SNUPI Technologies wants to help homeowners spot mold, leaks, and appliance failures before they become a big problem, but that’s just the beginning.
This Seattle startup is building a platform at the intersection of low-cost, low-power sensor networks, cloud computing, and machine learning, that is showing potential beyond the initial product and service that SNUPI is bringing to market.
SNUPI is also notable for its location in a surprising space on University Way—better known as The Ave—pioneering what city and University of Washington leaders see as Seattle’s next hot startup neighborhood.
SNUPI—an acronym for Sensor Nodes Utilizing Powerline Infrastructure—is getting plenty of attention, and for good reason, starting with its founders and leadership. The underlying technology was developed at Georgia Institute of Technology by co-founders Shwetak Patel, a MacArthur Genius Grant Fellow (pictured above, middle), and Matt Reynolds (above, bottom), now computer science professors at the UW. UW PhD student Gabe Cohn (above, top) helped refine the technology and is another co-founder. Jeremy Jaech, one of Seattle’s top serial entrepreneurs, is CEO.
In a little more than 18 months, SNUPI has taken an academic prototype and turned it into a product it’s marketing as Wally. The Wally name is meant to be friendly and approachable, and is a reference to the fact that the technology wirelessly taps the electrical wiring in the walls of a building, which serves as an antenna to link sensor nodes to a central hub. This innovation sips power compared to other wireless communications and allows the sensors to last for 10 years, the company says, on one battery.
That’s important because the sensor nodes—capable of detecting changes in temperature, moisture, and humidity—are designed to go in hard-to-reach places like attics, crawlspaces, and under appliances. Once in place, they quietly monitor for changing conditions that could indicate impending hazards. A spike in temperature around a water heater, for example, could indicate that the inner core has failed and tank rupture is days away. Wally would detect the spike and send an alert via text or email, potentially saving a big mess and inconvenience.
While the product development work has occurred quickly, it hasn’t always been easy.
“A lot of the work initially was trying to get the knowledge that had been learned in the lab into the heads of the guys who were trying to build production systems,” Jaech says.
“The theoretical models and simulations never actually match the real world situation,” Patel says, adding that it was critical in this process to have the company located close to the UW campus.
The door to SNUPI’s office is situated between a skate shop and a burger joint right on The Ave, complete with its grittiness, interesting people watching, and smells of food and incense. It’s not where you’d expect to find a startup, but climb the stairs to see the exposed brick and office dog, and you might as well be in Pioneer Square.
The university and the city are in the midst of a major planning effort aimed at turning the neighborhood west of 15th Avenue into “a regional center for innovation, knowledge, and creativity.” There’s lots to recommend the area, principally its proximity to a UW campus increasingly focused on spinning out startup companies, and the arrival of the Sound Transit light rail stop in 2021, which will strengthen connections to the rest of the city. (The neighborhood’s development is also a stated focus of the Startup Seattle initiative launched by the city in May.)
“It’s happening,” Jaech says. “I think this is going to be the next Pioneer Square or South Lake Union.”
The company has about 25 people contributing right now, including 14 full-time employees, seven part-time, and several interns. And, of course, it’s hiring: in operations and engineering. The company raised $1.7 million from Madrona Venture Group, Radar Partners, and the founders last year and is in the midst of raising its next round from existing and new investors.
SNUPI is about more than sensors and low-power wireless communications. The bigger challenge is doing something useful with the sensor data, and making it all easy and reliable for consumers.
“You can really start collecting a lot of interesting data over time as you get a number of these sensors out there,” Jaech says. While individual sensors might have a fairly narrow purpose, “as you start cross-correlating the data, you can draw some really interesting inferences about what’s going on in the house.”
As SNUPI accumulates this data, it plans to publish a series of generic home-health suggestions. Eventually, as enough data is gathered about an individual house outfitted with these sensors, it could offer specific suggestions. Indoor air quality readings might suggest it’s time to replace your furnace air filter, for example. After many years, the system could even be used during a home sale to prove to the buyers that the home was well cared for, with sensor data demonstrating that there were never mold problems, or that if there were, they were rectified promptly, Patel says.
There are lots of new approaches to home automation, security, and hazards detection. Best known is the learning thermostat from Nest, a Palo Alto, CA-based company whose vice president of technology is Yoky Matsuoka, a former UW computer science professor and MacArthur Fellow. In recent days, Xconomy has written about Canary, a New York company that aims to do home security monitoring through a device that connects with a smartphone; and Revolv, a Boulder, CO-based startup trying to make managing the proliferation of smart home devices easier. Something called Spotter, from New York-based Quirky, appears similar to Wally, but it requires a pair of AA batteries and a Wi-Fi connection.
At the high end of the market, sensor systems—required by some insurers to reinsure after a leak—can cost $5,000 to $8,000 and require a plumber and electrician to install, Jaech says.
Insurers will also typically offer a discount to homeowners with these systems. Asked whether SNUPI has made progress in negotiating insurance-company incentives for its systems, Jaech says only, “Yes.”
The company believes its wallyHOME product will stand apart, opening a new middle-tier market with its $299 price (for a starter kit with a central hub and six sensor nodes, as well as the cloud-based service that alerts you when something’s amiss), easy setup, modern communications tools, and long battery life. The product is expected to hit the market in spring; SNUPI is accepting order reservations now.
Jacquelyn Jaech, SNUPI vice president of marketing and spouse of Jeremy, says they initially plan to market directly to consumers at home shows before delving into retail sales. The company has also been approached by property managers who have a major incentive to detect potential hazards that tenants might ignore, and to spot them before they do damage to adjacent units. They are considering a commercial version of the product, she says.
In an era of growing public awareness about surveillance, big data, and its privacy implications, SNUPI’s leaders say their products and services pose little risk. For starters, the transmission of data within the home itself is done using very low power and is difficult, even for the system itself, to detect. When that data is sent to the cloud, it is encrypted. Patel says the company will be transparent with customers about how it uses their data, and why.
“As a company and as a brand that has the power to look into the data, we have to be explicit about what we’re doing with it,” he says.
While we were talking, Patel showed me a text message alert from Wally indicating moisture was detected on the main floor of his home. He was not dismayed. It turns out the white, curved, computer-mouse-sized nodes are a favorite toy of his 1-year-old daughter.
“That’s her licking it,” he says, adding, “I have probably eight or nine of them. A couple in the car. A couple in her little toy bin. And there’s one in the diaper bag. This calms her.”
Patel never imagined this use for the technology. The company has heard from people who want to use it to monitor wine cellars, rare-book libraries, fish tanks, and art collections.
This is an important indicator of a product that has real potential, says Jaech, a 30-year startup veteran with category-defining products like Aldus PageMaker and Visio business drawing software to his credit.
“In both cases, there were way more uses of the technology than we envisioned,” he says. “And I think we’re going to find the same thing with this.”