The first phase of Equinix's $190 million, eight-floor data center in Amsterdam Science Park (AM4)—a key interconnection gateway to Europe and the world—will open on July 5, according to a company blog post.

This initial buildout of AM4 will provide space for 1,550 cabinets, and upon the completion of four future expansion phases, the 230-foot tall facility will support a total of 4,200 cabinets with a total usable floor area exceeding 124,000 square feet.

Read more: Equinix to Build Huge Amsterdam Data Center

This is the Silicon Valley-based behemoth’s second data center in that location, one that it claims is within a 50-milisecond network reach of 80 percent of Europe. It would have been the company’s third facility there; however, after Equinix acquired TelecityGroup early last year it was forced by antitrust regulators to sell it to rival and landlord Digital Realty Trust.

“Our new data center in Amsterdam is part of Equinix’s efforts to help our customers thrive at the digital edge. Our global colocation and interconnection platform, Platform Equinix, already includes 175-plus facilities in 44 markets on five continents,” said the company.

High connectivity makes Amsterdam Science Park an attractive location and cloud hotspot in Europe, which handles around 38 percent of the Dutch data traffic, making it the second-largest data route in the Netherlands.

Customers at AM4 will also have direct network connectivity to cloud service providers as well as the AMS-IX and NL-IX exchanges.

While an Amazon Web Services Direct Connect node resides in AM3--its existing facility in the Amsterdam Science Park--customers in other Equinix facilities in the city will have access to it through Metro Connect, the Equinix network that links all of its sites in the metro area.

Amsterdam is the third European market where Equinix offers direct private access to Amazon’s cloud.

As the data center expansion frenzy by the largest cloud providers in top markets becomes less of a frenzy and more of an incremental growth story, Equinix is positioning itself to capture the next wave of cloud data center growth, which, according to its executives, will be driven by the next generation of applications and opportunities in new markets around the world.

For the world’s largest data center provider that means making some changes to the way it develops business strategy and the way it structures deals with some of its clients.

“We’ve lived with six years or seven years of a wave — we call it kind of the first wave in cloud — and we think there’s another wave coming,” Stephen Smith, Equinix CEO, said on a call with analysts Wednesday. “Obviously, the big hyper-scalers and the cloud providers have learned from their first wave of deployments,” so the next wave “is going to be different.”

To make sure its strategy is in tune with those changes, the company has created a new business group led by one of its top executives. The new Strategy, Services, and Innovation unit consists of the office of the CTO, business development, product management, and product engineering, as well as several new business teams, all charged with “evaluating and translating key market, competitive, and technology trends into actionable business requirements.”

Charles Meyers, who’d been Equinix’s COO for the past four years or so, was appointed last month to lead the SSI unit as President of Strategy, Services, and Innovation.

New Cloud Architecture Driving Data Center Decisions

As cloud providers have learned from the first several years of deploying cloud availability zones, network nodes, and access points, Smith said, they’re changing both the architecture of their cloud infrastructure and the edge of their deployments. “Edge” in this case means all the points where an individual cloud company’s network ends, handing off traffic to another service provider who takes it to the end user.

“We’re watching all that; we’re working very closely with all of them; and yes, we want to position to move quickly as they continue to deploy, and we think it’s still very, very early,” he said.

While there isn’t a comprehensive list of all the changes taking place – the new unit was formed to figure that out – at least some of the drivers are clear.

A More Distributed Cloud

The Internet of Things requires a more distributed data center topology, where many relatively low-capacity nodes are deployed closer to where many devices generate data. Some companies have already started investing in this edge computing infrastructure – AT&T and NTT Communications are two of the more prominent examples – but that investment is bound to accelerate once the anticipated 5G wireless standard comes out, enabling the kind of data transfer rates required for next-generation applications like self-driving cars and augmented and virtual reality.

Read more: What’s Behind AT&T’s Big Bet on Edge Computing

IoT edge control points “are going to be connected to the cloud and are going to require storage, networking, and server infrastructure all over the world to drive connected cars, connected tractors, sensors, cameras — all these things that are going to get connected to the cloud and to the internet,” Smith said.

How Equinix’s strategy is going to reflect this remains to be seen. The more likely scenario is its site-selection thinking will be heavily influenced by the ability to either host edge computing nodes or quickly aggregate data from as many of them as possible; while its technology investment decisions will be focused on enabling interconnection of networks that carry that data.

“You’ll probably see Equinix taking a more active role in positioning ourselves for extending the interconnection footprint to where these aggregation points matter, and how they evolve … and how the edge moves,” Smith said.

That may mean expanding in secondary data center markets or entering new ones Equinix hasn’t traditionally pursued. The company has already seen demand from cloud providers pick up in places like Sydney, Melbourne, and Osaka after they built out their initial points of entry into Asia-Pacific — the top-tier markets like Singapore, Hong Kong, and Tokyo.

See also: Edge Data Centers in the Self-Driving Car Future

New markets this next wave of cloud deployment may take Equinix into some emerging metros in Southeast Asia, South Korea, India, South Africa, and additional Latin American countries (it already has data centers in Brazil and Colombia).

Preparing for Bigger Deals

The next wave may also mean making different types of deals with some of Equinix’s customers. While the company doesn’t do wholesale leases in their traditional form, it has made large-capacity deals with several clients it considers important strategically.

Because it gets higher returns on data center space where many players with smaller footprint interconnect, Equinix has been reluctant to do large single-tenant wholesale deals. While it doesn’t sound like it’s going to start doing those now, the company is going to be more flexible in the way it structures deals with cloud providers who need lots of capacity in particular markets.

What shape those deals will take is unclear at this point; figuring that out is one of the things the new SSI group has been charged with. Equinix is seeing more interest from hyper-scale cloud providers in larger-capacity deals, and it wants to be able to deliver.

“It is something that we are actively looking at within the context of the new SSI organization,” Meyers, the SSI group’s president, said on Wednesday’s call. “It’s to say, hey, are there a very select set of customers and requirements on the sort of hyper-scale side that we view as strategic, that we would like to increase our appetite for? And then, what are the ways that we can do that both from a design and engineering and deployment methodology, as well as how we would finance those transactions?”

Equinix this week announced partnerships with two very different cloud service providers.

One is with the German enterprise software giant SAP, whose Customer Relationship Management (CRM) cloud infrastructure can now be accessed directly via private network links inside Equinix data centers. SAP has second-largest CRM market share after Salesforce.

The other deal is with Rescale, a provider of high-performance computing infrastructure as a service. Rescale’s IaaS cloud is also now accessible by Equinix customers through direct connections that bypass the public internet. Rescale will offer access to its new ScaleX HyperLink infrastructure to meet enterprise users’ burst requirements. In other words, if your on-premises data center cannot handle a spike in demand, you’ll be able to add cloud capacity temporarily using this service.

A Big Deal for SAP

The partnership with SAP brings direct connections to the enterprise SaaS at eight Equinix data centers. This is important for SAP, considering that its biggest competitor in the CRM arena — industry leader Salesforce — began offering direct connections through Equinix in February.

“SAP joined the Equinix Cloud Exchange platform to address customer requirements for enterprise hybrid architecture in an environment that lends itself to the very highest levels of performance and reliability,” Christoph Boehm, senior vice president and head of SAP’s Cloud Delivery Services, said in a statement. “With SAP’s traditional base of more than 300,000 software customers seeking ways to take the next step in a cloud-enabled world, SAP has established efficient capabilities to deliver on those requirements.”

Direct connections to cloud services target enterprises, who are being courted by both data center operators and cloud providers. Since these connections bypass the internet, they are seen as all but doing away with security, latency and throughput issues that are thought to be keeping many enterprise customers from embracing public clouds.

In June, Equinix’s major rival and landlord Digital Reality Trust released results of a study of direct connections between its data centers and IBM Bluemix, which revealed 50X less latency and more than 55X greater throughput. Security is also improved with direct connections, because data doesn’t mingle with internet traffic.

SAP has much to gain through arrangements such as this one. The company has been paying a price for being late to the cloud game, which it entered in 2012, and doubtlessly hopes that partnering with Equinix will help its efforts to catch up. Citing 2016 figures, Forbes reported that Salesforce, which entered the cloud-based CRM space early, has seen its market share double since 2010 to 22 percent, while SAP has seen a decline from 15 percent in 2010 to around 11 percent.

For Equinix, the worlds leading global colocation data center provider, deals like this are now business-as-usual. The company has long offered direct connections to all major IaaS public cloud providers — Amazon Web Services, Microsoft Azure, Google Cloud Platform, IBM Bluemix, and Oracle Cloud — and has been on a tear signing direct-connection partnerships with major enterprise SaaS providers like Microsoft 365, Google G Suite, Salesforce, and now SAP. Going after top SaaS providers has been a big part of the company’s strategy since at least 2015.

