Jupiter Intelligence
“UNDErSTANDINg ClIMATE rISK IS ONE OF THE BIggEST CHAllENgES FOr THE PlANET.”
Founded: 2017 by ceO rich Sorkin
Combines weather, climate and terrain data to create geographically-specific risk profiles Number of employees: 50 Money raised: $33 million
Saildrone is already selling forecasts to sports teams, insurance companies and hedge funds.
More than a half dozen investors, including the foundation run by former google CEO Eric Schmidt, are backing the company with $90 million in venture capital. Pitchbook pegs its value at $260 million. Employee head count is at 100 and growing. In January, the company launched a free weather app and plans to charge for a premium version starting in June. Cheap computing power allows Saildrone to quickly test powerful weather models. But its vast trove of sensor data makes it a novel challenger to AccuWeather and its ilk. “We have a unique data source that they don’t have,” Jenkins says.
Other startups take weather data and go several steps beyond forecasting. Two-year-old Jupiter Intelligence, based in San Mateo, California, with offices in New York and Boulder, Colorado, combines weather data with information about an area’s environment and terrain to create “climate risk assessments.” It sells two services, short-term predictions of one hour to five days and long-range projections that look up to 50 years into the future.
Jupiter hopes that major cities planning for hurricanes, floods and fires, like Houston and los Angeles, will eventually become customers, but in the short run businesses affected by severe weather seem like surer sells. Any company with a warehouse in a low-lying area wants to know how many square feet it might lose to sea-level rise and when that loss might happen. The company’s insurer wants to know that too. QBE, a big Australian insurer, is already a Jupiter customer, as is Nephila, an insurance-focused investment firm. Both are also investors who have contributed to the company’s $33 million in backing.
“IBM and AccuWeather predict the weather,” says Jupiter founder and CEO rich Sorkin, 57. “We predict the impact of that weather.” Jupiter charges $200,000 to $500,000 to run a pilot for new clients. Yearly subscriptions cost $1 million and up. revenue to date is in the single-digit millions he says, but he projects ten times that for 2019.
A Yale grad with an M.B.A. from Stanford who started his career as a management consultant, Sorkin ran Elon Musk’s first breakout venture, Zip2, which developed online city guides, before it sold to Compaq for $300 million in 1999. In 2008 he made a first attempt at a startup to sell weather forecasts to businesses. His idea was to take publicly available meteorological models, pump them up with computing power and sell 30-day forecasts to energy commodities traders. But the forecasts were no better than the competition’s. He raised only $1 million for the company, Zeus Analytics, and it went out of business in 2011.
Jupiter looks more promising. The 50-person team includes talent from the federal government’s National Center for Atmospheric research and NOAA. like Zeus, Jupiter uses government-issued weather data, but Sorkin says artificial intelligence and detailed terrain-based information are producing risk projections that customers are willing to pay for. Without cloud computing, he says, Jupiter wouldn’t exist.
Another challenger with huge ambitions: ClimaCell, cofounded in 2015 by Shimon Elkabetz, 32, a former Israeli air force pilot, while he was earning his M.B.A. at Harvard. When he was in the military, he nearly lost control of his plane after a forecast failed to warn him that he was about to fly into a thick bank of clouds. At the time, he thought to himself, “Someone has to come up with a new tool.”
Together with two cofounders, he developed what he says are minute-by-minute, hyperlocal forecasts that he claims are 60% more accurate than those of competitors. ClimaCell’s edge, according to Elkabetz: In addition to the government weather data all the private forecasters use, it pulls weather inputs from new sources like cellphone signals and street cameras. “We call it the weather of things,” he says. “We turn everything into a weather sensor.”
The company has raised $77 million in venture capital, valuing it at a reported $217 million. In just the past year, Elkabetz has opened offices in Tel Aviv and Boulder, Colorado, to supplement his Boston headquarters. He is taking dead aim at many of the industries that AccuWeather serves. “We’re going to be bigger than AccuWeather,” he says. “We want to become the biggest weather technology company in the world.”
Already ClimaCell is making ground-operation forecasts for airlines including JetBlue, which is also an investor, and game-day forecasts for the New England Patriots. Via, a
ride-sharing company, uses its forecasts in five cities including london and Amsterdam. On a Sunday in mid-May, Via got an alert from ClimaCell about heavy rain in New York that would last from late morning to midday, with varying intensity in different parts of the city. Knowing demand would spike, Via made sure it had enough drivers in the wettest spots. “There is a significant monetary value to us in using ClimaCell’s platform,” says Ari luks, Via’s director of global marketplace economics.
Can ClimaCell best the incumbents? Marshall Shepherd, director of the University of georgia’s atmospheric sciences program and a past president of the American Meteorological Society, says he’s yet to see robust statistics proving the company’s claims, but adding lots of new data inputs could help produce more accurate forecasts. “I’m bullish on what they’re doing,” he says.
Others are skeptical. “ClimaCell makes a lot of claims, but I’ve never seen proof of anything,” says Clifford Mass, a longtime University of Washington atmospheric sciences professor. “Street cameras are not going to improve weather forecasting.”
Elkabetz counters that prospective customers are given proof of its claims.
ForecastWatch’s Eric Floehr is as close as it comes to an expert with a broad view of the private forecasting business. He says the jury is out on Saildrone, ClimaCell and Jupiter Intelligence. What about ClimaCell’s assertion that its forecasts are 60% more accurate than the competition’s? “Extraordinary claims require extraordinary evidence,” he says.