Plume, a Wi-Fi software company and operator of the largest software-defined network (SDN), has raised $300 million in a new round of minority equity investment led by Softbank Vision Fund 2.
It comes fresh off the heels of raising $270 million in February from Insight Partners, and the latest round brings the company’s valuation to $2.6 billion.
Plume, a mesh Wi-Fi innovator with a software-as-a-service cloud platform, has been growing all dimensions of the business fairly aggressively, according to CEO Fhari Diner. It’s a trend that’s accelerated since the funding earlier this year in terms of scale, employees and financials.
Plume’s Wi-Fi software platform grew from 22 million households in February to now 35 million. It now manages more than 1.2 billion devices – which makes its cloud controller the world’s largest SDN – and has added more than 60 new internet service provider customers. It’s averaged 2.2 million new home activations each month.
The Palo Alto-based company has plans for growth across R&D, sales and marketing, plus acquisitions. But with the latest round, financing largely had to do with who the investor was – i.e., Softbank’s Vision Fund – rather than seeking more money, Diner told Fierce.
“The way we [Softbank, Plume] think about the future in general, investing aggressively in the cloud, in data, AI, these are all kind of vectors we saw that were perfectly aligned,” he said.
Softbank for its part appreciated and acknowledged the value that has been built, and Diner said the fund has proven very well connected and influential around the globe. Already it’s introduced Plume to a number of potential partners and customers, he noted, expressing confidence they will be fruitful for Plume and therefore for Vision Fund.
The company primarily sells its cloud platform to CSPs, helping manage and deliver adaptive Wi-Fi services for smart homes and connected devices. Its open source OpenSync element, introduced in 2018, is device-level software that’s integrated into gateways, or extenders that essentially decouple software and hardware so providers can scale new services quickly and avoid hardware lock-in.
In the U.K., for example, Plume’s HomePass smart home service was made available to 25 million U.K. households with Virgin Media – even those not yet within the provider’s broadband footprint. Comcast, an early investor in Plume, was also its first major U.S. customer. Comcast uses Plume’s software integrated into xFi Pods for mesh Wi-Fi to eliminate residential dead spots.
Using data and AI for predictive tools and network analytics is a key feature of Plume’s technology. In total, it’s raised $697 million to date.
Mobile operators and 5G FWA
Plume has a large presence in the fixed broadband world, counting major cable operators like Charter, Comcast and Virgin Media, along with telcos and fixed wireless access providers among 240 service provider customers. But it also has an eye on mobile operators and telcos for FWA using licensed spectrum.
It’s a segment that is very interested in the in-home services, and a market where Diner expects to see more and more relationships for Plume (SoftBank itself has a mobile and fixed wireline business in Japan).
“We’re expanding our customer base. Mobile is a big one in the near term,” he said, as well as fixed wireless access with 5G. He noted telcos have been investing heavily in 5G and there’s been a harder push to compete with cable.
In addition to boosting Wi-Fi and smart home services, another benefit of partnering with Plume for FWA is managing licensed spectrum more efficiently.
“In the wireless world you can change the network parameters to optimize the network exactly the way we do in the home with Wi-Fi,” Diner said.
This largely comes from Plume’s scale and analytics on Wi-Fi in 35 million homes. With that data it can tell and anticipate the load a FWA network is carrying, Diner explained. For example, it can identify if there’s a cluster of homes where demand is likely going to increase on a Tuesday based on patterns but other homes where users might not be active. Plume can provide APIs with insights about which cell towers or small cells homes should connect to. Knowledge of that Wi-Fi environment gives it a unique capability for FWA operators in terms of load balancing, small cells, macro cells and radio resource management of a cellular network, he said.
“We’re not controlling the cellular network, but we provide information that enables the controller of the cellular network to use the licensed spectrum, which is very expensive, more efficiently,” Diner said.
He couldn’t provide more details about plans for the space but called out Qualcomm as a partner and investor in Plume, hinting at “some exciting initiatives” in and around FWA using licensed spectrum.
Using AI to reduce churn
Plume uses AI in a number of ways, such as to re-configure and preconfigure Wi-Fi networks based on real-time and predictive data (for example a busy Sunday evening with a family streaming 4K video in the living room). Or it can be used to secure IoT devices like baby cameras, doorbells or coffee makers with anomaly detection by recognizing what constitutes regular or irregular traffic patterns – based on indicators like IP addresses and when usage occurs – across large numbers of the same type of device.
A particularly interesting use of AI and data for service providers is the ability to help predict churn.
Since Plume knows the in-home connected device experience, it can correlate performance with customer call-in rates to their service providers to help determine if a subscriber is likely to leave.
“Churn is arguably the most expensive thing that happens to any subscription business,” Diner said. “If I were to tell my service provider customers that we believe there’s 80% chance that the following 400 customers will leave you based on the patterns we see, that’s hugely valuable information for them.”
Diner explained two of the main drivers are input Plume gathers on so-called “happiness” of devices – like if that Zoom call is getting minimum packet loss or 4K streaming is getting enough throughput. And it’s hooked into call centers based on APIs (Plume doesn’t know who is calling but knows the location and how many times they’ve called). So if there’s a group of customers that have called their service provider and they also have unhappy home devices, Plume’s algorithm could predict a percentage chance that a certain number of customers would churn in the next month. Providers could then use that information to offer a promotion or take other steps to “help get in front of a problem that is brewing,” Diner said.
It’s not 100% accurate, but Plume is able to predict churn surprisingly well, according to Diner.
Beyond residential, Plume has two other main segments, including very small businesses like coffee shops, restaurants or those that don’t have their own IT departments, which it’s already active in. A third is industrial environments, where it aims to expand.
When it comes to smart cities, Diner said Plume isn’t trying to sell directly to cities, but it presents a somewhat natural fit based on other segments.
“Public, smart city, utilities I think that will happen sort of as a blend,” he said, noting those kinds of environments want to take advantage of a mix of residential, small business and publicly owned capabilities.