![]() 13 A prime example is Nvidia, which supplies AI chips to data centers. These efforts create opportunities for chip vendors exposed to AI processing and data intensive hardware. Growing downstream applications of generative AI has demand for AI-first infrastructure rising, whether in the original designs of new buildouts or as part of infrastructure optimization. 11 Meta is increasing the footprint of its custom in-house built AI accelerator chip in new data centers, building new dedicated AI training clusters, and is investing to deploy supercomputers designed to support augmented reality, content understanding systems, and real-time translation technology. Meta, for example, is reimagining its data center architecture to cater to AI workloads and expects investments to accelerate in the coming quarters. 9,10 Also, the trend of optimizing data centers for AI is set to continue. 8 For hyperscalers, Microsoft leads the pack with 24 projects underway in eight countries, Google has 15 projects across seven countries, and Amazon has 15 projects in the works in nine countries. 7Ĭapacity expansion is a major target, and in 2023, hyperscalers and colocation service providers are expected to add a staggering 26 million square feet of data center space. 6 Currently, hyperscalers own and operate close to 900 data centers, accounting for 37% of worldwide data center capacity, according to one estimate. 5 Spending for the group is down 2.1% relative to last year but remains up over 25% from 2021 levels. In the last year, capital expenditures from Microsoft, Amazon, Google, Meta Platforms, IBM, and Oracle collectively topped $154 billion. Hyperscalers Ramp Up Data Center Investments The potential to command a premium for emerging services can boost earnings for data center companies in the near to midterm, in our view.4 Growing demand for computational power for AI, automation, and the internet of things (IoT) is projected to amplify this pricing trend. primary data center markets for colocation services. Constrained data center capacity contributed to a 20–30% YoY asking price increase in the eight U.S.Data Center construction in North America increased 25% year-over-year (YoY) in H1 2023, propelled by AI and capital expenditures from leading tech giants.In our view, these dynamics can bolster revenue and profits for the entire data center ecosystem, including large cloud giants, colocation service providers, and suppliers of essential components such as chips, storage, and networking equipment. ![]() 2 Data center capacity and pricing for space are at a premium, and we believe that demand is only likely to increase as cloud services adoption and generative AI-based products and services continue to gain traction. 1 With data center vacancy rates at unprecedented lows, the meaningful uptick in new construction projects in 2023 comes at a critical point. Data center investment in new capacity construction and absorption is surging in North America.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |