
Technology stocks have faced bouts of selling as concerns grow over rising borrowing costs, infrastructure expenses and the time required to monetise AI investments. Liu dismissed concerns that the sector is experiencing a bubble, arguing that demand for computing infrastructure remains strong. He also expects Foxconn to maintain double-digit revenue growth next year, supported by its expanding AI server business.
Humanoid robots enter Texas facilities
Separately, Foxconn is stepping up its automation efforts, with more than 10 humanoid robots already deployed at its Texas facilities for internal use, according to Reuters. Brand Cheng, chairman of Foxconn subsidiary Foxconn Industrial Internet, told Reuters that the company plans to expand robot deployment beyond Texas.
The subsidiary has introduced two wheeled robot models designed for tasks such as lifting, assembly, and loading and unloading materials. It is also developing AI models for humanoid robots using Nvidia's GR00T foundation model.
The developments highlight Foxconn's broader push to position itself across the AI ecosystem, from manufacturing the hardware powering AI systems to using robotics in its own production facilities.
This article is for informational purposes only and does not constitute financial advice.