“Equinix recognizes that enterprise cloud needs vary, and by aligning a company’s business requirements to the best cloud services they can create a more agile, flexible, and scalable IT infrastructure,” Charles Meyers, Equinix’s recently appointed president of strategy, services and innovation, said in a statement.

See also: Equinix Positioning for Next Wave of Cloud Data Center Deployments

Direct connections with SAP will initially be offered in Amsterdam, Frankfurt, Los Angeles, New York, Silicon Valley, Sydney, Toronto, and Washington, D.C., with more to be added later this year.

Compute-Intensive Hybrid Cloud

The partnership with Rescale also targets the enterprise, primarily by offering hybrid cloud customers a way to easily scale out from their on-premises infrastructure to the public cloud during times of peak demand.

“Although many enterprises have already made the move to cloud for certain data services, many are reluctant to make the move for one of their largest investments: on-premise, high-performance computing,” Tyler Smith, Rescale’s head of partnerships, said in a statement. “Rescale helps ease the migration process with a hybrid and burst solution, making best use of existing and cloud resources in tandem, and the partnership with Equinix adds a compelling layer of ultra high-speed data transfer and best-in-class security, while allowing the enterprise to keep their preferred private/public cloud model.”

Steve Steinhilber, VP of business development at Equinix, indicated that this partnership might be seen as adding to their data centers’ infrastructure in a way that will be enticing to potential enterprise customers.

“By working with Rescale, we are able to add a big compute stack to our existing interconnection offerings and IaaS capabilities,” he said, “and it enables our customers to run high-performance computing jobs in a turnkey and on-demand fashion. Through colocation or just leveraging the Equinix platform, Rescale and Equinix are providing a holistic approach to compute intensive workloads.”

Traditionally, most of Equinix’s users have been service providers.

Two years after it did a pilot fuel-cell installation at one of its Silicon Valley data centers, Equinix is making a big bet on the technology, which uses a chemical process to convert natural gas to electricity.

The Redwood City, California-based data center provider today announced a deal that will see fuel cells installed at 12 additional US data centers. The deal will be financed by the utility Southern Company, with whom Equinix signed a power purchase agreement for a total of about 37MW of generation capacity, which will be the largest single deployment of fuel cells for data centers to date.

The fuel cells will be supplied by Bloom Energy, a leader in the market whose energy servers have been deployed on corporate campuses, at data centers by eBay, Apple, NTT, CenturyLink, and at two Verizon data centers Equinix took over when it acquired the large data center portfolio from the telco earlier this year.

Fuel cells produce significantly fewer carbon emissions than traditional gas-fueled power plants; they also don’t require water, unlike power plants, which consume massive amounts of water to generate energy. Apple has made fuel cells part of its corporate sustainability program, using Bloom’s technology to provide 10MW of capacity for its Maiden, North Carolina, data center campus.

The technology, which produces energy on-site, is also considered more reliable than America’s old, outage-prone electrical grid. eBay is using Bloom fuel cells as a primary energy source for its data center in Utah, relying for backup on the electrical grid instead of diesel generators one sees in most data centers.

The most commonly cited drawback of fuel cells is their high upfront cost, which has been addressed with various financing schemes, such as Equinix’s agreement with Southern. In 2013 Bloom created a leasing program for its energy servers, giving customers the option of paying for the product over time, financed by Bank of America.

The deal will bring Equinix’s total fuel-cell capacity to more than 40MW, including its existing Silicon Valley installation and the two Verizon sites, in Los Angeles and New York.

New fuel cells will be installed at the following Equinix data centers:

• Silicon Valley: SV1 through SV6, plus SV10;
• New York metro: NY2, NY4, and NY5;
• Los Angeles metro: LA3 and LA4

Customer appetite for colocation data center services powered by renewable energy has been on the rise in recent years, and providers like Equinix have been responding by making public commitments to renewable energy goals and by striking the types of utility-scale power purchase agreements that Google pioneered to clean up its data center energy supply.

Equinix, the world’s largest colocation and interconnection company, has a long-term goal of powering its data centers with 100 percent renewable energy. According to its most recent sustainability report, it was at 56 percent renewable as of last year.

Total bandwidth on the private links companies use to interconnect their networks directly is growing faster than total bandwidth of the public internet and three years from now will outpace growth of the internet’s total bandwidth by almost two times, reaching about six times the volume of global IP traffic, according to a study by Equinix, the world’s largest data center colocation and interconnection provider.

In the first study of its kind, Equinix has attempted to quantify private interconnection and create a forecast model for growth of this type of network links. For the company, which charges customers monthly fees for every such link provisioned inside its data centers, interconnection has been the fastest-growing source of revenue; and it projects this business will only continue to accelerate, as digital business transformation marches on, spurring companies to seek more and more partners and service providers to link to without involving the internet, which is considered slower, less reliable, and less secure than private connections.

“The worldwide reliance on interconnection bandwidth is growing rapidly as businesses competing within this evolving landscape mature from exchanging traffic over a few connections through a single network carrier as the intermediary, to activating multiple direct interconnection exchange points among a global ecosystem of providers and counterparties,” the analysts wrote.

What kind of bandwidth are we talking about here? The report projects private interconnection to reach north of 5,000 Tbps by 2020. That’s enough to transmit all printed content in the US Library of Congress three times within one second, according to the analysts; or process more than half a million electronic payments per minute:

“Assuming the average value is $50, that’s $27.5M per minute or $1.6B per hour.”

So who’s linking to whom exactly? At the highest level of abstraction, Equinix splits the participants into enterprises and service providers.

Enterprises are described as organizations that aren’t at the core in the business of providing infrastructure-oriented digital services; these are your banks, insurers, manufacturers, traders, professional services firms, energy companies, government agencies, etc.

Service providers by Equinix’s definition are companies whose primary business is telecommunications, cloud and IT, as well as content and media.

Here’s how both segments’ interconnection bandwidth needs will change between 2016 and 2020, according to Equinix (click chart to enlarge):

Both categories, however, have use cases for private interconnection with network providers, cloud and IT providers, financial services, supply chain partners, and content companies. Not surprisingly, the amount of bandwidth for interconnecting with network providers far outstrips all other use cases today, and in the analysts’ 2020 projections (click chart to enlarge):

Interconnection bandwidth demand by enterprises — considered the largest growth area for all manner of service providers, from data center companies like Equinix to network, IT, and cloud providers – is projected to skyrocket over the next few years, going from 12 Tbps of installed capacity in 2016 to 547 Tbps in 2020, growing 160 percent annually (click chart to enlarge):

US has the largest portion of total interconnection bandwidth today and is expected to retail that lead over the next three years, while Latin America will have the highest annual bandwidth growth rate within that timeframe: 62 percent per year, reaching 620 Tbps of installed capacity by 2020.

Download the full report here

While data center providers in Houston weathered Hurricane Harvey and subsequent flooding without any publicly disclosed outages, Hurricane Irma, a Category 5 storm headed for Florida, will be another stress test for the internet and private network infrastructure in the South. And because of Florida’s strategic importance to network connectivity, the stakes will be higher when Irma makes landfall in the Sunshine State, which the National Weather Service says will happen Sunday.

If some buildings in Irma’s path lose power, the effects on connectivity could ripple well beyond the region that immediately surrounds it. One particular building is especially critical.

NAP of the Americas, the Miami data center and carrier hub, is the biggest network gateway between the US and Latin America, and companies in the US that rely on it alone to serve customers south of the border would not be able to reach those customers if it goes offline. In addition to being a cross-continental gateway, Miami, and especially the NAP, serves as the primary interconnection hub for most Latin American networks.

“Miami appears to be the only strategically critical communications node in the hurricane’s path,” Jon Hjembo, senior analyst at the telecommunications market research firm TeleGeography, said. “From a network perspective, what’s so worrisome about Irma targeting Miami is that, not only does it present a risk to network infrastructure in the southeast coastal region, but it threatens communications across Latin America.”

Submarine cables move Latin American traffic up to Miami, he explained, and from there traffic gets distributed back south to its final destinations in Latin American markets. Some cables that land in South America, such as the new Seabras-1 system, land in the New York-New Jersey area on the US side, “but for most LatAm networks, Miami is the place to peer with one another and get access to major content providers.”

According to PeeringDB, content and cloud service providers that host network nodes in NAP of the Americas include Amazon, Apple, Dropbox, Facebook, Google, LinkedIn, Microsoft, Riot Games, IBM’s SoftLayer, Twitter, and Tencent, among others.

Equinix's NAP of the Americas in Miami

Many companies that exchange traffic in Miami have alternate routes for their Latin American traffic, and many have network nodes in the Latin American markets they serve, so an outage at the NAP wouldn’t mean a complete cross-border network blackout, but not everybody has built such resiliency into their network infrastructure.

