Foxconn just closed 2025 with its best quarter on record, and the numbers make one thing clear, the AI bubble is nowhere near bursting. In fact, the company posted $83 billion in Q4 revenue, a 22% year-on-year surge that topped analyst expectations of $77 billion.
On top of that, full-year revenue hit a record of $8.09 trillion, with December alone setting the highest monthly revenue in company history.
Taken together, these are not coincidental figures from a company sitting at the physical center of the AI economy.
How AI Servers Took Over Foxconn’s Revenue
For decades, Foxconn built its identity around the iPhone but that identity has undergone major changes.
In Q2 2025, the Cloud and Networking division claimed 41% of total revenue while consumer electronics fell to 35%, marking the first time AI infrastructure outsold consumer devices in company history.
More importantly, AI server revenue grew more than 60% year-on-year in Q2 and by Q3 the company projected AI server revenue growth of 170% with rack shipments growing threefold quarter-on-quarter. All this goes to show that the AI momentum is not slowing down anytime soon.
The AI Bubble Runs on Steel and Concrete
Unlike mere speculation, Foxconn ties its growth to physical commitments.
Specifically, as the world’s largest contract electronics manufacturer, Foxconn builds the servers that hold chips in data centers which means every hyperscaler capex round translates directly into production orders.
Beyond that, Foxconn is backing a $2 to $3 billion annual capital expenditure plan focused entirely on AI, with new facilities in Texas and Wisconsin already underway. That capital is already deployed and already generating returns.
Why the AI Bubble Keeps Attracting Blank Checks
The demand sustaining all of this starts at the very top.
For example, Foxconn Industrial Internet’s cloud computing segment grew 88.7% year-on-year in 2025, with AI server revenue for cloud service providers increasing more than threefold and revenue from high-speed switches above 800G surging thirteenfold.
To put it simply, that kind of growth does not happen without massive committed spending from hyperscalers. As long as Amazon, Microsoft, Google, and Meta keep expanding data center capacity, Foxconn’s order books stay full.
The Risks That Could Finally Pop the AI Bubble
Nevertheless, real risks exist. Geopolitical friction, US tariff policies, and chip supply constraints all threaten momentum.
Furthermore, Foxconn’s fortunes tie closely to Nvidia’s dominance, meaning any architecture shift or export restriction hits Foxconn directly. The company itself acknowledged that its record Q4 creates a tougher comparison base heading into Q1 2026.
Even so, all of this points to one clear direction, the AI bubble is not losing momentum anytime soon.
