Quanta Computer Inc. announced a 40% increase in capital expenditure for this year.

The company is allocating NT$20 billion (US$660.4 million) to meet rising demand for AI servers powered by Nvidia Corp’s chips.

These AI servers accounted for about 70% of the company’s revenue last quarter.

Production of new AI servers using Nvidia’s GB200 chips began late last quarter, with demand expected to surpass that of the previous-generation Hopper-series chips.

Quanta expects AI servers to make up over 70% of its revenue this year, up from 60% last quarter.

The company anticipates pressure on gross margins due to the transition to next-generation AI servers. Gross margin improving slightly to 7.92% last quarter but down from 8.48% a year earlier.

🔗 Source: Taipei Times

🧠 Food for thought

1️⃣ AI server market boom drives unprecedented capital investments

Quanta’s 40% capital expenditure increase is part of a massive industry-wide investment trend that extends far beyond a single company.

The global AI server market is projected to grow from $31.2 billion in 2023 to approximately $430 billion by 2033, reflecting a 30.4% compound annual growth rate1.

This helps explain why Quanta’s business model has transformed so dramatically, with AI servers now representing 70% of revenue while notebooks have dropped to just 25%.

McKinsey research indicates this surge is driven by the explosive compute demands of generative AI applications, with total compute requirements potentially reaching 25×10^30 FLOPs by 20302.

Cloud service providers’ continued heavy investment in data center infrastructure further supports this trend, creating a cycle of investment across the AI hardware ecosystem.

2️⃣ Component shortages creating bottlenecks despite massive spending

Despite the substantial capital investments, Quanta explicitly cited chip supply constraints as limiting their AI server output—a pattern reflected across the industry.