Nitin Rao, head of infrastructure strategy at CloudFlare, which speeds up web content delivery for companies and helps them withstand cyberattacks, said companies that have network nodes in multiple facilities inside and outside the Miami metro would most likely not experience major performance degradation had NAP of the Americas or other interconnection facilities in Irma’s path gone down.

“We can’t speak for every other network,” he said, but “a number of the larger networks will have a lot of redundancy.”

Still, “the NAP is critical enough that any impact to the NAP would result in a drop in traffic,” Rao added. “There isn’t an obvious number-two.”

NAP of the Americas is owned and operated by Equinix, the Redwood City, California-based data center and interconnection services provider who acquired it as part of a large data center portfolio from Verizon in a $3.6 billion deal that closed earlier this year. Because of its strategic importance to connectivity in the Americas, it is the crown jewel in the former Verizon data center portfolio.

The building is designed to withstand Category 5 hurricane winds and stands 32 feet above sea level, according to Equinix’s website. The company has three data centers in the Miami area and an office in Tampa, about 300 miles to the north.

In a statement, an Equinix spokesperson said the company is following its protocol for “weather-related events” this week, as it prepares for Irma. “As with Harvey, the data center staff will be prepared to spend an extended period of time in the buildings throughout the duration of the storm to ensure operational continuity. We have prepared with food, water, cots, etc.”

The company planned to close its office in Tampa mid-day on Thursday to allow employees and their families prepare for the storm.

Many of the networks exchanging traffic at the NAP of the Americas Internet Exchange, or NOTA, also use the Florida Internet Exchange, or FL-IX, for redundancy. FL-IX is distributed across eight data centers, which include NAP of the Americas and other buildings in Miami, as well as facilities in Boca Raton and Ft. Lauderdale.

If NOTA went down, their traffic would be automatically rerouted to FL-IX, and, depending on the service provider, to one or several other major interconnection cities in the South, such as Dallas or Atlanta, whichever location a particular provider has its closest network node in, Rao said.

Irma causing an outage at the NAP of the Americas is far from a given. Besides being located and designed with natural disasters in mind, such facilities have layers upon layers of infrastructure redundancy, backup generators, and on-site fuel supplies.

The effectiveness of all these precautions was demonstrated in the aftermath of Harvey’s landfall in Houston, but things do go wrong every so often, and it’s always wise to architect a network in a way that doesn’t incapacitate it if one site goes down.

Equinix has struck a deal to acquire Itconic, an Alcobendas, Spain-based data center services provider, which operates five data centers in Spain and Portugal. The American colocation and interconnection giant is buying the company from The Carlyle Group, a Washington, D.C.-based private equity firm, for €215 million (about $295 million). The deal, announced Monday, includes Itconic’s subsidiary CloudMas, which helps enterprises move to the cloud.

Itconic -- which is similar to Equinix in that its customer base includes lots of network, cloud, and IT service providers who interconnect their networks inside its data centers – brought in about $67 million in revenue in the second quarter.

The deal, signed last week and expected to close in the fourth quarter, will give Equinix new enterprise data center customers (a hotly contested market) and expands its footprint to new markets, where its clients can now extend the edges of their networks to and tap into submarine cables that will carry their traffic to and from Africa and, to a much smaller extent, Northern Europe. It also boosts the colocation provider’s pool of experts who can help enterprise clients stand up modern hybrid cloud environments.

New Edge Markets

The new markets where Redwood City, California-based Equinix will have presence once the acquisition closes will be Madrid (two data centers), Barcelona, Seville, and Lisbon.

In recent months, the company’s CEO Stephen Smith has been publicly talking about plans to expand into new markets as it positions itself for what he refers to as the next wave of cloud data center deployments. An important characteristic of that wave will be a push by cloud providers to deploy network and content caching nodes in more places around the world; physical network presence closer to more users means better user experience and lower data transport costs. This is often what companies mean when they’re talking about expanding the “edge” of their networks.

From the Equinx/Itconic press release:

By providing additional capacity in key markets, Equinix continues to play an important role in helping companies to extend their IT operations to the digital edge through the interconnection of people, locations, clouds and data.

Lots of West Africa Bandwidth

Not only do the new data centers give Equinix clients the ability to expand the edges of their networks to serve customers in Southern Europe, two of them will also give them access to submarine cables that will enable them to reach customers on other continents, primarily Africa. As more people in developing countries get online, demand for digital content in new places is quickly rising, in turn stimulating demand for edge nodes in developed countries from where American and European companies can serve those emerging markets.

Itconic’s Lisbon data center is connected to the ACE (African Coast to Europe), WACS (West Africa Cable System), and Tata Global Network submarine cable systems. ACE reaches north, to France, and south, all the way to South Africa, landing along the way in almost every country along the West African coast; WACS has a somewhat similar route, but lands further north, in England, and has fewer West African landings; the section of Tata’s TGN system that serves Western Europe lands in the north of Spain and in England, but the system is global in nature, its other strands landing in the US, Europe, Middle East, and Asia. Access to Main One, a cable that lands in Ghana and Nigeria, is in the works, according to Equinix.

The Seville data center is connected to the Canalink cable, which lands in Morocco and the Canary Islands.

Ticking the Enterprise Box

Itconic data centers in the Iberian Peninsula serve some 400-plus customers, among them enterprises with globally recognizable brands, such as L’Oreal, Deloitte, BNP Paribas, Real Madrid, and Bank of America. Attracting more enterprise clients to what has traditionally been a service provider-heavy base has been a key strategic initiative for Equinix in recent years, which the deal helps push along.

Equinix last year rolled out an aggressive plan to court enterprises, targeting companies with at least $10 million in annual revenue and 500 or more employees, of which it said there were about 350,000. Smith said then that he believed the company could add anywhere between 20,000 and 60,000 enterprise customers over the following 10 years. (The company had about 6,800 customers total when the plan was announced.)

Expanding Professional Services

Addition of the CloudMas team, which specializes in hybrid cloud architectures, cloud adoption, and migration, also in theory pushes Equinix’s enterprise initiative forward. Many traditional (non-tech) enterprises don’t have the expertise to develop the most effective IT infrastructure strategy that takes advantage of modern cloud options in-house and typically need professional help.

This isn’t the first company that provides cloud professional services to enterprises Equinix has acquired. In 2015, it bought New York-based Nimbo, which, as an official Amazon Web Services and Microsoft Azure partner, also helped enterprises shift from a traditional on-premises data center strategy to one that leverages cloud.

The stocks of all seven US data center REITs (there are now six, following a merger that closed Thursday) slid down simultaneously this week, after a well-known venture capitalist and hedge-fund owner said at an investor conference that advances in processor technology will eventually lead to the demise of the data center provider industry.

But industry insiders say his views are overly simplistic, and that history has shown that advances in computing technology only create more hunger for data center capacity, not less.

Since server chips are getting smaller and more powerful than ever, companies in the future will not need anywhere near the amount of data center space they need today, Chamath Palihapitiya, founder and CEO of the VC firm Social Capital, who last year also launched a hedge fund, said Tuesday afternoon, according to Seeking Alpha, which cited Bloomberg as the source:

Word that Google may have developed its own chip that can run 50% of its computing on 10% of the silicon has him reading that "We can literally take a rack of servers that can basically replace seven or eight data centers and park it, drive it in an RV and park it beside a data center. Plug it into some air conditioning and power and it will take those data centers out of business."

Following the event, called Delivering Alpha and produced by CNBC and Institutional Investor, stocks of data center providers Digital Realty Trust, Equinix, QTS, CyrusOne, CoreSite, DuPont Fabros Technology, and Iron Mountain were down, some just over 2 percent and others over 3 percent.

Alphabet subsidiary Google did release a paper this past April that said its custom Tensor Processing Unit chips, developed in-house, allowed it to avoid building additional data centers specifically for executing neural networks (the dominant type of computing system for AI), but the company said nothing about the implications of TPUs for other types of workloads, which collectively far outstrip neural nets in terms of total computing capacity they require.

But it also revealed in April that it's been using TPUs to run machine learning workloads in its data centers since 2015. Meanwhile, cloud companies as a group (which includes Google) are spending more on Intel chips. Arrival of the TPU has not slowed Google's investment in data centers; quite the opposite. Since the release of the paper, Google announced new cloud data centers in Northern Virginia, Oregon, Singapore, Australia, England, and, just earlier this week (on the same day Palihapitiya made his remarks) in Germany. The company uses a mixed data center strategy, building some of its data centers on its own and leasing the rest from the types of companies whose stocks Palihapitiya’s remarks set in motion.

One of those companies is San Francisco-based Digital Realty, whose shares were down 3.6 percent at one point Wednesday. John Stewart, the company’s senior VP of investor relations, said that nearly every phone call and meeting with institutional investors Wednesday and Thursday started with the investor asking about what the VC had said.

