Editor’s note: Minutes before this story was published, Facebook posted a blog conceding that it was terminating its Aquila program, closing its drone factory, and focusing on partnerships instead.
Facebook’s plans to beam high-speed Internet from enormous solar-powered drones in the stratosphere appear to be in disarray. Two key engineers behind its Aquila drones have left the company, and it recently canceled plans for a secret high-altitude flight campaign at Spaceport America, possibly because Facebook no longer has any aircraft available to deploy.
A trove of emails between Facebook and Spaceport America, obtained under New Mexico public records law and first reported by Business Insider, details the painstaking process of turning a site for rockets and space planes into a test bed for some of the largest drones in the world.
That years-long project came to an abrupt halt earlier this year, when Facebook decided not to move forward with a high-altitude flight campaign. At around the same time, Andy Cox, the British engineer Facebook hired in 2014 to develop its drones, and Martin Gomez, Facebook’s director of aeronautical platforms, left the company.
Federal Aviation Administration (FAA) records also show that Facebook has let the registration of one of its two full-size Aquila (“eagle” in Latin) drones lapse. The other was badly damaged on its maiden flight and does not appear to have flown since.
The question now is whether Facebook has abandoned its ambitious plans for solar-powered Internet drones altogether—perhaps in favor of a satellite-based system—or is simply redesigning and building a new generation of aircraft. The company would not immediately provide an answer to that question.
Facebook has been working on Aquilas for years, with a first full-scale test flight in 2016 and a second reaching nearly a kilometer in altitude on 22 May last year, from a U.S. Army facility in Arizona.
The challenges of building an aircraft designed to remain aloft for months are legion. One in particular that dogged Facebook was how to bring an aircraft the size of a 737 safely back down to Earth when it does not have wheels or landing gear.
The emails between Facebook and Spaceport America show that its drones would have to contend with stray animals, archaeological remains, protected birds, and even the wrong kind of dirt.
For much of 2016 and 2017, the social media giant quietly planned a futuristic launching/landing zone just north of the (as yet unused) Virgin Galactic terminal for Richard Branson’s suborbital spaceplanes. A specially built hangar planned for the site was envisioned as home to Facebook’s equally innovative Aquila drones, enabling the first flights at their planned operational altitude of 18 to 27 kilometers, and testing their ability to provide Internet connectivity to a 50-km radius beneath them.
Aerial view of the runway at Spaceport America Spaceway, 22 October 2010. Photo: Spaceport America
Just nine days after Aquila’s first flight in 2016, a site selection expert from Facebook’s global connectivity company internet.org asked Spaceport America officials to sign a nondisclosure agreement about building a facility there, emails show. In an effort to preserve its anonymity, Facebook would be referred to by the code name Denali in many public documents.
Problems quickly emerged. For a start, although Spaceport America has 3.6 kilometers of runway and the Aquila needs only 335 meters to take off, it must launch into the wind, from a wheeled dolly that remains on the ground. (The lack of landing gear is so the aircraft can save weight.)
Proposed launching and landing zones for Facebook's drones at Spaceport America. Image: Facebook
The geometry of the Spaceport runway would have prevented the drones from being launched when the wind was blowing from certain angles, unless extra pavement was added to the east or west sides of the runway. “It would be a major and probably expensive construction complication, but it could be done,” wrote Bill Gutman, director of aerospace operations for the New Mexico Spaceport Authority (NMSA).
Waiting for the wind to align with the runway was fine for Aquila’s initial tests but would hardly work for a sustained campaign of flights. At the Spaceport, Facebook decided to prepare its own launch and landing area instead, directly on the bare ground.
Landing is the most hazardous time for the Aquila. On its maiden voyage in 2016, Facebook’s original aircraft encountered turbulence shortly before landing. The shifting air currents deflected and then broke its right wing. In the subsequent crash, the drone sustained “substantial damage to [its] primary structure” as well as charring to one battery, according to an accident report filed with the U.S. National Transportation Safety Board. That aircraft does not appear to have flown since.
For the 2017 test in Arizona, Facebook’s second Aquila prototype took off from a runway but landed successfully on a 150-meter circle of carefully groomed sand about 15 centimeters (6 inches) deep, christened the ‘Aquila Beach.’ Kevlar pads bonded to the plane’s motors reduced damage, although the plane’s propellers can be seen hitting the ground before the aircraft itself. In New Mexico, the aircraft would be descending from much higher altitudes and Facebook engineers wanted a correspondingly larger launching/landing circle, up to 600 meters wide.
