If anyone is looking for an ideal place to invest right now you could do a lot worse than data center REITs. They have cooled off to a safe entry point and the dividends they generate can give you some semblance of safe haven with minimal FOMO should the bull keep raging.
Specifically I recommend DLR. If you want an ETF they have one called SRVR, but I prefer DLR and CONE.
I've always wondered how these relatively small operators will fare against Amazon, Alibaba, and Microsoft. Seems at least as risky as office and retail REITs.
That's a pretty good observation. CONE is smaller but DLR is pretty big. Regardless, DLR's size still cannot insulate it from the increase in purchasing power that outfits like AWS and Azure are increasingly able to flex on them. My understanding is that in some ways AWS and Azure are acting as middlemen, housing the information on behalf of their many customers who might have previously held leases directly with REITs like DLR at higher rates.
Growth in FFO (funds from operations, the most relevant REIT metric) for data center REITs has lagged over the past few years due to this phenomenon. So the bear case is that the the glory days are over and their valuation will need to come back in line with the rest of REITs. The bull case is that industry consolidation (M&A) can somewhat remedy this power imbalance and data center REITs will continue growing as the world increasingly becomes more digital, with 5G, blockchain and IOT coming to fruition.
So I'm under no impression that DLR and their peers can reverse the downward pressure on their pricing overnight, but I don't view the trend as inevitable and something that can make the land that DLR owns less valuable. Regarding the land value, I'll paste this snippet from a SA article:
The value of each data center is largely a function of its position along the internet backbone, the physical fiber-optic network that links every connected-device across the world. Properties within the backbone, or more precisely at the "intersection" of various networks, are able to provide higher-value network-based colocation and interconnection services, which command higher rent-per-MW and generally have significantly higher barriers to entry due to the inherent "network effects."
DLR owns many of these backbone locations, whereas CONE does not. However I am partially bullish on CONE because I am bullish on Texas. (about half of CONE's data centers are in TX) I think the California exodus is not simply a clickbait celebrity trend - companies are moving, friends of mine are moving, and I think CONE's TX properties will become more valuable as data demand increases there.