“Andy [Power, the company’s CFO] and I are in New York, meeting with our largest institutional investors, and this topic has come up as basically the first question every single meeting,” Stewart said in a phone interview Thursday.

Worries about advances in computing technology driving down demand for data center space aren’t new; it’s a concern that data center company executives have had to address periodically for many years. Computer chips powering data centers that were built in the last several years are denser (in terms of the number of cores per square centimeter) and more powerful than they’ve ever been; but during the same time, data center providers have seen a boom in demand unprecedented in scale, as companies like Google, Microsoft, Amazon, Oracle, and Uber have been ramping up investment in new data center capacity, some to support their quickly growing enterprise cloud businesses, and some to support growth in the number of individual consumers who use their apps.

Customers including IBM, Google, Apple, Microsoft, Oracle, and Amazon “are spending billions of dollars on incremental new data center CapEx, and they are doing that and signing leases with us for 10 to 15 years,” Power said. “They don’t think their data center’s going to go away.”

Bill Stoller, a financial writer and analyst and regular DCK contributor, said people who run data centers for these large companies are in the position to know the most about their companies’ future demand for data center capacity. “They are entering into long-term contracts for facilities built with today’s technology for cooling and electrical capacity,” he said. “Why would they be entering into 10-plus-year leases if this technology was obsolete. They are on the cutting edge.”

Technological progress has created numerous massive leaps in improving computing efficiency – even outside of semiconductor progress described by Moore’s Law (a growth curve that’s in fact flattening) -- the most recent ones being server virtualization and cloud computing. Neither of those leaps caused a drop in demand for data center space. Outcome of such leaps has been the opposite: more efficient computing has opened up possibilities for new applications that can take advantage of the improvements, driving demand further.

Recent advances in AI, driven to a great extent by the lower cost of processors that can run neural networks, are creating more demand for computing capacity. Servers filled with specialized chips used specifically to train and/or execute neural networks, such as Google’s TPUs, or Nvidia’s GPUs (the most widely used processors for training workloads) require more power per square foot in a data center than CPUs that run most of the world’s software. They are not replacing regular servers in data centers; they’re being installed in addition to them.

“Those higher-density racks generate more heat; they require more cooling; and these are special applications for high-performance computing,” Stoller said. In the vast majority of cases, rack densities are much lower.

Rack density indicates the amount of computing power that can be housed in a single rack and has direct implications for the amount of real estate required to host software. The data center provider business isn’t just about selling space, however; it’s also about selling power, the ability to cool equipment (the higher the density, the more cooling capacity is required for a single rack), and access to networks.

Steven Rubis, VP of investor relations at DuPont Fabros Technology, the data center REIT that specializes in providing wholesale data center space to hyper-scale giants like Facebook, Microsoft, and others, said Palihapitiya’s statements were “an oversimplification. There’s probably more nuance to it; we get this argument from investors all the time.”

Equinix has substantially expanded its presence in Istanbul, acquiring a data center and buildings for expansion from the British data center provider Zenium Technology Partners for $93 million.

The existing data center, which generates $2.5 million in annual revenue, is 16,000 square feet, but the additional shell buildings gives the site potential to be developed into a data center campus, with 130,000 square feet of data center space and 22MW of power capacity, Equinix said in a statement.

Istanbul is not a top data center market, but it is a “strategic gateway” between Europe and Asia, meaning it acts as an interconnection point for some networks that carry traffic between the two continents. A Mediterranean submarine cable system, called MedNautilus Submarine System, lands in Istanbul, linking the city to Italy, Greece, Cyprus, and Israel, according to the Submarine Cable Map by Telegeography.

Equinix’s previously existing Istanbul data center is called IS1, and the Zenium site will be renamed to IS2, the company said. In its announcement, the Redwood City, California-based data center provider said there is opportunity in helping Turkish businesses reach outside markets by offering interconnection services in Istanbul and in selling connectivity services to multinationals that use Istanbul as headquarters for the region.

Zenium was founded by its CEO, Franek Sodzawiczny, who also co-founded Sentrum, a London colocation provider whose three-site greater-London portfolio was acquired by Digital Realty Trust in 2012 for about £716 million.

Istanbul was the first data center market London-based Zenium entered, followed by expansion into Frankfurt via acquisition in 2014. It entered the London data center market last year, acquiring a site in Slough. It’s also advertising a second London facility on its website, saying London Two will be available starting this month.

Google's cloud services team has made improvements to the way traffic of customers using its Dedicated Interconnect service (launched in September) gets routed across Google's global network and signed a new partnership with Equinix to get the service into more data center locations around the world.

As the name implies, Dedicated Interconnect offers dedicated direct connections via fiber cable between on-premises data centers and Google Cloud Platform. Because they're popular with enterprise customers, all of the major public cloud providers offer direct connections, both from on-premises and select colocation centers.

Direct connections offer the advantages of speed and security. Since traffic doesn't have to travel across the public internet, it's not susceptible to being snagged by black hats, nor does it have to fight for available bandwidth. For companies that regularly transfer massive amounts of data there can be financial benefits as well, in the form of lower bandwidth costs.

Although speed and latency are automatically improved with a direct connection, the first of today's announced updates deals with the speed data travels across Google's cloud. This is important both to accommodate connected devices and because the cloud is now often a large part of an organization's private network.

Dedicated Interconnect now supports Cloud Router Global Routing, which allows subnets in GCP to be accessible from any on-premises network through the Google network. In simple English, this makes it easier to find the shortest route from start to finish across Google's cloud. This was always possible through manually configured static routing, which worked well for things like office-to-office routing but not so well for connected devices with constantly changing end points. With Cloud Router, the routing happens automatically and on the fly.

The second item is more of an expansion than an upgrade. GCP has added four colocation data center locations to the Dedicated Interconnect circle: Mumbai, Munich, Montreal, and Atlanta. In addition, the cloud provider has entered into a partnership with Equinix, the world's leading colocation data center provider by market share, to bring GCP direct connection access to areas where the cloud provider doesn't have a physical presence.

"By providing direct access to Google Cloud Dedicated Interconnect, we are helping enterprises leverage Google’s network -- the largest in the world -- and accelerate their hybrid cloud strategies globally," Ryan Mallory, a VP at Equinix, said in a statement. "Dedicated Interconnect offered in collaboration with Equinix enables customers to easily build the cloud of their choice with dedicated, low-latency connections and SLAs that enterprise customers have come to expect from hybrid cloud architectures."

According to John Veizades, the direct service's product manager, Dedicated Interconnect can be configured to offer a 99.9 percent or a 99.99 percent uptime SLA, and is now available and ready for production-grade workloads.

Equinix has created an empire by building data center hubs where companies can interconnect their networks – nerve centers of the internet and private corporate networks. In many places around the world, an Equinix facility is where a single customer can access an unrivaled amount of carriers, cloud companies, CDNs, and all other types of service providers who help ensure their applications or content reaches their intended end users.

And while simply being inside such a nerve center already makes a network architect’s life easier (because they can reach all the networks you need from one place), turning every link up is still a complicated and lengthy process. After they’ve negotiated the terms and signed a contract with the service provider, the colocation provider has to provision a physical interconnect, and it may take days, weeks, and sometimes months, for the link to go live.

In today’s world, where customers can provision cloud servers almost instantly, and where applications are so highly distributed that companies need to interconnect with more networks and in more places than ever, those lengthy connection provisioning lead times have become a problem. It’s a problem a handful of startups recently formed to address, and it’s a problem Equinix itself is hoping to solve for its customers with a new service it announced today.

The service is called Equinix Cloud Exchange Fabric, and it aims to do to connection provisioning what Amazon Web Services did to installing servers in a data center. Using new software-defined networking capabilities in the previously existing Equinix Cloud Exchange platform, customers can choose the network they want to connect to, make a few mouse clicks, and have the link up and running in a few minutes, James Staten, Equinix’s global head of market development, said in an interview with Data Center Knowledge.

Like they do for cloud infrastructure services, customers pay only for the amount of time they use the Exchange Fabric. A basic 5 megabit connection from Silicon Valley to Ashburn, for example, would cost about $350 per month, he said. Inside a single data center, a 5 megabit link will cost about $150 per month. The price goes up the more bandwidth you use and the longer distance your traffic needs to traverse.

Not only does it make network interconnection within a single facility faster and easier, it does the same for linking network nodes that sit in different cities and in different metro areas. Instead of setting up an agreement with AT&T, for example, to carry traffic between your storage cluster in Dallas and your cloud servers in Ashburn, you make a few selections through Equinix’s online portal (or your own interface that uses Equinix’s API), and Equinix handles the rest.

