Benedict Evans’ Post

Benedict Evans do these trends look different in Europe (with differing cultural / RTO mores) or Asia (where the growth story continues)?

These seem independent of each other and more correlated to macro economics. Could just as easily be music tastes. Maybe warehouse money becoming datacenter money? But other than Amazon, the spenders would be very different, just like the spenders for offices and retail are very different.

What is the correlation between these? Do you expect a collapse of data centers? I would have thought that with the power hungry AI boom it will grow even further and put even more pressure on power consumption and water needs?

Not surprising that surge in DC is occurring at the expense of warehouse capacity. For the same area, DC is more profitable even after fitting power & cooling, which explains why big players like Prologis are actively converting warehouse to DCs.

My two cents on this: https://www.linkedin.com/posts/davidtimis_ai-futureofwork-technology-activity-7363132666906378241-R63f?utm_source=share&utm_medium=member_desktop&rcm=ACoAABGsJuIBpH058iTyChDrY305_8-V4G622XU

The 2000-era datacentre bubble looks surprisingly small. I remember working as an equity analyst & my first research note involved covering what was then called colocation facilities rather than DCs. Buildout was measured in sq metres rather than megawatts. I called overcapacity when I realised nobody had accounted for massive imminent rises in server density per rack…

When i started in Real Estate in London in 2009, the most prestigious jobs were in London Offices and Prime Retail, and that is where the best graduates flocked. Nobody wanted to work in Warehouses (unless you were very forward thinking or lucky). That status quo quickly got up-ended. Maybe its being up-ended again. Its also fascinating how what happens in CRE is such a mirror of the mega trends in the broader economy.

Death of distance. Been professed for a long time, most famously by Cairncross in 1997. And all that time, it was proven false (cities boomed like never before). But it seems that first covid, and now AI, are finally putting a dent. Working from home has taken root and while many companies are back to the office, most work from home at least a few days per week. High street retail never recovered. The curious line is that for warehouses - I wouldn’t have expected that to drop as much as it did. Maybe a trade policy effect?

Tom Humphreys

Much appreciated Benedict Evans. Two questions come to mind: 1) Overall construction volume seems to be down a bit. Is this a positive for further data center growth? 2) How far do you think it can run? Assuming that AI > e-commerce and the ascent being not as steep, one could extrapolate that it will continue to rise until 2029/30.