Hyperscale operators maintained their record-setting year of capex spending, which was up 59 percent in the second quarter from the same period last year, according to a new report from Synergy Research Group. In the first half of 2018, the hyperscale operators spent $53 billion on capex, up from $31 billion in the first half of 2017.
Synergy Research analyzes the capex and data center footprint of the world’s largest cloud and internet service firms at the end of each quarter. This includes the largest operators of infrastructure-as-a-service (IaaS), platform-as-a-service (PaaS), software-as-a-service (SaaS), search, social networking, and e-commerce.
According to John Dinsdale, chief analyst and managing director at Synergy Research, its previous analysis data of the hyperscale market included 24 companies. However, the research firm reduced this to 20 companies because LinkedIn has been integrated into Microsoft and three other companies weren’t spending enough on capex to warrant inclusion, he said.
Hyperscale operators refer to those that operate “huge data center networks that are used to support a multitude of external clients,” Dinsdale said. “Hyperscale operators also plan and operate data centers very differently from the norm, having different use cases and using different operational metrics.”
The research found that the top five spenders have remained the same for the last 10 quarters. These big spenders are Google, Microsoft, Amazon, Apple, and Facebook — and combined they account for more than 70 percent of the total hyperscale capex. According to Synergy Research, three out of these five experienced their highest capex levels in the most recent quarter.
The large spending by these operators, particularly in cloud computing, has become a competition killer said Dinsdale. This means that small operators wanting to join the market have a large — potentially impossible — uphill battle in order to compete.
Outside of these five operators, other top spenders were Alibaba, Baidu, IBM, JD.com, NTT, Oracle, SAP, and Tencent.
Hyperscale capex spending primarily includes spending on data centers, which Dinsdale said accounts for a “very large chunk of the capex.” This includes the land, the buildings, the servers, and the gear associated with data centers. Outside of data centers, hyperscale capex numbers also include other property, equipment, construction, and development costs.
Synergy Research found that hyperscale capex has continued to grow and shows no sign of slowing anytime soon. Combined, the hyperscale operators analyzed spent more in capex during the first half of 2018 than they did for all of 2015. Dinsdale noted that the most recent quarter would have been the highest quarter yet if it wasn’t for one purchase in the first quarter of 2018: Google’s $2.4 billion purchase of Chelsea Market buildings in February.
Dinsdale said that the immense growth is driven primarily by demand and changes to technology and business models — all of which are long-term drivers. As public cloud services, social networking, and other burgeoning technologies grow to address these changes, “All of these have to run on huge data center networks,” he said. “But it goes further than hyperscale networks being built to satisfy demand for existing services: As the hyperscale networks get bigger and more sophisticated they can enable new services and features which in turn stimulate whole new areas of demand.”