“It’s that facility-to-facility, metro-to-metro [connectivity] that’s really interesting and powerful,” Eric Hanselman, chief analyst at 451 Research, said, commenting on the announcement. “The challenge right now is that for existing companies to be able to construct connectivity even facility to facility is relatively complex.”

Enabling Edge Strategies for IoT

Staten anticipates the most common use cases for the new service will be Internet of Things applications: connecting data sources, such as sensor-bearing equipment or connected cars, to cloud service providers and Equinix enterprise customers’ own systems.

Manufacturers need to collect data from their products in every metro those products are sold into for quick analysis. They don’t usually do analytics in-house, so they need to connect to cloud services like Salesforce, AWS, or Microsoft Azure in those metros, he explained. Hypothetically, data from GM vehicles on the road in Dallas would be aggregated in an Equinix data center in the Dallas-Fort Worth metro, where it would be ingested by a cloud provider for analytics. Some data would then travel back to the vehicles and some would end up on GMs servers, either in the same Equinix data center or elsewhere.

Other potential use cases include movie studios, which nowadays can use dozens of contractors around the world to make a movie happen. They use private interconnection to ship content from contractor to contractor during the production process, when security and performance are essential, Staten said.

Financial services companies need to connect not just to exchanges but to all other players in the ecosystem; healthcare organizations make extensive use of private connectivity because of the strict privacy rules they have to comply with; and any company that needs to connect to devices at the edge of its network, be it to support mobile apps for its employees or for its customers, should be able to use the new Equinix service as part of that edge connectivity strategy, he said.

New Channel for Carriers

For now, the service is available for interconnecting data centers in nine markets within North America and 11 within Europe, with more locations on the way. Similar capabilities are slated to become available in Asia and Latin America next year, and eventually, Equinix wants to enable the same easy interconnection provisioning across countries and across continents.

While at first glance the offering may appear to compete with the likes of AT&T, Verizon, Comcast, and other service providers whose presence in Equinix facilities is what makes those facilities attractive to other customers in the first place, it actually makes those carriers’ services easier to consume. “It can actually be a plus for the network operators,” Hanselman said. “Equinix simply becomes a channel” for their products. Some of them are already using the new capability to extend their geographic reach, according to Staten.

The service does compete more directly with companies like Megaport and Epsilon Telecommunications, whose business models are centered on making network interconnection easier by using SDN technology. There is still value to Brisbane, Australia-based Megaport’s platform, which is focused on interconnection between facilities via Wide Area Networks than on intra-facility LAN links, and which links to many more data center providers than just Equinix, users who have no need to step out of the Equinix ecosystem may no longer be reachable for the startup. Similarly, Equinix’s new offering competes with only a part of the value proposition of Epsilon, which provides full solutions, including connectivity to branch offices and as far out to the edge as home workers.

The Equinix Cloud Exchange Fabric makes life easier for network architects by giving them an easier way to connect to service providers, partners, and clients while giving carriers a powerful channel partner and a platform that makes their services easier to buy, but at the end of the day, the biggest winner here is Equinix, which by enabling easier interconnection makes its facilities both more desirable and more “sticky” for its customers.

“All of this is helping to go feed a greater motivation to move into collocated facilities in some form or fashion,” Hanselman said. “Equinix wins when people are connected.”

Equinix has agreed to buy Metronode, one of the largest data center providers in Australia, for about US$792 million from the Ontario Teachers’ Pension Plan.

The acquisition gives Redwood City, California-based Equinix – already one of the largest colocation providers in Australia – data centers in four markets where it didn’t have footprint before, but it also expands Equinix’s capacity in the country’s two largest and most important data center markets: Sydney and Melbourne. Australia’s entire colocation industry is centered around these two cities, Jabez Tan, research director at Structure Research, who recently authored a report on the Australian data center market, said.

Sydney is the larger of the two. The Sydney colocation market generated about $380 million in 2016 and is projected to grow about 12 percent this year, according to Structure, which expects it to expand at a compound annual rate of 13 percent between now and 2021. The analysts expect the Melbourne market – which generated $124 million in 2016 – to grow faster. Structure projects a 16 percent compound annual growth rate in the Melbourne colocation market from 2016 to 2021.

The largest data center providers in Australia besides Equinix and Metronode include Global Switch, Fujitsu, Telstra, NextDC, Macquarie Telecom, Vocus, AirTrunk, and Digital Realty Trust. Equinix is the second-largest provider behind Global Switch in Sydney, and NextDC dominates the Melbourne market.

Cloud Giants Drive Demand in Australia

As of late, demand for data center services in Sydney and Melbourne has largely been driven by hyper-scale cloud providers, according to Structure. With that in mind, it comes as no surprise that Equinix in its announcement signaled that the acquisition gives it the kind of product in Sydney and Melbourne cloud giants may be looking for:

These new campuses are hyperscale ready, enabling Equinix to support requirements from high-growth global cloud service providers.

High-square footage single-client leases aren’t the type of deal Equinix usually prefers. Its strategy revolves around getting as many companies as possible to interconnect their networks inside its data centers. This interconnection-ecosystem strategy yields more revenue per square foot and makes the facilities more valuable to existing clients and more attractive to new ones.

That doesn’t mean Equinix shies away completely from wholesale deals. It does make that type of deal occasionally if the client is large and “strategic” enough. The biggest cloud service providers in the world are those strategic clients, and they are all growing in Australia.

Microsoft already operates Azure data centers in the Sydney and Melbourne markets and has a major new region in the works in Canberra. Amazon Web Services, Google Cloud Platform, and Alibaba Cloud operate cloud data centers in Sydney. Structure expects Google and Alibaba to accelerate expansion in those markets in the coming months.

Structure:

Australia, like other mature markets in the Asia Pacific region, is a magnet for massive-scale cloud and this is even more crucial for Australia given its isolation, which has a profound impact on latency and performance. Getting closer to end users and an increasingly strict data sovereignty regime makes in-country infrastructure basically a must-have.

Local Clients Important Group in Asia

Hyper-scale cloud companies aren’t the only point of the deal for Equinix, which has generally shifted its approach to Asian markets from focusing on enabling outside companies to expand in Asia to serving more local clients, who now account for a “sizable amount of its portfolio” in the region, Structure’s Tan told Data Center Knowledge Monday.

Metronode’s assets and customer base help Equinix “make meaningful in-roads with government and local businesses,” he said. “Metronode is one of few colocation providers with a truly pan-Australian footprint.”

Once its acquisition of Metronode closes, Equinix will have two additional data centers in Melbourne (where it currently has one) and three in the greater Sydney area (where it currently has four). Perth, Canberra, Adelaide, and Brisbane are all markets Equinix does not currently serve but will once the deal is complete. Metronode has two data centers in Perth, and one each in the other three locations.

Metronode will add about 860,000 square feet of land and 215,000 square feet of data center space to Equinix’s Australian footprint, with opportunity to build more facilities in some of the locations. Metronode owns most of its land, which is an important factor for a real estate investment trust like Equinix.

Modular Data Centers

One factor that may make the integration of Metronode more complicated for Equinix is that the Australian provider uses container-like data center modules by the UK company BladeRoom to expand capacity inside its facilities. But Equinix already has some experience here. Earlier this year it acquired the UK data center business of IO, which also used the modular approach.

“Equinix will likely use the BladeRoom modules to support existing customer contracts and expansions” but probably go back to its typical data center design for new deployments, Tan said.

DCK Investor Edge is a weekly column about investment in the data center market, covering both publicly traded data center REITs and privately held players in the space. More about the column and its author here.

Global interconnection giant Equinix is not changing its business model. However, the line is blurring between what constitutes wholesale data center and retail colocation deployments. This is especially the case when it comes to the largest consumers of data center space, the hyper-scale cloud service and Software-as-a-Service (or SaaS) providers, social media giants, and other planet-scale tech platforms.

Winning the battle for data center customers now requires having secure, fast, and cost-effective alternatives to access multiple public cloud services and managed services vendors. Equinix is looking to get a competitive advantage by working closely with hyper-scale customers to attract highly profitable traditional enterprise customers for colocation cabinets and private cloud deployments.

Catching the Next Wave – Part II

Speaking at the Citi Internet, Media, and Telecom conference in Las Vegas earlier this week, Smith reviewed a busy 2017 and shared the company's near-term strategy, which includes "catching the next wave of cloud deployments"– something he’s been talking about frequently as of late. This was similar to the theme from last year's conference presentation, where cloud giants like Amazon Web Services, Microsoft Azure, and Google Cloud Platform were mentioned as potential anchor partners to help with international expansion by Equinix into new countries like South Korea, India, and South Africa.

But there was a new twist this time around. According to Smith, Equinix intends to proactively develop owned facilities that will incorporate 1MW to 2MW hyper-scale nodes -- described as "collapsed architecture" -- for the dozen largest cloud service and SaaS providers, plus Facebook and Apple.

