While some analysts wonder whether the AI boom will fade away in the coming months, others are looking further into the future. According to experts PricewaterhouseCoopersBy 2050, under the basic scenario alone, the global data center construction cost will increase to US$31.6 trillion, and will reach US$50 trillion under the more optimistic scenario.
Image source: NVIDIA
The authors of the forecast say that this will be the largest investment boom in human history, surpassing the development of railway construction, network and electrification. For reference, the United States’ current GDP is about $30 trillion, so the AI boom should represent a completely different scale of investment in computing infrastructure than it has hitherto.
Can’t say everything is going smoothly in this area. According to Data Center Watch, at least 75 data center projects worth a total of approximately US$130 billion have been canceled or will postpone their original plans due to local protests, and the data only covers the first three months of this year. In addition to the impact on the environment, protesters are often concerned about people leaving the labor force.
By 2050, the United States will account for approximately half, or $15.1 trillion, of data center construction design investments. The Asia-Pacific region will spend $8.2 trillion, ranking second, while Europe ranks third with $5.6 trillion. The fourth place is the Middle East, with investment of US$1.1 trillion; Africa ranks fifth, with investment of US$255 billion. At least, that’s the forecast announced by a PwC representative.
Most of the funds will be used for regular updates of data center hardware within the specified period, and direct construction of new capacity or purchase of land will have less cost impact on market participants. On an annual basis, global data center spending will grow from US$800 billion this year to US$1.1 trillion in 2030, and to US$1.8 trillion by 2050. Experts say major growth will come from China and India, as countries with large populations and dynamic economies have the greatest potential for growth. Infrastructure development in some regions may be limited by a lack of necessary components and energy sources as well as initial data requirements for AI training.
In the event of periodic disruptions in the supply of chips to a specific region, the amount of investment in data center infrastructure could be reduced by nearly 20% during a given period. At the same time, countries’ desire for “digital sovereignty” will only redistribute global investment, rather than reduce it. The authors of the forecast say there is no doubt about the availability of capital and the demand for computing power. The whole question is which organizations and regions will be able to express themselves most actively in this area over the next twenty-five years.
If you find an error, select it with your mouse and press CTRL+ENTER.
