How a Boston private equity firm became a major fiber and data center player

A conversation with Gillis S. Cashman, managing partner, M/C Partners


M/C Partners is a Boston-based communications-focused private equity firm, originally the media/communications arm of TA Associates and spun out in the mid-80s. The firm was an early investor in Western Wireless, Nextel, and a host of smaller wireless companies across the country. Its most recent big wireless deal helped to fund the initial build out of MetroPCS in 2001-2002.

On the wireline side, M/C was an early investor in Brooks Fiber Communications, one of the original competitive access providers, and 12-15 local carriers or CLECs. The firm followed a focused strategy of keeping those companies very regional, companies like Phone Michigan, Florida Digital, Cavalier Telephone based in mid-Atlantic.

Active in data center deals as well, it started Fusepoint in Canada in 2001, subsequently investing in Attenda in the UK, and most recently adding Involta to its portfolio last year.

During a recent conversation with managing partner Gillis S. Cashman, he said, “On the fiber side, what’s a bit unique about us, is given our sector focus, we’re flexible in our stage and style. When the markets fell apart in telecom, we did a lot of distressed investing. We took ICG private with Dan Caruso back in ’04, turned that company around.”

In late 2006, Cashman said, the fiber market started to turn around, and that was the impetus to form Zayo Group. Since launching in 2007, Zayo Group has made 35 acquisitions, done more than $1 billion in revenue, and went public last year.

“At the same time in 2007, we also formed an entity called Lightower, which was the spin-out of the fiber and tower assets of National Grid here in the northeast,” Cashman said. “National Grid was getting out of the telecom business. We split the business into two – tower assets and fiber.”

About 18 months into the Lightower venture of $290 million, the firm sold the towers off for $225 million and invested those proceeds in the network, including 7 acquisitions, before selling off the assets to Berkshire in 2010.

“We’re pretty focused on broadband infrastructure, broadband services, IT infrastructure and services, and then we’ll do what we consider tech services and communications services, which are really services that relate back to the broadband networks,” Cashman said.

Currently, Cashman said, M/C is confident in a data center company called Involta in its portfolio. “They focus in on tier-2 and smaller markets, so markets like Boise, Idaho and Akron, Ohio. We have an ERP consulting managed services and hosting business called Denovo and then we spun out the IT and data center assets of Acxiom back in July.” He also liked the potential for the firm’s latest deal, a company called Everstream, a fiber business in northeast Ohio.

Where are you seeing the growth and are you still investing heavily in fiber assets?

Our strategy is really more on the underserved markets. The interesting stat that has been out there on the fiber side…. there has been a lot of investment obviously with Zayo, and with Lightower there has been a lot of consolidation. But when you look at enterprise buildings connected to fiber, it’s about 58-60% of buildings with 20 or more employees are still not connected with fiber….and that is really pronounced in the tier-2 or smaller markets. And so from the infrastructure perspective on both data centers and fiber, that’s really been the focus is finding platforms that have built out a network and providing capital for those core backbone networks to build the lateral to connect hospitals, towers and enterprise buildings. And that was really the strategy around Everstream. They had built a very nice network. They went into northern Ohio, some of the smaller markets there. They’re in Cleveland, Akron, and in Canton, and in some of the smaller markets. They’re building connectivity. The strategy is to densify the network and connect more buildings, more towers. The backbone had been built, and needed capital to scale.

On the data center side, that’s an interesting market. You really have to parse both the market you’re in and then you as a provider – the vertical focus. I think data centers in tier-1 markets we’re staying away from. But in markets where the infrastructure doesn’t exist, we see a huge opportunity. And that was the thesis with Involta. What we’re seeing on the data center side is the change in the architecture, whereas before you had the focus on these big mega data centers in remote areas with low cost power. There was a focus on maximizing economies around power. But as you think about more applications moving to the cloud, the focus has really shifted to maximizing application performance and security, and that really requires a different architecture where the proximity of the data center becomes increasingly important for latency issues. Having those servers closer to the end user and edge, and having multiple points of presence, becomes more important. On a very macro level, instead of 50 servers in one data center, you would rather have 50 servers in 50 data centers, very close to the edge.

