The Age of Fake Datacenters Part 3 - Customers Speak Up

[The Age of Fake Datacenters and The Age of Fake Data Centers, Part 2 - Faster and Faker]

For those who have read my previous research notes on Fake Data Centers, welcome back. For those who haven’t: in brief, the data center market is stuffed with entrants trying to take advantage of a unique confluence of supply shortage and demand uncertainty. These new entrants? can’t really deliver what they’re promising - they are a mix of the well-meaning, the rent-seeking, some genial layabouts, and a group of outright grifters.?

At PTC a few weeks ago, I spent time gathering actual data on Fake Data Centers. And not by talking to the faux developers - instead, I’ve been talking to real data center end-customers. The “Hyperscaler Plus” group of at-scale buyers. And I asked them about Fake Data Centers. From my unscientific survey of decision makers, strategic negotiators, infrastructure business developers, and due-diligence Technical Program Managers, I’ve gotten some very similar answers.?

75% to 90% of data center projects being proposed can not pass the most cursory due diligence and fails at the “I read the pitch deck and this makes no sense” stage. That number goes to a straight 9 out of 10 when any actual due diligence is applied. That means out of 100 emails in the mailbox of a typical buyer, 90 of them are useless and worse than useless - they are a waste of time.? Entire teams of contractors, believe it or not,? have been deployed to separate the wheat from the chaff.

This has caused the contractual equivalent of a distributed denial of service attack on datacenter and colocation acquisition teams across the industry.? Amongst the tales of woe that I’ve heard, the easiest situations involve the worst sorts of unforced errors allowing rapid elimination - “100 Gigabit Internet Available” being the one that causes the most mirth. The more complex the pitch, the more difficult it is to eliminate the fakes and the more time it consumes.?

That being said, now that we have data, the Fake Data Centers fall into one of several categories from the buy-side perspective. They include:

• A good or at least acceptable team (meaning actual subject matter experience) with a project which isn’t capable of development. This is always unfortunate because a team typically is only going to get one bite at the apple - no one has the time to iterate on projects that can’t successfully pass through the wickets. That being said, this sort of scenario can sometimes work out because a high quality team could successfully develop a poor site by identifying a specific buyer who is willing to ameliorate or ignore a specific issue. This is, however, a very poor wager to make.?

• A weak team with a potentially successful project - of course, you would never know if the project would actually be successful because the buyer would have to supplement the seller’s project team and no one has the time to make it work. At the same time, those teams generally want a premium on the site, not understanding that a significant part of the value is the ability to advance the project to the “goal status” - powered land, powered shell, or build-to-suit. Unsurprisingly, I spend a lot of time working on these as there is an obvious economic inefficiency.?

• Inexperienced teams with poor projects - in some ways, this is the best scenario for the buyers because they are obvious and can be rapidly discarded. The “tells” are numerous and they tend to be the same across such projects. This is sadly the most common scenario. As is sometimes said, if the CEO of your ostensible data center developer has never actually been inside a data center, this may be you. I do admire your confidence and perseverance, truly. But maybe it's time to partner with someone who has done this before, once or twice?

I keep mentioning the teams. And the presence of an experienced reality-based team is absolutely vital. Without a great team, there is no possibility of success - as Machiavelli said,? Gold will not always get you good soldiers, but good soldiers can get you gold. And the emphasis is on the plural - you don’t become a real data center project or company with a single individual no matter how skilled or experienced. Only an experienced and well-crafted team can make the magic happen on a significant scale. And I’m not talking about the existing players, only - there are plenty of newer developers who have hired experienced teams - Crusoe’s hiring of the old 365 Main team comes to mind.?

The other thing you can do, if you don’t want to hire real data center professionals - and I do beg you to do so - is to partner. If you’ve got a likely piece of land, perhaps approaching a QTS, Compass, EdgeConnex, Vantage, Iron Mountain, NTT, Digital Realty, or the host of other companies with real experience “going vertical” will be helpful. Because, in addition to that experience, they have MSAs, contract vehicles, and most elusive of all, end-user trust - the one commodity that coin cannot buy.?

Finally, when we talk about Fake Data Centers, several readers have been concerned that I'm talking about their specific project or that they will be singled out. Ironically, this is a self fulfilling prophecy. If you think you're running a Fake Data Center project, you either have the world's worst case of Imposter Syndrome or, it's true. It might be worth some thought on that.