Equinix is not looking to compete for the multi-megawatt super-wholesale deals. However, this new initiative does underscore the need for Equinix to own more of its facilities to compete for these hyper-scale customers on long-term leases (10 or more years).

Smith said Equinix expects to underwrite these hyper-scale deals at a mid-to-high teens project internal return rate (IRR). This is a much lower margin compared to the low-30s IRR underwritten for the development of Equinix data center space intended for retail cabinet deployments.

Other "next wave" initiatives discussed during the session included: Software Defined Networking and Network Function Virtualization, 5G, Big Data, new applications (use cases), major changes in the security space, and future storage trends. The focus on winning the submarine cable derby remains a corporate priority, with 28 to 29 Equinix data center locations mentioned by Smith as being "perfectly positioned" to win contracts for providing access to new cables being deployed.

These initiatives are being spearheaded by a group of about 100 employees working as the "Next Wave" team led by former Equinix COO Charles Meyers. Essentially, this team is tasked with "future proofing" Equinix. The balance of the company’s 7,000 employees remain focused on daily operations and delivering the quarterly performance that Wall Street has come to expect from the connectivity-focused Equinix retail colocation business model.

Expect M&A to Continue

Over the last two years, Equinix acquired TelecityGroup in Europe, Bit-isle in Japan, completed and integrated the $3.6 billion Verizon data centers in the Americas last year, and finished 2017 with the announcement of a $792 million acquisition of 10 Metronode data centers in Australia, to be purchased from the Ontario Teachers Pension Fund. It had previously announced a $93 million acquisition of a Zenium data center in Istanbul, Turkey.

Equinix now has a market cap of $34 billion, making it one of the ten largest US real estate investment trusts (REITs). After the Metronode acquisition is completed, it will operate a network of 200 or so data centers in 52 markets, located in 24 countries.

Metronode Was Pricey

Notably, Smith confirmed that Equinix paid up in order to win the Metronode deal. I have heard the number of 31x EBITDA kicked around as the multiple. Smith said the deal was "in that zip code." It was by far the most expensive acquisition of 2017, in large part because of new institutional players and sovereign wealth funds looking to find a home for capital in the data center space.

The chief exec said that in addition to data center operators, Equinix had to beat out dedicated infrastructure funds, "and this new capital raised the bar." Keep in mind, Equinix trades at about 25x 2018 AFFO per share multiple, making the Metronode acquisition less of a stretch.

Additionally, Smith detailed that the Verizon Americas portfolio was fully integrated in December and confirmed that it continues to perform better than it was underwritten. Verizon Americas will be accounted for separately for two more quarters as a "carve-out," so performance can easily be tracked by analysts and investors. Verizon was a much larger acquisition, and the portfolio’s outperformance could in a sense help offset the high Metronode-deal multiple.

Investor Edge

The Republican tax bill passed in the US at the end of last year has changed the playing field for investors. There are concerns about too much of a good thing (GDP growth) and increased government borrowing to fund the deficit which is anticipated to grow significantly over the next 10 years.

The interest rate on the US 10-year Treasury Note spiked up above 2.5 percent last week. Many economists are predicting a 3 percent interest rate on the 10-year Note by the end of this year. The price of a barrel of oil has recently spiked up to highs not seen since December 2014. This is inflationary and could translate into a "shadow tax" on the consumer paid at the gasoline pump and for winter heating bills.

These macroeconomic factors have led to a sector rotation where materials, transports, manufacturing, and other sectors that benefit from the tax bill (and/or higher interest rates) are being bid higher. This has led to a solid start for the Dow 30, which continues to set new records almost daily. But dividend-focused sectors like utilities and REITs have not fared as well.

Data center REITs are no exception. They have sold off 4 to 5 percent during the first two weeks of 2018. Notably, Mr. Market is throwing out the million-dollar babies (data centers and wireless towers) with the REIT bathwater (slower-growth sectors). Data center landlords are putting new development capital to work at double-digit returns. Several, including Equinix, have guided to grow revenues and earnings (FFO/AFFO per share) at a double-digit clip as well to support exceptional dividend growth.

I believe that data center REITs are currently on sale, which I doubt will continue after the fourth-quarter and full-year 2017 results are announced in February. Therefore, this would be an excellent time to put some idle cash to work if you believe the secular drivers of cloud, Big Data, streaming video, IoT, edge computing, and data center outsourcing will continue.

Steve Smith, who has been CEO of Equinix for the last 11 years, resigned abruptly after mishandling an employee matter, the company announced Thursday without revealing any more details. Smith is also stepping down from the company’s board of directors.

“Steve Smith has made the difficult decision to resign as CEO, and from the Equinix Board of Directors, after exercising poor judgment with respect to an employee matter,” Equinix, the world’s largest data center colocation and interconnection provider, said in a statement.

The board has appointed Peter Van Camp, Equinix’s executive chairman and Smith’s predecessor in the CEO chair, as interim chief exec before it launches a formal process to identify a replacement.

In a statement, Van Camp said the board gave the matter “the deepest consideration” and acknowledged “the many contributions” Smith made to the company during his tenure.

He also stressed that Smith was not leaving because of the company’s performance. “I also want to emphasize that this action was not related to the company's operational performance or financial condition, both of which remain strong,” Van Camp, who was Equinix’s chief exec between 2000 and 2007, said.

Equinix has acquired the Infomart Dallas, the iconic carrier hotel that serves as the primary interconnection hub for networks in the Southern US.

The Redwood City, California-based colocation and interconnection giant said Wednesday it will pay $800 million in cash and debt to the building’s owner, ASB Real Estate Investments, a division of ASB Capital Management. ASB is the owner of Infomart Data Centers, the wholesale data center provider that's been operating the building.

“As Dallas becomes an increasingly strategic North American interconnection hub for businesses making the shift to digital, the acquisition of the Infomart will enable existing and new customers to scale their operations with Equinix,” Karl Strohmeyer, Equinix president for the Americas, said in a statement.

A report that Infomart Data Centers was exploring a sale of some or all of its assets surfaced in December. The company also has data center assets in San Jose, California; Hillsboro, Oregon (leased to LinkedIn); and Sterling, Virginia. On an earnings call Wednesday, Equinix executives said they had evaluated Infomart's other assets but found the Dallas facility to be the most valuable part of the portfolio.

The amount of network carriers present in the building make it a highly strategic asset for a company like Equinix, whose business revolves around providing companies access to networks they need to link their systems to.

The company also gets room to expand capacity in one of the country’s most important and fastest-growing markets. The 1.6 million-square foot building has enough “underdeveloped capacity” to add about 11MW of power, according to Equinix. Adjacent land on the property has room to accommodate another 40MW, the company said.

Finally, as a real estate investment trust, Equinix has to own a substantial portion of the footprint it uses to provide its services. Acquiring the building, where it already operates four data centers as a tenant, ticks the REIT box.

This is Equinix’s first acquisition of 2018, which follows an especially active year for the company in terms of M&A. Last year, the company announced acquisitions of the Australian data center provider Metronode, a Zenium data center in Istanbul, the Spanish provider Itconic. It also closed the blockbuster $3.6 billion acquisition of the Verizon data center portfolio, which was announced in late 2016.

The Verizon deal included another major southern interconnection hub: NAP of the Americas. The Miami carrier hotel is the primary interconnection point for most networks in Latin America and serves as the main network gateway between the US and Latin America.

On Wednesday, Equinix also reported full-year business results for 2017. Its total revenue for the year was $4.4 billion, up 21 percent from 2016. Net 2017 income was $233 million. The company said it expects to reach $5 billion in revenue this year.

This is Equinix’s first earnings report since the abrupt resignation of its former CEO and board member Steve Smith. He left the company in January “after exercising poor judgment with respect to an employee matter,” according to the company’s official statement.

The Infomart building has about 45 tenants, including networks, office tenants, and colocation providers like Equinix.

The building’s tenants generated about $50 million in revenue last year, Equinix said. The Redwood City giant was responsible for $20 million of that revenue, including rent and “maintenance recoveries.”

Equinix expects to close the transaction by mid-2018.

*Updated 03/08 with comments from CoreSite

A power outage at an Equinix data center is at least partially to blame for connectivity problems some of Amazon’s cloud customers experienced last Friday. The Amazon Web Services hiccup has been blamed for temporarily silencing the company’s smart assistant Alexa and disrupting operations for AWS customers including Atlassian, Twilio, and Capital One.

In a report on the AWS service health dashboard, Amazon representatives said its Direct Connect service had lost connectivity to Equinix data centers DC1 through DC6 and DC10 through DC12 in Ashburn, Virginia (all these facilities comprise a single Equinix campus), and CoreSite’s VA1 and VA2 data centers in nearby Reston.

