Foxconn can make 1,000 AI racks a week, to increase next year, chairman says

Foxconn said it can now build about 1,000 AI server racks per week and expects production capacity to rise next year, underscoring its role as Nvidia’s primary maker of AI racks and reflecting strong demand from the data‑centre boom; the company told partners at its tech day that AI will be a major growth driver in 2026 and plans to invest $2–3 billion a year in AI. Management also said EV volumes have reached a level where automakers can outsource more production to Foxconn and it will showcase vehicles for the Japanese market, while the firm is pursuing quantum computing initiatives. A Foxconn unit, Visionbay.ai, plans to deploy a 27‑megawatt data centre using Nvidia’s GB300 GPUs in H1 2026 — billed as the largest advanced GPU cluster in Taiwan and Asia’s first GB300 AI data centre — positioning Foxconn to capture further cloud and AI infrastructure spending and deepen its Nvidia partnership.

Analysis

Foxconn said it can manufacture about 1,000 AI server racks per week and expects that capacity to rise next year, establishing itself as Nvidia's primary maker of AI racks and a direct beneficiary of the current data‑centre expansion. Management projected AI as a major growth driver for 2026 and committed to annual AI investment of $2–3 billion, signaling sustained capital allocation to support rack production and related services.

A Foxconn unit, Visionbay.ai, plans to deploy a 27‑megawatt data centre using Nvidia's GB300 GPUs in H1 2026, described as the largest advanced GPU cluster in Taiwan and Asia's first GB300 AI data centre; that project both deepens the Foxconn‑Nvidia relationship and creates a visible revenue/timing milestone to track. The firm is also pushing into EV manufacturing — saying volumes have reached a level to attract outsourced production — and is exploring quantum computing, reflecting a strategic diversification beyond handset assembly.

The announcement is moderately positive for Foxconn's medium‑term revenue mix but heightens execution and capital‑intensity risk: realizing $2–3 billion annual AI spend, completing the Visionbay GB300 deployment on schedule, and converting rack capacity into confirmed customer orders will determine margin recovery. Investors should focus on shipment cadence, contract wins with cloud customers and evolving margin disclosure as near‑term performance indicators.

Key Decisions for Investors

• Consider selectively increasing exposure to Foxconn (2317.TW) given the 1,000 racks/week capacity and $2–3 billion annual AI investment, but size positions modestly until shipment and order evidence confirms demand
• Monitor progress on the 27‑MW Visionbay GB300 deployment (H1 2026) and publicized customer contracts as key catalysts and short‑term proof points of revenue conversion
• Watch Nvidia GPU demand and partnership developments as a correlated lead indicator for Foxconn's AI revenue, and maintain readiness to hedge or pare exposure if timelines or margins slip