Data infrastructure: Steel and thin air

The growth of infrastructure as an asset class in recent years has predominantly been a story of demand. Looking for returns in a low interest rate environment, investors have focused more on alternatives – and infrastructure offers an attractive risk-return mix, given its typically long-term, yield-driven income streams. The supply side of the equation, however, is far more interesting. 

This is where a number of trends around industrial-scale digitization, including smart cities and the internet-of-things (IoT), are blurring the lines that define real assets and indeed creating an entirely new category of infrastructure in data. 

Projects such as renewable energy installations or internet-connected transport facilities represent next-generation advances on legacy assets, but data infrastructure, which now binds and supports almost every business process imaginable, has proven to be more than just a modernization of copper wire. 

Many telecoms-focused infrastructure investors now favor the expression data over communications, seeing the latter as a misnomer that fails to encompass the scope of the assets involved. Matt Evans, a fund manager at AMP Capital, is one such investor and notes that as certain data subsectors become indispensable to everyday life in areas beyond their traditional communications use-cases, they are beginning to open a broader range of asset classes to the infrastructure investor.

“There is a recognition that some data-related assets previously regarded as private equity-style investments now have all the characteristics that you look for in infrastructure,” he says. “But in addition to understanding if those assets are replicable or if they have strong market position, investors must understand what makes them essential to the operation of society.”

Going mainstream

Over the past five years, AMP has built out a global data portfolio of two telecom tower projects and three fiber optics platforms. Like much of the traditional infrastructure-focused investment industry, the firm has gradually begun to see these assets in recent years as essential to people’s livelihoods, sufficiently proven in technological terms and reliable enough in contracted cash flows to be considered core.   

Some of the most recent proof of this effect includes infrastructure specialist I Squared Capital agreeing to buy a fiber optics unit of Hong Kong’s Hutchison Telecommunications for about $1.8 billion, as well as KKR and Affinity Equity Partners bidding $1.7 billion for Australian fiber network provider Vocus. But in a possible sign that levels of comfort are still being defined in the fiber space, KKR and Affinity retracted their offer last year, saying only that they were “unable to support a transaction on terms acceptable to the board.” 

KKR has nevertheless classified data as one of its two most active infrastructure areas alongside midstream energy, especially in the towers space. The GP’s key play in the segment to date has been the purchase of a 10.3% stake alongside Canada Pension Plan Investment Board in the cell tower arm of Indian telecom company Bharti Airtel for $950 million. In April, Bharti merged with Indus Towers, a domestic rival backed by Providence Equity Partners, in a deal that valued the combined company at $14.6 billion.

Notably, the investment was based largely on downside protection, with Bharti claiming about $8 billion in backlogged long-term contractual revenue at the time of KKR’s entry. According to Raj Agrawal, global head of infrastructure at KKR, if a tower business fails, it’s typically due to the asset having only short-term contracts remaining or factors such as the physical positioning of the towers not being competitive. 

“That’s because all the data technology itself is well understood, so the issue is mostly the commercial arrangements around it,” Agrawal explains. “Among the fundamental ways we get comfortable with telecom infrastructure investments is buying long term contracted assets with strong competitive positions and making sure we recuperate enough cash flow in the contracted period such that the risk is around rate of return rather than whether or not we’re going to get our capital back.”

Risk profiling

Setting maxims around the pricing of risk profiles is difficult in data, however, since the assets are increasingly being approached from different angles by investors ranging from pure infrastructure, real estate, growth capital, and even venture. While screening based on the criticality of any given asset to everyday life is a fairly common starting point, the details of due diligence will vary dramatically on investor expectations around entry timing, returns, and duration of holding. 

At the same time, data is generally believed to be more vulnerable to disruption risk than many other infrastructure categories. In addition to almost universal risks related to changing user patterns, data is seen as contending with a more volatile innovation factor, where upcoming technologies tend to obviate once cutting-edge investments at a relatively quicker rate. This can be a particularly sensitive issue for pure infrastructure investors since they will plan around a longer horizon on any given asset. 

Archana Hingorani sees disruption risk around data projects from a unique vantage point, having focused on traditional infrastructure as CEO of IL&FS Investment Managers for eight years until 2017 when she founded Siana Capital, a venture firm targeting some of the sector’s underlying technology. She says data infrastructure plays must be approached like traditional assets due to their high upfront capital requirements, but differentiators such as a lighter regulatory burden reveal they are venture and growth oriented by nature.  

