It seems clear to me that those projections for data center growth are overly optimistic and highly uncertain. There is a lot of FOMO by the hyperscalers and others going on; they seem to me to be unlikely to miss out on the crash.
NAREIT lists 3, Leyla, Digital Realty (DLR), Equinix (EQIX), and Iron Mountain (IRM). The only one I've looked at closely is IRM and not recently. Do no consider their financials appropriate. Of course the stocks are all soaring since the arrival of Chat GPT. And IMO they will all join in the AI crash to come.
Before AI it was clear that one can make money building data centers, but less clear about operating them. Sustaining capex and obsolescence were uncertain. Those worries have proven out. Most real estate is priced on decades of cash flows. This has never made much sense to me for data centers and it makes less sense now.
The PE chart and the exit gridlock tell the same story from opposite sides. Multiples stay near 2021 highs because nobody is selling. Nobody is selling because transacting at a lower multiple would crystallise a loss the quarterly NAV hasn't acknowledged yet. The 16,000 companies held for 4+ years aren't maturing. They're waiting. And the longer they wait, the wider the gap grows between the mark on the books and the price a buyer would actually pay.
The UBS/Blue Owl concentration stat is the quieter bomb. 60% of a fund raised through one distributor, and that distributor changed its allocation guidance. When a single channel controls that much of your capital base, a portfolio rebalancing memo becomes a redemption crisis. And as an outside investor, you can't see the concentration until it's already a problem. That's the opacity that makes private markets private.
Private markets can look smooth on paper until everyone wants liquidity at the same time. The real test of an investment isn’t just the return you’re promised, but how easily you can actually get your money back when you need it.
It seems clear to me that those projections for data center growth are overly optimistic and highly uncertain. There is a lot of FOMO by the hyperscalers and others going on; they seem to me to be unlikely to miss out on the crash.
Paul, do you know of any public REITs actively investing in data centers?
NAREIT lists 3, Leyla, Digital Realty (DLR), Equinix (EQIX), and Iron Mountain (IRM). The only one I've looked at closely is IRM and not recently. Do no consider their financials appropriate. Of course the stocks are all soaring since the arrival of Chat GPT. And IMO they will all join in the AI crash to come.
Before AI it was clear that one can make money building data centers, but less clear about operating them. Sustaining capex and obsolescence were uncertain. Those worries have proven out. Most real estate is priced on decades of cash flows. This has never made much sense to me for data centers and it makes less sense now.
The PE chart and the exit gridlock tell the same story from opposite sides. Multiples stay near 2021 highs because nobody is selling. Nobody is selling because transacting at a lower multiple would crystallise a loss the quarterly NAV hasn't acknowledged yet. The 16,000 companies held for 4+ years aren't maturing. They're waiting. And the longer they wait, the wider the gap grows between the mark on the books and the price a buyer would actually pay.
The UBS/Blue Owl concentration stat is the quieter bomb. 60% of a fund raised through one distributor, and that distributor changed its allocation guidance. When a single channel controls that much of your capital base, a portfolio rebalancing memo becomes a redemption crisis. And as an outside investor, you can't see the concentration until it's already a problem. That's the opacity that makes private markets private.
Greta post! What’s your take on the newly created AMAPS??
https://www.apollo.com/wealth/insights-news/insights/2026/05/introducing-amaps
Private markets can look smooth on paper until everyone wants liquidity at the same time. The real test of an investment isn’t just the return you’re promised, but how easily you can actually get your money back when you need it.