In a stunning display of AI ambitions, Microsoft has emerged as the leading customer for NVIDIA’s cutting-edge GB200 AI servers, with demand exceeding all other cloud service providers combined. According to renowned industry analyst Ming-Chi Kuo, Microsoft’s orders for GB200 servers in Q4 2024 are expected to surge 3-4 times compared to previous quarters, marking a significant escalation in its AI infrastructure investment.
This massive demand is driving a rapid acceleration in production and shipments of key components for the GB200 servers, with suppliers beginning mass production in Q4 2024. This expedited timeline reflects Microsoft’s eagerness to deploy these powerful AI servers, giving it a head start in the race for AI supremacy.
Kuo’s analysis reveals that Microsoft’s initial GB200 deployment will focus on data centers located in cooler climates, including Washington state in the US, Quebec City in Canada, and Helsinki in Finland. This strategic approach aims to proactively mitigate potential cooling issues associated with the high-performance servers, ensuring optimal operation from the outset.
Kuo anticipates the shipment of 150,000 to 200,000 Blackwell AI GPU chips in Q4 2024, with a staggering projected growth of 200-250% quarter-over-quarter to 500,000 to 550,000 units in Q1 2025. This explosive growth further underscores the escalating demand for AI computing power and Microsoft’s commitment to building a robust AI infrastructure.
Microsoft’s aggressive pursuit of NVIDIA’s GB200 servers highlights the fierce competition in the AI landscape. By securing a significant portion of the initial production, Microsoft positions itself as a dominant force in the rapidly evolving AI industry. The company’s strategic deployment in cooler climates also demonstrates its proactive approach to managing the challenges associated with high-performance computing.