Foxconn AI Server Demand Drives 35% Profit Jump

Foxconn, a renowned device assembler, has experienced a 35% surge in profit, propelled by a surge in AI server demand. The company's AI server business has grown to the point where it now significantly impacts the financial results. Foxconn's net profit for the second quarter rose 35% year-over-year to NT$59.97 billion, or approximately $1.86 billion, while revenue climbed by 41% to NT$2.526 trillion.

Cloud and networking products, which now account for more than half of the revenue, saw substantial growth. Foxconn anticipates a high double-digit percentage increase in AI server rack shipments for the third quarter of the current year, and more than double the shipments for the entire year. This marks a shift in focus from being Apple's largest device assembler to being a substantial contributor to the AI hardware sector.

Analysts reported that Foxconn's profit exceeded the NT$58.22 billion expected by analysts surveyed by FactSet. Chairman Young Liu has highlighted that AI infrastructure demand is coming from model developers, cloud providers, governments, and enterprises, and expects the infrastructure build-out to continue for at least the next three to five years.

TrendForce projects that Google, Amazon, Meta, Microsoft, Oracle, ByteDance, Tencent, Alibaba, and Baidu will invest over $886.7 billion collectively in AI servers by 2026, which is nearly 90% more than the previous year, with the top five North American hyperscalers accounting for nearly 90% of that spending. Foxconn's earnings align the manufacturing results with cloud infrastructure spending projections, demonstrating that AI hardware demand is already translating into supplier sales.

The upcoming racks will feature a mix of hardware configurations, with Google expected to expand its TPUs, AWS leveraging Nvidia systems combined with its in-house ASICs, and Meta utilizing Nvidia, AMD racks alongside proprietary silicon. This diversification increases the complexity of manufacturing and procurement, shifting some pressure beyond GPUs to advanced chip packaging, high-bandwidth memory, networking, and other rack components.

Power and cooling are also key constraints, as new rack-scale AI systems require distinct facilities compared to conventional servers. While rising production provides stronger evidence of supply reaching manufacturing lines, buyers should still verify the availability of complete rack capacities, memory, packaging, cooling, and power before concluding that supply will be easily secured.