Why Cisco Bought Viptela for $610M

Cisco has pulled the trigger on one of the hottest software-defined wide area networking (SD-WAN) startups, announcing that it plans to buy Viptela for $610 million in cash and equity.

Viptela, based in San Jose, was founded in 2012 by Amir Khan and Khalid Raza, who are currently President and CTO respectively. It's backed by top-tier Silicon Valley investors, including Sequoia Capital, Northgate, and Redline. According to Futuriom research, it has raised $108 million in four rounds of funding -- with the last round of $75 million raised in May of 2016.

Did Cisco overpay? Probably. Most of my sources say revenues for most of the SD-WAN startups, including Viptela, were under $100 million in 2016. It's possible that Viptela is ramping close to $100 million for 2017 -- it's got a big service provider contract with Verizon -- but Cisco no doubt paid a premium for Viptela. But it did so to acquire one of the leaders in the SD-WAN space, which has now become red hot and may include IPO candidates. Viptela could have gone public in the 2018 timeframe if Cisco did not buy it, so it was a major threat.

Cisco says the Viptela purchase "will be able to accelerate the path to developing next generation SD-WAN solutions, by combining Viptela’s cloud first network management, orchestration and overlay technologies with industry-leading routing platforms, services, and SD-WAN capabilities from Cisco."

The translation is: Cisco still doesn't have a full software-based WAN solution, so it had to buy one.

SD-WAN has been gathering momentum for some time because it's an easier way for enterprises to manage their complicated WANs, which connect branch offices or datacenters. In the case of a large Fortune 500 company, which is Viptela's target customer, it can be very complicated to tie together corporate networks in a clean and secure fashion.

Other powerful trends are driving SD-WAN: Many enterpries are now connecting directly to the cloud, to access cloud applications or services, which means that the traffic patterns on corporate networks are changing as they become more cloud-facing rather than internal datacenter-facing. This requires a different set of routing technologies that can manage connections to the cloud and internal datacenters at the same time.

Viptela is part of a whole batch of startups -- which include Aryaka Networks, Cloudgenix, Silver Peak, and VeloCloud -- that have been trying to solve the "WAN in the cloud" challenge.

Viptela has high pedigree. Khan and Raza are highly respected routing engineers (with Cisco and Juniper experience behind them) who wanted to introduce efficiencies in large corporate network by building a software-based routing engine that could handle many types of protocols, and handle them to build secure virtual private networks (VPN), without all of the configuration hassles of complicated hardware-based solutions (in other words, Cisco and Juniper gear).

I once spoke to a user who described the Viptela auto-configuration routing properties as "magical." You can imagine how much pleasure a network engineer at a bank can take in pressing some buttons that automatically configure a secure, global WAN, rather than having to manually configure a bunch of routers. That's how Viptela technology has been described.

This demonstrates the power of the SD-WAN trend and why Cisco is threatened by it. WANs used to be built in a bespoke fashion, with network managers installing vendor gear at branch offices and configuring specific connections and VPNs; they'd use private line connections they might buy from service providers, using MPLS technology. SD-WAN has shifted the game to commodity hardware that can be plugged in and configured more dynamically with a cloud application, so that management does not require as much manual manipulation of specialized routers -- and in some cases eliminates the need to buy an MPLS line from a service provider.

This approach, of course, was a threat to Cisco. Cisco sells routers. The idea of anybody being able to buy a commodity hardware box and install Viptela's routing software from the cloud was dangerous. Cisco, over the years, has been trying to put together a WAN story with a product combination called iWAN, but iWAN has mostly been ineffective in the market, according to my sources, and it has always involved a complicated and difficult-to-understand combination of many Cisco products -- often Cisco routers themselves. That kind of defeats the whole purpose of SD-WAN.

Cisco wasn't ignoring the whole SD-WAN trend -- it's been hedging all along. Earlier this year, Cisco invested in VeloCloud's Series D round. Cisco and VeloCloud had been developing a partnership over time -- in 2015, Cisco demonstrated interoperability between VeloCloud and Cisco's iWAN, and VeloCloud was given certification status to work with Cisco services routers.

So what about that valuation? Ah, yes, the valuation is rich. It's at least five times forward revenue, perhaps more. But it could also represent a markdown from Viptela's last private round of investment. According to a link from a Viptela blog, Viptela was valued at $875 million in a round in May of 2016. But Cisco has close to $70 billion in cash and short-term investments, so haggling over price may not have been the priority -- getting the right startup was. It bought one of the stronger options.

This deal now narrows the field of SD-WAN startups as the game of musical chairs is underway. Another SD-WAN startup, Pertino, was purchased in 2015 by another startup, Cradlepoint. But that deal was considered by many to be an asset purchase and probably in the double-digit millions. Cisco's purchase of Viptela at $610 million is by far the most substantial deal in the SD-WAN space, and it could lead other startups in the space to seek deals.

This deal will also set the stage for some interesting games by the other premier startups in the space, some of which are considering a run at IPOs in 2018. I'd say keep your eye on Aryaka, Silver Peak, and VeloCloud. But what is Juniper's SD-WAN strategy? Maybe we'll investigate that next.