Editor’s Note:

What’s happening in the data center market is becoming more important to those of us in the submarine cable industry with each passing day.  Gone are the days of a simple formula for bringing submarine cable traffic to the national network : landing station-to-backhaul-to-PoP.  Now, each month brings new announcements highlighting the convergence of submarine cables with the broader domestic fiber optic network in the United States.  Data centers and colocation facilities are becoming as much a part of the submarine cable network as the cable station.  In some cases, they are one in the same.

Given this, it is important for the submarine cable industry to keep abreast of developments in the data center market.  At the recent Telecom Exchange (TEX) conference in New York City, I had a chance to sit down with Keith Olsen, the energetic CEO of vXchnge; a rapidly growing data center operator.  As it turned out, we didn’t discuss submarine cables at all, but Keith’s insights into the data center market are fascinating and our industry needs to hear the viewpoints of executives like Keith as the submarine cable and data center markets become further enmeshed with one another. 

“vXchnge is a relatively young company when you think about the data center business.  The company’s management team had built a successful data center business in the 2000s to 2010 with 34 data centers. We took the company public, ran a very successful operation and then were bought by Equinix. 

While most of the management team went to Equinix, I retired.  Then the team came to me and said, “We want to do it again and you have to come out of retirement.”

So we sat down and studied the ebb and flow of things that were going on in the industry and what our strengths were.  Our model the previous time was really interesting.  It was counter to some of the big thinking that was going on that said that there were six mega markets for data centers in the US -- the Bay Area, LA, Chicago, Dallas, NYC and DC.  And yet, we built data centers in markets like Seattle, Toronto, and Atlanta outside of the big markets. And people said, Keith, why do you build in these secondary markets?  I don’t look at these as secondary (or primary or tertiary), I look at them and say, “Are they vibrant?”  Why would I want to build just where the puck has been?  I consider myself a telegeographer, not a technologist.  I study trade routes, telecom routes – looking for patterns of progress. 

As we looked at the next iteration, we felt that our thesis was pretty strong.  Atlanta, Seattle and Toronto had turned out to be among Equinix’s fastest growing markets.

So where did we want to fit in?  What’s the next group of markets?  We looked at fundamental research on demographics.  What is the population density?  This is very important to any network-based service.  How many businesses in the Metropolitan Statistical Areas (MSAs) have over 100 full-time employees (FTEs), 500 FTEs.  How many have strong metro, regional and local fiber networks.  Where’s the wireless investment.  Of MSA’s 5 to 35, which ones have strong GDP growth and are in need of data center capital deployment?  We presented our thesis to our primary financial sponsor and said here’s the business model that we’re thinking about.  They said it resonates. 

It’s all about the edge.  There are multiple great operating models in the data center business.  There are these big footprints.  They make sense for large deployments.  Then there’s the edge.  We take those demographics and say we think this is where the next iteration of deployments are going to be necessary. 

So we went in thinking about the Portlands, Austins, Pittsburghs, Nashvilles, etc.  We also believe that the former Rust Belt of the US will be in the next decade a very significant manufacturing area – high-tech manufacturing with robotics.  The area has inexpensive power, water resources and strong transport networks.  Trade requires networks, networks require locations to operate from. 

vXchnge started a little over three years ago.  We’re in 14 markets right now; some acquired, some built, some enhanced.  And we’ll continue to look at the right types of properties.  Not necessarily buying companies.  We may buy assets.  Because right now it’s a very hot market and you have to pay very high premiums if you want to buy companies. 

We’ve brought things to the market, especially enterprise and energy.  We have strong competitors.  They have big brands.  We show people the nuances - where we fit and where they fit, and neither one wants to do each other’s role.  We have terrific clients that range from the globals to the emerging growth startups.  We’re scaling up.  We started with 4 employees, now we are at 200 and we have a strong track record of recruiting good people.”

The key to running a good data center business is to be the air -- invisible.  I want to be invisible.  We put in 10-year batteries and swap them out every 5 years.  Why?  Because as long as we’re providing good, reliable services, we are invisible.  If you have good clean air, the customers come.”