The company’s idea was to completely clear an area northwest of the main runway of mesquite bushes and rocks, then replant it with native grasses to form a smooth landing bed. But Gutman foresaw some issues with that approach. “We believe the circle misses known archaeological sites, but we’re verifying that with our consultant,” he wrote to Facebook in July 2016. “We [also] potentially need to time the surface work to avoid disturbing nesting birds.”
Christine Anderson, executive head of NMSA, hoped that Facebook was not in a hurry to begin testing. “It may take quite a while for seed to grow...like six months to a year or more, even if we water it once per week,” she wrote to Facebook. “Can you land on bare dirt if you have to?”
Facebook’s site expert, Kevin Slover, confirmed that the Aquila could land on dirt, although “we would want to do what we could to ‘fluff it up’ as much as possible by clearing, ploughing, and harrowing.”
What’s more, Gutman said, the large landing circle would need a game fence to keep out livestock and other large animals. The first fence proposed by Facebook would have run directly through the Spaceport’s fuel farm. Gutman noted that this is the planned home of a “fairly large” storage vessel containing the nitrous oxide propellant for Virgin Galactic’s SpaceShipTwo vehicle. Accommodating such restrictions could add up to $400,000 or 14 weeks to Facebook’s construction plans, Anderson wrote.
Facebook’s hangar would be a tentlike tensioned membrane structure, supplied by Sprung, the same company that built Tesla’s new Model 3 production-line building in California.
Facebook wanted to test the drones’ ability to relay Internet signals in a number of frequency bands, including the Ka band (27.5 to 29.7 gigahertz), V band (47.2 to 48.2 GHz), and some narrow K and Ka bands proposed by the International Telecommunication Union specifically for high-altitude platforms like the Aquila. A future fleet of Aquila drones would connect with each other using free-space laser optical links, centered around 1,550 nm.
Facebook is not the first tech company to express an interest in testing Internet drones at the Spaceport. In 2016, Google tested a 5G system called SkyBender there, using an optionally piloted aircraft and possibly its own high-altitude drones. Google’s parent company, Alphabet, terminated that ‘moonshot’ program in 2017.
In February, FCL Tech filed an application with the Federal Communications Commission (FCC) for an LTE-based system with a 50-km radius, probably for use with the Aquilas. Slover had previously written to the NMSA: “Our IT folks would like to [coordinate] with you to get the connectivity we’re after. We have been considering using our own equipment to provide the backbone to our hangar, but our team would like to explore your solution as well.”
However, neither FCL Tech nor Facebook sought permission from the FCC to test at the Spaceport using the drone’s E-band or other high frequency signals. Neither did Facebook make any serious progress with building its hangar or preparing the launching and landing area. Soon after the FCC application, Facebook told Spaceport America that it would not be testing there after all.
A Facebook spokesperson told Spectrum: “Beginning in 2016 we worked with Spaceport to see if their facility could be a potential test site for our connectivity efforts. Because site investments like these require a long lead time, we often pursue multiple options at once and make initial investments and preparations so we can move quickly if we end up needing to use that site. We ultimately signed a short-term lease with Spaceport and worked with them on some lightweight site prep, but we have no plans for further investment or any operations at this time."
Lacking a test site, two of its founding engineers, and probably even aircraft to fly, Facebook’s Aquila program looks to be on pause, at least.
But perhaps Facebook is only waiting for the next generation of aircraft. Way back in the summer of 2016, Bill Gutman wrote to Kevin Slover, “I believe you told us that the direction the Aquila project will be taking in the future will be for the vehicle to take off and land using a runway rather than a prepared circle.”
“You’re correct,” Slover replied. “The plan is for a typical runway take-off and landing.” If the next flock of Aquilas sports tougher underbellies or landing gear, Facebook will no longer have to worry about fluffing up dirt or disturbing their feathered brethren after all.
Mark Harris is an investigative science and technology reporter based in Seattle, with a particular interest in robotics, transportation, green technologies, and medical devices. He’s on Twitter at @meharris and email at mark(at)meharris(dot)com. Email or DM for Signal number for sensitive/encrypted messaging.