Skybox view of the Ras Tanura Najmah compound in Saudi Arabia

Skybox

Skybox view of the Ras Tanura Najmah compound in Saudi Arabia

General


So you heard that Google is buying a satellite imagery company called Skybox Imaging for $500 million? Will the world’s most powerful search engine start snapping paparazzi shots the next time you’re sunbathing nude in the backyard? Will the NSA start taking notes on the exact time you left the office? Will Amazon need to find a safer hiding spot for its delivery drones?

Well, not exactly. But you can be excused for being creeped out by the notion of frequently updated, sub-meter, high-resolution satellite photos of everything in the world not covered by a roof — and 90-second black-and-white videos, too.

So what exactly does Skybox do, just how much can those satellites see and how will that be harnessed by Google?

For Google’s part, the company says that if and when the deal closes, Skybox will first help with updating map imagery, and possibly later Internet access and disaster relief.

To better explain what’s going on, we talked to Skybox investors, former employees and competitors, and watched a video where the founders explain themselves. Here’s what we learned:

The five-year-old company, which was founded by four Stanford grad students, has successfully launched a single satellite, and it already has paying customers. The next satellite is slated to launch later this summer. So no, there are not thousands of Skybox satellites with one of them poised precisely over your head at this very moment, in case you were worried.

Its first satellite, SkySat-1, was sent up in a Russian rocket in November 2013. The satellites cost about $2 million to $5 million to build, said a source. Larger traditional satellites can run more than $1 billion each.

How detailed are the images? The company abides by earlier government restrictions that don’t allow public access to satellite imagery that captures details smaller than 50 centimeters on Earth, said Ethan Kurzweil, a partner at Bessemer Venture Partners and a board observer for Skybox. That means anything smaller than roughly a foot and a half is an indecipherable blur. (Though those rules have been loosened — just today.)

“You can’t actually see anything anyone is doing,” Kurzweil said. “You can see a car, but you can’t make out a person. I don’t think you can count people, but you can count cars. So I think the privacy implications are not so severe as people might otherwise think.”

Still, with ever-improving camera resolution, slackening government restrictions and Google’s mixed record on privacy, the capability of these satellites is sure to be a subject of ongoing concern.

Skybox may be able to analyze retail trends by counting cars in parking lots over time.

Skybox Skybox says its imagery could help customers analyze retail trends by counting the number of cars in parking lots over time.

Satellite-building isn’t traditionally the realm of tech startups, and Skybox had to fight to convince investors and potential customers, its founders recalled on a Stanford panel. Co-founder John Fenwick, a former U.S. Air Force officer who leads flight programs, recalled one of the wackier proof-of-concept stories out there. “Finally we said screw it, and went and bought a Lear jet, and drilled a hole in the bottom, and put a camera to fully replicate what we were going to see from space. That helped, actually quite a bit.”

Satellite imaging companies like DigitalGlobe and AirBus have long been launching expensive spacecraft with powerful cameras into high orbits. But the high costs mean expensive images that can only be updated every so often, when the slowly orbiting satellite hovers above the spot again. That has largely limited the market to big institutions with deep pockets, including the military and, at least until now, Google.


“This is exactly what Google does: obviate the data middleman. And now they’re doing it in space.”

 

Tyler Bell, VP of product at the location data company Factual


The critical leap with Skybox is that the startup has developed an inexpensive satellite that flies at a lower orbit.

“The only thing incumbents could do was launch these massive satellites that cost 10 years to make. By the time you launch, you’re three cycles behind,” said Kurzweil. “By taking off-the-shelf parts [and] putting them together in a rapid design and test methodology, you shorten up those cycles.”

It also means the company will eventually be able to launch a fleet of them, driving down prices, updating the frequency of images and opening up a host of new commercial applications.

The company has current plans to build and launch 15 more satellites, effectively creating a constellation that can surround and move in sync around the globe, said Promod Haque, senior managing partner of Skybox investor Norwest Venture Partners and a Skybox board observer.

Skybox’s big goal and milestone would be to visit everywhere in the world once a day, Dan Berkenstock, co-founder and chief product officer, said at Stanford.

“The challenge for us in making this big idea a reality is that we’re trying to put a lot of things together,” Fenwick added. “We’re designing and manufacturing spacecraft. We’re launching and flying them. We’re putting together a huge data infrastructure. We’re doing crazy image processing algorithms. And a UI to be able to have folks play with the data. So it’s really more like five or six startups in one.”

Skybox's SkySat-1 before it was launched into orbit on a Russian rocket.

Skybox Skybox’s SkySat-1 before it was launched into orbit on a Russian rocket.

Sure, Skybox imagery might be more of a snapshot than a more detailed picture with features like multispectral resolution, like what DigitalGlobe would offer — but the point of small satellites is to be much more current. Skybox aims to take pictures with greater frequency than rivals.

Regularly revising Google Maps with fresh imagery could lead to “a whole bunch of new applications,” said Haque.

That includes crop monitoring, weather analysis, carbon credit accounting, disaster response, search and rescue and many other use cases, all of which can be baked into mapping applications.

Which brings us back to Google.

“This is exactly what Google does: Obviate the data middleman. And now they’re doing it in space,” said Tyler Bell, VP of product at the location data company Factual.

“This is a company that got the hangar from NASA, now it buys a drone company and a satellite company, and it has the best maps in the world,” said Eric Gundersen, CEO of Mapbox, which develops software that makes use of mapping data. “It is really a fascinating play.”

The fact that Google and Skybox are such a natural fit for each other may have hurt the startup, said one person familiar with the company. Earlier rumors had floated the deal at $1 billion, but that price clearly didn’t happen. “If there are no other buyers, there’s no leverage. Only Google makes sense. It was always the only buyer for Skybox since day one,” said this person.


Related Article

Gundersen, whose company competes with Google’s paid mapping developer products, said that Google is trying to dominate the market for satellite imagery, and all the data and services that can be extracted from it or layered on top of it.

“Google needs the maps to unite all their services,” he said. “This is about keeping everyone in their ecosystem.”

Skybox stands to benefit, too. Google’s established strength in mapping, software, cloud processing power and vast financial resources could turn Skybox’s imagery into even more powerful products. It has lots of smarts around things like entity recognition that Skybox could make use of to make sense of all that imagery. And more money most likely means more satellites sooner.

Before Google swept in, Skybox had about 150 employees. It was working with Space Systems/Loral to build additional satellites that use its designs.

Skybox isn’t the only fresh and exciting satellite startup. Competitor Planet Labs is also pursuing satellite imagery, but its approach is a little different. Planet Labs is putting up smaller satellites with lower-resolution cameras, and many more of them.

In February, the company deployed 28 satellites from the International Space Station. It said in March it planned to launch 100 more over the next 12 months.

At least one neutral observer (i.e. not an investor) said the difference really shows in the respective image quality from each company, with Skybox offering superior resolution.

Planet Labs satellites are the size of a shoebox, while Skybox’s are one cubic meter, or about the size of a mini fridge. With more size comes more capability, but also less flexibility. And bigger payloads are a bigger deal to send into space.

Skybox also stands out from other satellite imagery companies by offering video — which can show things like ocean wave dimensions and traffic patterns that wouldn’t come through in a still image.

Competitive distinctions aside, Google’s planned deal is a positive sign for the space in general, said Scott Nolan, a partner at Founders Fund, which has invested in Planet Labs.

“It’s really validating that there’s this space renaissance under way right now on commercial satellites,” he said. “Not only are these VC-backed companies viable in their own right, but large companies are seeing it’s really valuable for their businesses.”




1 comments
@sardire
@sardire

 terrific explanation of the intent and deliverables