Shelter-in-place orders in 2020 gave rise to previously unforeseen demand for video conferencing, video streaming, gaming, and e-commerce. During this challenging period, it was data centers that connected us to each other and maintained productivity.

 

Data center equipment spending recorded particularly strong growth in 1H20, followed in 3Q20 by an absorption phase. Absorption phase is the period when the utilization across data center equipment continues to increase until it reaches the expected max threshold before requiring additional capacity. Based on our high frequency data, construction of capacities grew in high teens during 3Q20 compared to the previous quarter. Our algorithms predict spending bounced back 5-10% QoQ in 4Q20, resulting in an overall solid growth for 2020 of ~25-30%. 2020 almost managed to reach the record absorption of 2018 during its second quarter but fell short by less than 3%.

 

Big Data Federation now predicts data center equipment (compute, networking, and storage equipment) spending1 in 2021 to slightly grow compared to 2020. Overall, 2021 will be another positive year for the world of the data center, cloud providers, and their suppliers. Approximately 65% of the total data center equipment capital expenditure will come from three hyperscale cloud service providers (Amazon, Microsoft, Alphabet/Google), ~25% from Facebook and Apple collectively, and ~10% from China’s hyperscalers (Alibaba, Baidu, Tencent). For 1H21, our algorithms predict single digit growth for hyperscale data center equipment spend (with the second quarter experiencing a slight slowdown), and an absorption phase in 2H21. Growth is especially expected in Northern Virginia which comprises approximately 50% of the U.S. data center market due to large hyperscaler presence in that market.

 

 

Big Data Federation employs a comprehensive dataset based on US and international hyperscale companies’ supply chains, quarterly/annual filings, and investor presentations to accurately assess data center equipment expenditures. A real-time currency conversion analysis is conducted for international companies. Continuously-learning proprietary algorithms that select and interconnect data generate real-time predictions.