Friday File: Social SPAC, Criteo, Bitcoin and Data Centers

By Travis Johnson, Stock Gumshoe, September 15, 2017

I’ve already shared the major portfolio change I made this week, adding that IPO to my Real Money Portfolio yesterday, and I have also separately posted my notes and thoughts from the CNBC Delivering Alpha conference for my favorite people (yes, that’s you!)…. but I have a couple other things to note as well, so I’ll post those thoughts separately here, for the sake of clarity and convenience.

Here are the other two pieces Irregulars received from me as we close out this week:

My Notes from Delivering Alpha
Buying an IPO. Again.


Social Capital Hedosophia SPAC (IPOAU) — I should reiterate that because of the nature of this IPO buy I made yesterday, the Social Capital Hedosophia SPAC (IPOAU), there’s likely to be no real rush — I bought it because it seemed fairly priced at just a 3% premium to the cash in trust that will be used, one hopes, for a good acquisition. But it should be reasonably priced, these kinds of issues do occasionally get to silly and overpriced valuations before there’s real news if they’re connected with a high-profile manager, but more often they just fluctuate around pretty close to their $10 IPO (and redemption) price until an acquisition or merger decision gets announced (or rumored). I do think it’s perfectly reasonable to buy up to $10.50 because of the value of the warrant, and because I’m willing to risk the fact that my money won’t earn anything for two years in exchange for getting access to this particular venture capital investor as he pursues his plan to acquire one of the relatively mature public tech companies (like the billion-dollar plus “Unicorns”), but more than that is tougher to swallow.

Really, I just want to be clear that this IPO shouldn’t change much in price until they decide on an acquisition offer to make. The story was well-covered by CNBC and did make the Wall Street Journal last night, so it’s certainly no “deep secret” that you have to get into before everyone else, and it’s reassuring that the SPAC units haven’t popped too foolishly, but this will be a “sit and wait” story for quite a while unless Chamath Palihapitiya moves a lot more quickly than most SPACs… and it’s quite possible that the shares could drift under $10 if we have a long wait with no news or rumors.


Criteo (CRTO) — I’m a little perplexed at this one, it took a really sharp drop this week on no news, other than the “news we already knew” that the next version of the Apple operating system will make things tougher for ad-tech companies (particularly independent ad tech companies like Criteo who don’t have “captive” daily users like Google or Facebook). But we knew that was coming, and that was why the stock took a hit in June on the announcement… and why I sold half of my position, as the company’s future became riskier to project. But why is it down another 10% this week?

I have absolutely no idea. It seems that it’s just the ad-blocking story returning — mostly because the ad industry got together this week to “blast” Apple for the cookie-blocking settings in the newest version of Safari. I don’t think the actual ad blocking news is particularly different than what Apple announced back in June, but it is at least getting a lot more attention now… and is more imminent. So I’ll reiterate what I said back in June when I took profits on half of my position to reduce my risk exposure:

“I don’t think this will destroy Criteo’s business — but I do think that the increasing focus on anti-tracking regulation and on default cookie-destroying or anti-tracking settings in new browsers is a growing problem, and I think that trend substantially increases the risk to a small company like Criteo that doesn’t have very much leverage over the large players in the industry. Ironically, it may also increase the odds that Criteo gets acquired by someone, since their customers and systems would probably be more valuable in the hands of either a massive ad network or a telecom company.”

And I’ll still hold, at least for now. The risk is higher than it was when I first bought Criteo a few years ago, but the company remains cheap in an industry where most smallish competitors aren’t even profitable (CRTO has a forward PE of about 14), so the reward, if these risks turn out to be less horrible than feared by investors, should be substantial. I’d still put far more of my “internet advertising” investment into Facebook and Alphabet, who are effectively a duopoly in the business, but there’s a bit of space in the corner for our little French friend, too.


And though I’m throwing in my lot with Chamath Palihapitiya in some part, by buying into his Social Capital Hedosophia SPAC this week, I’m also betting against him in some ways.

How so? Well, he had some quotes at the Delivering Alpha conference that were bearish on data centers, and I still own data centers… and bullish on bitcoin, and I sold essentially all my cryptocurrencies this week.

First, the data center REITs.

I still hold CoreSite (COR) and call options on QTS Realty (QTS), both data center REITs, despite the fact that part of his talk at the Delivering Alpha conference included some musing about the fact that data centers are on the verge of becoming extinct.

It’s a measure of how much attention a high-profile and charismatic investor gets when he receives CNBC time, but that sentiment — almost a throwaway in his discussion with Barry Sternlitz — caused many data center REITs to drop 3-4% or more almost instantly, and CoreSite, as the best performer in that group (by a long shot) fell the worst, down almost 9% since the conference.

Part of that’s because Barry Sternlitz is a very widely respected real estate investor, and he agreed with Chamath on the lack of appeal for the data center REITs… and part of it, I expect, is simply the fact that folks look for any reasonable reason to take profits after a run like these stocks, especially COR, have had in recent years.

The quote from Sternlitz was that he’s been avoiding (not shorting) the data center space since talking it over with Palihapitiya: “It’s not going to happen overnight, but investing isn’t about overnight, and if you can see the train wrecks coming you can take advantage of them.”

Is that sour grapes for a real estate investor who missed out on a huge run, a bet that doom is coming in the near future for COR and DLR and EQIX and the others, or just a rationale for avoiding an expensive sector that has probably already had it’s biggest move?

There’s nothing wrong with avoiding a sector where you think the tailwinds are going to turn into headwinds. I’d agree that data centers are going to be less needed over time as chips get more and more powerful and intelligent, particularly as the rate of improvement in chips outpaces the rate of growth in cloud services and storage demands and distributed AI services like Alexa and Siri — and most of Chamath’s argument is that Google, in particular, has developed its own proprietary chips, at least for the deep learning/AI processing they depend on, that are far more efficient than standard CPU or GPU chips and require far less data center space than their growth in processing needs would otherwise require. And really “far less” is not an exaggeration, Google was looking forward toward the need to double its datacenter footprint because of the computational demands of “always on” machine learning for core products like search, and this faster Tensor Processing Unit they developed over the past couple years cuts that demand dramatically.

So that’s all true, but I’d argue that it’s not necessarily an imminent catalyst — to me it seems more like a continuation and perhaps acceleration of the trend toward more efficient data centers, which has been meaningful for a decade or more but certainly hasn’t cut into demand for data centers yet (demand has continued to grow, too). Google is still building giant data centers and still buying data center chips from Intel… Amazon and Apple and Microsoft and Facebook are still buying and leasing tons of data center space.

Perhaps these advanced technologies and AI processors will decrease the demand for data