Foxconn's chairman has set a confident target of over 10% revenue growth for 2026.
This optimism is overwhelmingly powered by a major new product cycle: NVIDIA's next-generation AI server platform, codenamed 'Vera Rubin' or 'VR'. This isn't about virtual reality headsets; it's a fundamental shift to highly integrated, rack-scale systems where NVIDIA designs the entire unit, from chips and networking to cooling. This new approach heavily favors large, vertically integrated manufacturers like Foxconn, which possess the expertise to build and deliver these complex, all-in-one systems at massive scale.
The forecast isn't just speculative. First, the product roadmap is becoming very real. NVIDIA officially unveiled the Vera Rubin platform at CES in January and recently confirmed that initial samples have already shipped to customers. With mass production slated for the second half of 2026, Foxconn has a concrete timeline to ramp up its manufacturing operations, turning future potential into a tangible production plan.
Second, the broader market provides a powerful tailwind. Market research firm TrendForce projects that overall AI server shipments will surge by more than 28% in 2026. While Foxconn's smartphone business also shows near-term stability, there are growing warnings of a potential industry downturn. This dynamic places the burden of growth squarely on the AI server division, making it the critical engine for hitting the company's ambitious target.
However, a significant hurdle exists that has less to do with demand and more with deployment. Data centers are incredibly power-hungry, and recent reports indicate that 30-50% of new projects face delays due to grid limitations. This means that even if Foxconn ships a record number of AI racks, its customers may struggle to actually install and power them on, creating a lag between shipment and revenue recognition. Geopolitical factors, such as U.S. export controls on China, also influence the landscape by redirecting infrastructure spending to other regions where Foxconn is expanding.
- ODM (Original Design Manufacturer): A company that designs and manufactures a product that is eventually rebranded by another firm for sale.
- Rack-scale platform: An integrated computing system where the entire server rack—including servers, storage, networking, and cooling—is designed and sold as a single unit.
- HBM (High Bandwidth Memory): A high-performance type of computer memory used alongside GPUs to accelerate AI workloads.