Foxconn's AI Expansion Plans and OpenAI Partnership
Foxconn is strategically expanding its US manufacturing presence with a multibillion-dollar investment and a key partnership with OpenAI to innovate AI data center architecture, supporting large-scale infrastructure projects and reducing reliance on traditional consumer electronics assembly.
Hon Hai Precision Industry Co., commonly known as Foxconn, announced plans to invest between $1 billion and $5 billion initially to expand its manufacturing operations in the US, targeting the growing demands of AI companies like Nvidia and OpenAI.
The announcement came on November 21, 2025, alongside a new partnership with OpenAI focused on co-designing and producing AI server racks and related equipment, such as cabling and power systems, to overcome early-stage challenges in AI data centers. Foxconn Chairman Young Liu highlighted in an interview that the collaboration draws on OpenAI's experience with AI compute issues, stating: “For OpenAI, they’re the largest user of AI compute and they have a lot of experience with what can go wrong... So because of that experience, they learned that there must be a new way or new architecture for the AI data center.”
The company aims to assemble up to 2,000 server racks per week in the US by 2026, though no firm purchase commitments were specified. This builds on prior interactions, including hosting OpenAI CEO Sam Altman in Taiwan, and ties into broader initiatives like the Stargate project with OpenAI and Oracle, involving a $500 billion US data center investment plan. The expansion is part of Foxconn's strategy to diversify beyond iPhone assembly for Apple, strengthen its position in AI hardware, and hedge against potential tariffs amid US political changes.
Foxconn Presence in the USA
Foxconn has maintained a presence in the United States since at least 2017, with operational facilities and ongoing investments in multiple states, though some projects have faced challenges and scale-backs. Key locations include:
- Wisconsin: Foxconn established a science and technology park in Mount Pleasant, breaking ground in 2018 after a 2017 agreement. The company has invested over $1 billion in the 3,000-acre campus, supported by more than $1.2 billion in public infrastructure spending. However, the project fell short of initial promises, creating fewer jobs than anticipated and shifting focus from large-scale LCD manufacturing to smaller-scale operations.
- Ohio: Foxconn operates a facility in Warren, which serves as part of its North American manufacturing footprint.
- Texas: The company has two sites in Houston (8303 and 8801), with plans announced in August 2025 to invest $450 million in AI-related facilities, potentially creating about 600 jobs. This expansion targets production for clients like Apple and supports AI hardware needs.
Additionally, Foxconn's US headquarters is located in the country, and recent developments include a November 2025 partnership with OpenAI to co-design and manufacture AI server components domestically, alongside broader plans for $1 billion in US investments over the next decade to strengthen supply chains amid tariffs and geopolitical shifts.
OpenAI's Project Tigris
OpenAI is developing custom AI chips under the codename Project Tigris, with plans for mass production starting in 2026, as part of efforts to reduce dependency on dominant suppliers like Nvidia and, to a lesser extent, Google's TPUs. The initiative, led by CEO Sam Altman, involves partnerships with Broadcom for chip design and TSMC for fabrication, focusing on specialized semiconductors for AI training and inference to address compute demands and costs. Reports indicate fundraising efforts in the tens of billions to support this, positioning the chips as a rival to Nvidia's GPUs, while OpenAI has also begun using Google's TPUs for some operations as a short-term diversification strategy.
Key Points
- Leadership and Origins: Led by OpenAI CEO Sam Altman, the project emerged around 2023 as part of efforts to build a chipmaking venture, with Altman fundraising in regions like the Middle East.
- Funding Ambitions: Altman has sought investments ranging from billions to trillions of dollars to overhaul global semiconductor production, including partnerships for fabrication.
- Technical Aspects: The chips are planned to use advanced processes like TSMC's 1.6nm A16 node, positioning them as lower-cost competitors to Nvidia's GPUs and potentially similar to Google's TPUs.
- Strategic Goals: The project addresses AI compute shortages and costs, with mass production targeted for 2026, amid OpenAI's broader push for sovereign AI infrastructure.