“When the weather impacting the entire East Coast affected power at a property we lease, we activated our contingency plans,” an Equinix spokesperson said in an emailed statement sent in response to an inquiry from Data Center Knowledge about the AWS outage. “We regret that these actions didn’t prevent a service interruption for some customers. The data center affected is currently operating on normal utility provided power.  We’ve been in touch with our customers who were impacted, and we will investigate this matter, so we can prevent something like it from happening in the future.”

In an emailed statement, Greer Aviv, CoreSite's VP of investor relations and corporate communications, said the CoreSite data centers mentioned did not go down but did not explain why Direct Connect links to the two facilities had dropped.

"CoreSite did not experience a power outage at either VA1 or VA2 in terms of our uptime," Aviv said. "The event was isolated to AWS services in the US East Region. Customers must plan for unscheduled service disruptions, and CoreSite helps them by providing access to AWS services across several geographic Regions. We are proactively investigating a root cause to ensure events like this don’t impact customers going forward."

The region experienced widespread power outages starting Friday as a result of the massive nor’easter cyclone that went through it over the weekend. In Loudoun County, home to one of the highest concentrations of data centers on the planet, 15,000 customers were without power Friday, utility Dominion Energy reported. Another nor’easter was expected to slam the East Coast Tuesday night.

Direct Connect allows companies to link to AWS servers via a private network instead of using the public internet. Both cloud providers and data center providers like Equinix and CoreSite pitch such services as a more secure and reliable way for enterprises to use cloud infrastructure. A company like Atlassian, for example, may lease some space at an Equinix data center to house its servers in Ashburn and link those servers via a private network to an AWS data center in the same region. (The example is purely hypothetical; we don't know how exactly Atlassian uses AWS.)

Direct cloud links have been a core business focus for Equinix in recent years. The company has positioned its colocation facilities as hubs where enterprises can get this kind of network access to all the major cloud providers.

It’s unclear whether the power outage affected the entire Equinix campus in Ashburn or a single data center that happened to house critical infrastructure for all Direct Connect links on campus. We’ve asked Equinix to clarify and will update this story once we hear back.

Direct Connect issues at the AWS US-EAST-1 region started around 6:20 am Eastern on Friday March 2 and went on for close to four hours. During that time, “some customers” in the Equinix and CoreSite data centers had lost Direct Connect links to the Amazon data centers in the region, according to the AWS status report.

AWS does not offer a Service Level Agreement (SLA) for Direct Connect, according to an FAQ on the company’s website. It does recommend that customers set up redundant Direct Connect links to prevent outages and enable Bidirectional Forwarding Detection, which ensures that drops in connectivity are detected quickly and the redundant links are used.

Enterprise customers and service providers running hybrid cloud environments -- some in Equinix data centers, and some on Amazon Web Services and other public clouds -- can now keep all their keys encrypted with a single system.

Equinix recently announced a new SmartKey solution for customers in its 190 data centers, public cloud providers, including AWS, Google Cloud Platform, IBM Cloud, Microsoft Azure, Oracle Cloud, and Alibaba Cloud, as well as SaaS providers, such as Salesforce, SAP, and ServiceNow.

They keys are encrypted using Intel SGX technology, and the solution is powered by Fortanix Runtime Encryption.

That means the keys are secured while at rest, in transit, and even while being used.

The Intel SGX technology came out about a year ago, and the Fortanix product was released last fall. Equinix is the first data center provider to build a key encryption product on top of the technology.

"I do believe it to be the first of its kind, and it's a very useful tool," said Christina Richmond, program VP for IDC's security services research practice.

Enterprise customers and service providers commonly use Equinix for hybrid cloud deployments, leveraging the company's high-speed connections between its data centers and the public clouds, said Lance Weaver, VP for product strategy and emerging services at Equinix.

"One of the challenges is how they can maintain control of the information they put into cloud providers," he said.

The SmartKey is a hardware security module that keeps the encryption keys close to where they are needed and allows customers to use a single platform for all their keys, no matter what cloud they're in, including on-premises clouds.

"And we do it in an as-a-service model, so it's easy to deploy and no equipment is needed," he added.

Weaver sees the encrypted-key functionally as a differentiating feature for his company's data centers.

However, while Equinix may be the first to offer such a service, it probably won't be alone for long, said Bob Laliberte, senior analyst at Enterprise Strategy Group. Although Equinix' global reach will be hard to match.

"Keep in mind, many of the large cloud providers provide a service similar for their own cloud – but not multiple clouds," he added. "So this would help reduce management overhead and separate the keys from the data."

According to ESG, organizations are increasingly moving product applications and data not just to cloud providers but to multiple providers. In addition, GDPR and other regulations make it important to ensure that data is protected and encrypted, Laliberte said.

"Organizations remain largely responsible for the security of their data in the cloud, yet they often find themselves making tradeoffs between security, simplicity, and scalability," Ketan Shah, VP of products at Fortanix, said. "This presents an opportunity for data center and cloud interconnectivity providers such as Equinix to offer unique value-added security solutions."

Meanwhile, both public clouds and third-party vendors offer their own key management solutions, though none so far using the Intel SGX hardware-based security feature.

Alibaba Cloud Key Management, Amazon Key Management Service, Google Cloud Key Management Service, Huawei Cloud Key Management Service via Cryptsoft, and Microsoft Azure KeyVault are all options available from the major cloud providers, according to Robert Westervelt, research director in IDC's data security practice.

"The key difference between cloud service provider key management services and those from Equinix and others is that cloud service provider solutions like Amazon's KMAAS do not typically support multiple clouds," he said. "Equinix may be especially attractive, because it can support a variety of cloud services."

Equinix has hired former Digital Realty CTO Jim Smith to lead its new effort to sign large data center leases with the world’s biggest cloud platform operators such as Amazon and Microsoft.

Smith, who left Digital in 2015 after 11 years with the San Francisco-based giant, later joined Microsoft for a 14-month-long stint as its general manager of site selection and network acquisition but left last November, according to his LinkedIn profile. He now joins Equinix as managing director, hyperscale.

"Jim Smith is now on board as a full-time employee," Charles Meyers, Equinix's president of strategy, services, and innovation, said on the company's first-quarter earnings call Wednesday. "We're really energized about that he's bringing an exceptional experience and skill set to the table."

Equinix interim CEO Peter van Camp said on the call that the company continued “to see progress building our Hyperscale Infrastructure Team, also known as HIT, that will focus on developing facilities turned to hyperscale requirements as we discussed last quarter.”

Redwood City, California-based Equinix and Digital Realty are the world’s two largest data center providers. These days, after more than a decade of steering mostly clear of each other’s markets, they are increasingly in direct competition.

Equinix has traditionally focused on retail colocation and interconnection, while Digital’s focus through most of its history has been on large, wholesale data center deals – a market where Smith cut his teeth. Three years ago, however, Digital made a big move onto Equinix’s turf, acquiring US interconnection leader Telx.

Later, European antitrust regulators forced Equinix to sell several key interconnection facilities on the continent to Digital in exchange for approving a merger with the London-based giant TelecityGroup, giving Digital substantial presence in the European interconnection market. Digital has been laser-focused on growing its interconnection play ever since the Telx deal.

While all the major hyperscale platforms have presence in Equinix facilities around the world to take advantage of the rich network interconnection ecosystem it’s developed over the years, the cloud expansion rush of the last few years created a boom for wholesale providers, as cloud giants dumped billions of dollars each quarter into infrastructure expansion.

Last year, Equinix unveiled a plan to get in on the action. At an investor conference this January, its then CEO Steve Smith (who resigned abruptly later that month after reportedly mishandling an “employee matter”) fleshed out the strategy, saying the company would soon start building facilities that would include 1MW to 2MW “nodes” aimed at hyperscale clients.

On the earnings call this week, interim CEO van Camp said its first dedicated build for the hyperscale initiative was in early stages of construction in the Paris market (Paris 8), expected to open at the end of the year.

“We are also progressing well with financing structures that will allow us to pursue this important market with limited balance-sheet exposure,” he said. “We expect to add a handful of strategic builds across key markets over the next year, and we have a healthy pipeline of attractive hyperscale opportunities.”

It’s been a tough year for Equinix shareholders. Now, a lot is riding on strategy and growth plans to be revealed at the Analyst Day on June 20.

Equinix (EQIX) shares have been in a bear market since the beginning of the year. This is in sharp contrast to previous years, where stronger revenue and AFFO per share growth drove impressive performance for investors.

Results reported for Q1 2018 were a mixed bag. Strength in Europe and Asia-Pacific was offset by slower growth in the Americas. The main culprit was lackluster performance of the Verizon Americas data center acquisition. Also, full-year 2018 AFFO guidance was a bit lighter, coming in at $1.595-$1.635 billion versus consensus for $1.669 billion.