We’re seeing adoption of this architecture; we saw it early on with the fiber side with high frequency trading, where latency was a huge issue. We’re seeing it on the content side as more of these mission-critical applications are residing on public or private clouds, where it’ll be a bigger issue for a larger portion of the IT spend and environment.

So will that give rise to more of the companies that you’re investing in where basically theyre building out small networks/small data centers?

Yes, exactly. Sizing the footprint of your data center to address the market – like Involta, they’re going into these markets, and building out initially, 30-thousand foot data centers. They’re not building out 200-thousand foot centers. They’re having smaller footprints in multiple markets. It makes more sense than having a huge footprint in a large market.

How are they dealing with the issues around power?

They have the power they need in those markets. There are no issues there. That’s part of the criteria for entering markets. You examine power availability and cost when you look at specific markets.

Are you expecting to see and therefore invest in a lot of these smaller companies? Will we see more of a consolidation effort by larger companies to move to these smaller data centers?

Part of the thesis around Involta was that it was an acquisition platform, so we’ve made a couple of acquisitions and continue to look for them. The more distributed architecture is making single-site owners realize that being part of a larger footprint better positions them long term. So you will see more of these single-site, single market-focused data centers consolidated. And the deal activity that we’ve seen in the last 12- to 18 months we’re starting to see consolidation happening. A lot of this because of the need for a broader footprint.

Would you expect to see any of the big telecom companies do this?

They have done some of this. They were among the first – we saw CenturyLink buy Savvis, we saw Verizon buy Terremark, Windstream bought Host Solutions. The problem is it’s a very different business and customer support model. And so a lot of the bigger telecos have failed when they have tried to integrate the sales and support functions into their core businesses. When you have companies that sold voice and data trying to sell data center solutions, that just doesn’t work. It’s a very strategic product set for telcos, and it should be a part of their business when they think of addressing enterprise needs with their own customers. With the companies that have done it well are the companies that have left that as a separate business unit and a separate sales and support function.

Do you expect any major players to make moves in this area?

There are some who have been inquisitive in this area. We’ve seen TierPoint, who has been out there recently. CyrusOne bought Cervalis, saw Zayo with a lot of activity in the data center side. When we started that business, it was all fiber. Through a number of acquisitions, they now have a footprint of 40+ data centers in the U.S. and Europe. So you now have a handful of data center companies trying to consolidate a region. It’s been very active, certainly, larger platforms out there trying to consolidate the space.

What will M/C Partners be doing in the space in the coming months?

The focus for us is really building out the Involta platform. We have a pretty tight focus on the markets we want to go after. We always, in a given fund, do multiple investments in a given sector. But the primary focus is putting capital behind Involta to consolidate some of these smaller markets.

Is there a bit of a land grab in these smaller markets whereas the first mover captures that market and the second and third movers miss out?

That’s exactly right. That’s the nice thing about these tier-2 and smaller markets. If you are first to market there, it really doesn’t make sense for others to move in to that market. For example, like in Akron, Ohio or Boise, Idaho, these data centers are not cheap. And with Involta, they’ll go in and lock up the top employers as anchor tenants, and those anchors will fund the build out of the facility. Once that’s done, they’ll go after the next tier of employers. And so, it really is in those markets as we think about a proprietary market position, because when that next company comes in, their opportunity doesn’t justify the build out of the market.

Are you seeing municipalities becoming engaged in this shift at all?

We have seen some of that.  I know in Ohio, there were a lot of tax incentives given to Amazon for this. Municipalities are definitely seeing the benefit of having these data centers in their markets. Not all states, it’s not broad-based yet, but we’re seeing more of it.