“The regulation risk is significantly higher in classic infrastructure than in data because there are so many more interaction points and compliance points that need to be adhered to across jurisdictions, stakeholders and local municipalities,” Hingorani says. “There may be some contractual provision for compensation if the government intervenes, but that’s not enough when you’re running an operating asset. Either your costs could go up or you may have to stop operations.”

This point complicates the view that strong interest in data from governments represents a risk mitigation factor in the form of a stable source of long-term contracts. Although data has become a major national security concern, governments have shown that they can change the economics of a business by tweaking rules around issues such as information access rights. Further, government contracts related to more sensitive data systems can be exclusive, leaving businesses to hinge their top line on a single offtaker.  

Operational differences between traditional and data infrastructure also contribute to the way they are approached by various investors. For example, unlike most traditional infrastructure plays, the operator is not necessarily owner and manager of data projects.

The notion that developers tend to control data projects from the beginning rather than going through build-own-operate-transfer models with temporary public concessions suggests that they may represent the most inherently private category of infrastructure. 

Other differentiators for data infrastructure include a generally increased level of modularity, which in turn offers an operational flexibility and mobility essentially unknown in more firmly civic-oriented projects such as energy. 

Data is also arguably the only truly scalable infrastructure. Data projects can start small and gradually build out, whereas traditional assets – even those markedly modernized by technology – face substantial minimum size requirements for both technical and economic reasons. This effect holds true in bulky assets such as data centers, which are both individually modular and scalable in terms of building a portfolio of networked locations that ensure consistent connectivity through redundancy.

Tech or infra?

Data centers are currently the biggest gray area for investors trying to draw a line between technology and infrastructure-style plays. The hardware is not overly technical itself, but susceptible to being either outdated or supported by advances in fields such as cloud computing. At the same time, the potential for adding adjacent properties and serving multiple customers in a colocation context can fit a range of asset classes depending on case-by-case specifics. 

For pure infrastructure investors, questions include whether the income generating services are inherent to the real assets or something that can be provided by outside parties, and whether entry prices can be justified given contracts often fall under the 10-year mark. Typical data center deals are said to transact at multiples of 20x EBITDA or more with growth plans predicated on data usage increases. This risk profile can fluctuate on the credit worthiness of customers and contract length, but so far, it has kept the segment mostly in the private equity domain.  

Standout deals includes Quadrant Private Equity’s two-year, 3.4x turnaround of Canberra Data Centers (CDC) in Australia, which was acquired by infrastructure-focused HRL Morrison for about $565 million. CDC is considered one of Quadrant’s biggest successes to date and a microcosm of the evolving perception around data centers as an infrastructure investment.   

“We caught a unique asset that was not quite infrastructure investor-ready and helped transition it into infrastructure that was more attractive to a core-plus type of investor like Morrison,” says Justin Ryan, a managing partner at Quadrant Private Equity. “That’s the process that allowed us to make our return, and for private equity, that’s where the opportunity lies in this space.”

Value-add efforts included the appointment of a new CFO, the financing of expanded capacity in new facilities, and help with challenges typical of the segment such as security and optimal electricity usage for cooling servers. Quadrant also prompted the company to make the most of its government connections in the Australian Capital Territory by doubling down with a hyperlocal consolidation strategy. 

Interestingly, real estate was not a significant factor in the investment. Data centers have historically been viewed as real estate-type plays in which the economics essentially break down to rack space rental with a modest power services arrangement. But as cloud computing advances encourage more complex outsourcing models, the business model has gradually moved into the infrastructure column. 

Anthony Muh, a partner at HRL Morrison and chairman of infrastructure manager JIDA Capital, notes that a technology-driven explosion in data usage has increased demand for resilience and reduced latency, which has in turn resulted in data centers transitioning away from real estate and toward infrastructure. As tech-related services are increasingly integrated into data centers, the assets take on more of the stable cash flow characteristics that long-term, low-risk players in traditional infrastructure require.

“If we compare the risk of a road or bridge versus a data center, one thing you have to think about is the risk of disruption from shared economy services like Uber and the forthcoming autonomous vehicles in terms of what they will do to car ownership and traffic volumes,” Muh says. “But we do know with reasonable certainty that the volume of data being generated and stored is growing exponentially by almost every industry. That growth may level off at some point, but I don’t see it happening anytime in the next 10 years.”