Notably, Verizon results are expected to rebound in 2019, as space for up to 3,000 cabinets will become available at the network-dense Miami NAP of the Americas, where colocation expansions should also drive cross-connect revenues at an even faster clip.

Tale of the Tape – Tough Sledding

Equinix shares are currently down 14.8 percent from the beginning of the year, and 9.4 percent over the past 12 months.

During the past 52 weeks, Equinix has traded in a range of $371 - $495 per share. Friday's $386 closing price was just 4 percent above the 52-week low and 22 percent under the high.

The crucial question for investors has become: Are Equinix shares a falling knife or a coiled spring going forward?

M&A Has Been a Drag on Results

Arguably, the plethora of M&A deals signed over the last year is the main culprit. In other words, Equinix’s sluggish AFFO-per-share growth rate is self-inflicted. Those deals were:

• IO UK data center business in February
• Acquisition of the colocation business from Verizon in May
• Zenium data center acquisition in October
• Itconic acquisition in October

Going forward, there will be an additional headwind from the Dallas Infomart and Metronode (Australia) acquisitions, both of which closed last month. These two deals will be dilutive to existing shareholders in 2018. They were particularly pricey: at 30x EBITDA each, which is a significant premium over Equinix’s own current multiple.

What About Revenue Synergies?

The near-term headwind in earnings growth might turn out to be a sensible trade-off versus the strategic value to the global network from assets added to the Equinix platform.

Notably, in Dallas, a portion of the parking lot adjacent to the Infomart building has been entitled for a 40MW multistory data center annex. There is still 11MW of expansion capacity available in the iconic downtown carrier hotel, which already houses four of the eight Equinix data centers in Dallas. There is existing space and land for expansion in Australia for revenue synergies from Metronode as well.

Investors will have to wait until the Analyst Day in June for Equinix management to explain the timing of revenue synergies from selling into available space and new ground-up expansion opportunities.

Hyperscale Initiative Economics

Last quarter, Equinix revealed that its Paris 8 data center would be the first built as part of the strategic initiative HIT, or hyperscale infrastructure team. These build-to-suits are intended to service the needs of the top 12 global cloud, technology, and SaaS providers. During Q1 2018, Equinix did not book or report any additional hyperscale activity.

On the first-quarter earnings call there was discussion in general terms regarding hyperscale projects being financed off balance sheet, with various investors allowing higher leverage to drive returns. Again, expect more detail on this to be revealed on Analyst Day.

Another “Known Unknown”

There is an additional variable that’s hard to quantify. Former CEO Steve Smith unexpectedly stepped down in January 2018. The sudden departure of Equinix's former charismatic leader had nothing to do with executing on the business model and strategic plan. Equinix is generally viewed as having an exceptionally deep bench of C-suite executives, including long-time CFO Keith Taylor and Charles Meyers, former COO, now president of Strategy, Services, and Innovation.

The resume of Peter Van Camp (PVC), who took over for Smith in January on an interim basis, makes him an ideal transition CEO for Equinix. He also continues to serve as executive chairman, a position he was appointed to in April 2007. Prior to becoming executive chairman, he served as Equinix's CEO and director since 2000, and president since 2006.

Investor Edge

Equinix continues to deliver on mid-teens dividend growth. The shares currently yield 2.4 percent, which is on the high end for this global interconnection giant. But it’s certainly at the low end for many income-focused investors. Still, the growth rate is the prize here.

The low AFFO payout of 45 percent is coupled with mid-teens dividend growth. The current FactSet consensus is for another 15-percent boost to the dividend in 2019, to $10.49, from $9.12 per share. Few other REITs – in any sector – come close to this type of performance.

There is a lot riding on the upcoming Analyst Day. A five-year growth plan is expected to be revealed. Equinix has an unequaled global moat, well-positioned to take advantage of the digital transformation and distributed IT architecture trends. Now, the company is layering the HIT initiative on top of the retail colocation business model.

Investors should always be wary of falling knives. However, my sense is that initiating or adding Equinix shares at current levels will be rewarded by a bounce back following the Analyst Day catalyst.

There were two main themes to Equinix’s strategy for dominating the global data center colocation market laid out during the Redwood City, California-based company’s recent Analyst Day presentation.

On the technology side, a global software defined platform that simplifies enterprise cloud and SaaS. On the financial side, a balance sheet the company intends to use "as a strategic weapon," in CFO Keith Taylor’s words.

Here are some key takeaways from the presentation:

Equinix is Unique

Unlike most REITs, Equinix is not focused on a single real estate asset, working instead on building out a global platform. The company now has 7,000 employees supporting a portfolio of 200 data centers.

Equinix has had an interconnection focus ever since the company was formed 20 years ago. In contrast, traditional REITs are primarily focused on spread investing (the difference between cost of capital and initial cash yields on acquisitions).

Equinix has invested billions of dollars in technology to build a global network, consistently investing in upgrades and expansions. The latest evolution is the Equinix Cloud Exchange (ECX) Fabric, announced in December and launched this January. The next evolution will interconnect all three global regions and each of the 200 data centers, company executives said.

The Global ECX Fabric is SDN-enabled, allowing customers to self-provision new circuits in minutes. It’s also API-enabled, providing the ability to manage all connections from a single screen. The goal is for the network is to enable enterprises to migrate applications to public cloud and provision hybrid cloud deployments across the globe.

Beyond the immediate potential win of accelerating enterprise bookings, connecting all three regions on a software-defined network has a big impact to the bottom line. Samuel Lee, the company’s president for Asia Pacific, said monthly recurring revenue, or MMR, per cabinet is 50 percent higher when multi-metro customers deploy in all three regions compared to single metro deployments. "Interconnection is the secret sauce," he said, "as well as consolidated billing for global deployments."

Services "Easy Button"

The Cloud Exchange Fabric has both in-house and partner services layered on. So, in addition to being able to self-provision circuits and cross-connects, customers will have an "easy button" to select value-add services. 

Interconnecting multiple metros using the Cloud Exchange platform launched in 2014 has been a complex job for customers, Brian Lillie, Equinix’s chief product officer, said. Removing that complexity is critical to accelerating growth in the enterprise vertical. He expects SaaS-like point-and-click provisioning across the entire IBX platform to become possible before the end of the year.

Capex Initiatives

During the 20-year history of Equinix the company has grown both organically and through M&A. The company is currently integrating eight acquisitions, most recent ones being Metronode in Australia and the Dallas Infomart.

M&A deals tend to make great headlines, but Equinix also has well over $1 billion in organic expansions underway. The largest number of project is in the EMEA region, and Taylor expects that over time EMEA and Asia Pacific will grow faster than the Americas.

Equinix is focusing investment on building future phases in existing IBX data centers, owned assets, and developing owned land with 80 percent targeting major metro markets, both existing and emerging.

Dublin and Shanghai, for example, are small markets today, but they will likely evolve into major markets over time, Taylor said.

Weaponizing the Balance Sheet

Transforming the platform is the best use of CAPEX, Taylor said. He highlighted three examples of transformational deals: the Telecity acquisition made Equinix number-one in EMEA; Switch & Data and Verizon deals expanded scale in the Americas; the Metronode deal made Equinix number-one in Australia; and Bit-isle transformed Equinix into the second-largest player in Japan.

Key balance-sheet metrics are highlighted in the slide below. Taylor is hoping to reach an investment-grade rating during the next few years, but more strategic acquisitions could push it out farther.

For the time being, the guidance is $10 billion to be spent through 2022.

Equinix has invested a total of $25 billion in CAPEX over the past 19 years, so the current five-year plan is accelerating that growth. Taylor believes that "only Equinix can make the global investment at scale." He expects Equinix to become a $7 billion-dollar company by 2022 in terms of annual revenue.

Investor Edge

Taylor laid out an expectation of 8-10 percent revenue growth on average during the next five years. Additionally, shareholders should expect to see AFFO per share grow at 8-12 percent and a 2-3 percent dividend, for a total return of 10-15 percent each year. The small dividend yield is due to Equinix's low AFFO payout ratio: less than 50 percent and among the lowest of all REITs. The balance of free cash flow is invested into organic growth underwritten at 30 percent return at stabilized occupancy.

Those are strong numbers for a REIT. However, the chart below shows that it is a deceleration from the gaudy AFFO per share growth between 2016 and 2017.

Taylor explained several times during his presentation that management is looking to grow the Equinix platform for the long term. Strategic and transformative acquisitions will continue to be a priority going forward.

Hyperscale Plans

There was one more key area highlighted during the presentation: HIT, or Hyperscale Infrastructure Team. HIT’s revenue, EBITDA, and AFFO per share contributions are not included in current full-year 2018 guidance.

We’ll focus on Equinix’s plans in the hyperscale market in the upcoming second part of this two-part series.