Foxconn chairman Young Liu used the company's latest earnings call to deliver a message: artificial intelligence is no longer just a growth driver—it is now the most decisive force shaping Foxconn's future.
Liu said the company is preparing a deep partnership with OpenAI to support the world's fastest-growing demand for compute, and he promised that key details will be unveiled at Hon Hai Tech Day (HHTD) on November 21–22, 2025.
Liu said Foxconn "will absolutely work with OpenAI, the leading player in the AI industry," citing CEO Sam Altman's ambition to build one gigawatt of compute capacity every week.
At today's AI pricing curves, Liu said, that goal implies a recurring weekly market worth roughly US$50 billion—an unprecedented scale that "is only just beginning."
A relationship years in the making
The partnership is not emerging in a vacuum. Altman made a low-profile trip to Taiwan in late September and early October 2024, meeting privately with Liu to discuss chip supply, compute infrastructure, and server-deployment roadmaps.
Liu later confirmed the meeting and said he intends to hold further discussions with Nvidia CEO Jensen Huang as OpenAI's hardware requirements accelerate.
Behind the scenes, Foxconn has already been closely linked to OpenAI's most ambitious infrastructure push: the Stargate Project, a multi-hundred-billion-dollar effort to build the United States' next generation of hyperscale AI data-center campuses.
Oracle is the principal cloud provider for the project, and Foxconn is Oracle's largest AI-server supplier—effectively placing the Taiwanese manufacturer at the hardware heart of Stargate.
SoftBank's SB Energy is expected to provide portions of the power infrastructure, while Foxconn's former EV assembly site in Lordstown, Ohio, is being converted into a production base for data-center equipment. The repurposed facility signals Foxconn's direct entry into the US hardware supply chain for the project.
Rising shipments and entrenched market share
OpenAI's breakneck expansion has already lifted Foxconn's server shipments. Although competitors such as Quanta have increased supply, industry data shows Foxconn still maintains the largest share of AI-server output, benefiting from long-cycle demand tied to the Stargate buildout.
Liu said the company's advantage is structural: Foxconn is one of the world's leading hardware developers for large-scale compute systems and a critical partner to customers racing to deploy GPUs, accelerators, networking gear, and power-delivery solutions at unprecedented speed.
Preparing for the age of "physical AI"
In Liu's view, the compute race is only at its beginning as AI evolves from narrow intelligence to AGI and eventually ASI. He said the shift from conversational models to autonomous AI agents—and eventually "physical AI"—will cause hardware demand to grow "explosively."
Given this trajectory, Liu framed Foxconn as indispensable: "We will fully and comprehensively support our partners in achieving their compute-deployment goals," he said.
Before closing the call, he urged investors to watch HHTD closely: "We will have OpenAI-related announcements at HHTD. Please stay tuned."
Article edited by Jerry Chen
