Foxconn’s board has approved a $1.37 billion investment in an AI compute cluster and supercomputing center, scheduled between December 2025 and December 2026.
The development comes as a transformative shift from contract manufacturing into owning critical AI infrastructure.
This move marks the latest phase in the Taiwanese contract electronics giant’s pivoting away from assembly work (iPhones, consumer devices) toward building the computational backbone of the AI industrial revolution.
The investment raises a central question: can Foxconn translate manufacturing scale into an AI services provider, and what does its deepening AI footprint mean for the geopolitical contest between US cloud giants, China’s AI ambitions, and Taiwan’s strategic clout?
The business case: Why Foxconn is building compute
Foxconn isn’t entering untested territory.
The company already holds roughly 40% of the global AI server manufacturing market share in 2025, with AI servers now generating more revenue than smartphones for the first time in company history.
The $1.37 billion cluster investment extends this advantage: moving from producing AI hardware for others toward directly offering AI computing services as an NVIDIA Cloud Partner.
The strategic calculus is clear. Cloud giants like AWS, Google, and Microsoft face a GPU bottleneck, with hyperscaler data-center capital expenditure reaching $134 billion in Q1 2025 alone, a 53% year-over-year spike.
By building dedicated capacity, Foxconn captures higher margins, diversifies revenue beyond manufacturing, and locks in long-term customer relationships.
The company’s cloud and networking division already showed the strongest growth in Q3 2025, driving record quarterly revenue of $67.8 billion.
Additionally, the investment addresses government policy: Taiwan’s “Ten Major AI Infrastructure Projects” initiative aims to create $510 billion in economic value by 2040.
Foxconn’s local investment aligns with Taipei’s push to establish Taiwan as an AI hub independent of US or Chinese infrastructure.
Geopolitical stakes: Triangulating US, China, and Taiwan
Foxconn’s move reshapes three overlapping power games.
First, it strengthens Taiwan’s AI shield.
Taiwan already controls 90% of global AI server manufacturing capacity; adding owned compute infrastructure deepens the island’s indispensability to every major AI player, US hyperscalers, Chinese competitors, and startups globally.
Taiwan cannot be easily decoupled from the AI supply chain without crippling progress worldwide.
Second, it complicates the US dependency calculus.
The US has long relied on Taiwan for semiconductors and hardware, yet simultaneously seeks to onshore capacity (like TSMC’s Arizona fab) to reduce exposure.
Foxconn’s $375 million sale of its Ohio plant to SoftBank in August, now repurposed for Stargate data-center manufacturing with OpenAI and Oracle, shows Taiwan firms hedging geopolitical risk by anchoring critical AI infrastructure in America.
However, the core intellectual property and manufacturing prowess remain Taiwanese.
Third, it signals competition with China’s AI ambitions. China cannot easily replicate Taiwan’s design-manufacturing integration or NVIDIA partnerships due to US export controls.
Foxconn’s infrastructure play underscores Beijing’s structural disadvantage in the compute arms race.
The post How could Foxconn’s $1.37B AI bet shift AI balance between US, China and Taiwan? appeared first on Invezz