Ready for liftoff

According to Cisco Systems, global mobile data traffic is set to increase from 17 exabytes per month in 2018 to 49 exabytes per month in 2021. This usage is understood to be predominantly based on consumer-level trends. As quickly commercializing technologies such as artificial intelligence, blockchain and IoT ramp up industrial demand for data transferring and storage capacity in the years to come, the parabolic projections could be exaggerated further still. 

All this will put extreme pressure on existing data infrastructure, inspire the creation of new branches of the industry and continue to evolve the risk profiles of existing and emerging assets. This will in turn fuel more outsourcing activity, shared utilization, and completely overhaul infrastructure’s reputation as a staid asset class with its feet firmly on the ground.

“Until recently, most investors would think of satellites as a high-risk technology play that is difficult to understand, but some are now seeing it as an infrastructure asset because the industry is starting to offer some core stable services and income,” Muh says. “We now expect to be fully connected everywhere we go, even where fiber cable is not available or possible, so satellites have become the default delivery mechanism with little competition and reliable income that infrastructure players could see as investable.”  

One of the main concerns among investors seeking an infrastructure risk profile in orbit is the rapidly falling cost of rocket transportation, which implies a lower barrier to entry in an environment where a lack of real estate restrictions seems to encourage competition. Business model replication and proliferation could also be an issue in light of industry estimates that infrastructure will be needed to support 100 million remote IoT connections globally by 2025. 

Space, however, is still generally considered exotic enough to result in prohibitively complex scaling and maintenance problems. Meanwhile, the relatively short lifespans of individual microsatellites are expected to necessitate multiple launches – a potentially deal-breaking expense that will be needed to keep transmitter arrays shipshape over the long haul. 

Companies tackling this challenge include Australia’s Myriota, which received $15 million earlier this year from a group of investors including Boeing HorizonX Ventures, Singtel Innov8, and Right Click Capital. The company is one of only a few players in the IoT infrastructure segment with direct-to-orbit capabilities. This technology allows sensors on the ground to transmit back and forth the with satellites without having to relay the signal through a nearby tower. 

“Direct-to-orbit companies have a more global and scalable infrastructure, in contrast to other systems that require base stations or 3G-4G connections that are always restricted to specific areas,” explains Ulric Ferner, a venture associate at Right Click. “For customers of this technology, installation and maintenance costs are drastically lower as you literally strip out an entire layer of infrastructure. In the near future, customer deploying IoT networks won’t care which technology is driving their data backhaul – it will simply be a question of reliability, cost, and ease of use.”

Comfort zones

As demand for internet ubiquity and big data drives cutting edge projects into the comfort zone of infrastructure investors, there will be pressure to safeguard the terminal value of assets through regular upgrades. “Infrastructure is no longer something that you can just ride out for a 20-year horizon as long as the contractual aspect is solid and the counterparty risk is small,” says one industry participant. “If you don’t stay on top of the technology, you risk having a stranded asset because it won’t be equipped the way it needs to be.”

To some extent, managing data infrastructure in this future will be a matter of close contact with counterparties that can offer guidance on technical specifics, as well as heavy reliance on investee management credentials in more advanced assets. But ultimately, building up the appropriate internal bench strength will be the make-or-break factor for investors. 

Data talent will be necessary for firms to maintain a window of visibility in their long-term operational plans such that assets don’t find themselves on the wrong side of looming technology transformations. In the more immediate term, there will be no substitute for having the right engineer on hand when the plant unexpectedly goes dark. 

“For a traditional infrastructure player to simply launch themselves into this space as a low cost-of-capital direct investor thinking that they can do it because it’s just like a toll road is a very dangerous path,” says AMP’s Evans. “This space demands expertise, and if an investor can’t access that knowhow, they will be exposed to risk when operating issues come up that they are not used to dealing with.”  

• Tweet  
• Facebook  
• LinkedIn  
• Google plus  
• Save this article  
• Send to  

• Topics
• Infrastructure
• Technology
• Buyouts
• Expansion
• Investments
• Support services
• TMT
• AMP Capital Investors
• KKR
• Quadrant Private Equity
• professional services