Think of Earth observation from space and what comes to mind is probably spy satellites: big and exotic spacecraft, national security, vast and barely accountable sums of government money. Or maybe weather satellites: big, also very expensive, public service, government money.

But as anyone who has looked at Google Earth knows well, it is hugely interesting, even transformative, to see pictures of your house or factory, say, from space. It is not much of a leap to imagine the practical and commercial value of seeing images that are more recent or in higher resolution than those available free on Google. Not surprisingly, there is a growing band of public service agencies and private enterprises beginning to offer services and products based on Earth imaging data. Monitoring crop health, mapping the urban sprawl and searching for oil and gas are just a few nascent applications.

Also not surprisingly, two great technology trends are enabling this civil Earth observation industry. One is the increasing capability and ever-lower cost of modern electronics and data storage. The other is the rise of artificial intelligence and machine learning; a growing number of satellites providing data in multiple spectra make Earth observation (EO) a big data exercise.

But unlike many other realms where computing power is transforming our 21st century lives, satellite Earth observation is ultimately tied to what feels like a very 20th century consideration. The optical, infrared and radar sensors that provide all of the data have to be in orbit. To be in orbit is to have been launched. And, to quote Robbie Schingler of the San Francisco-headquartered Earth information provider Planet, “Launch sucks – it doesn’t work very well.”

Schingler, speaking in May at a conference examining the “new paradigm” in Earth observation organised by the European Space Agency (ESA) at its ESRIN EO centre in Frascati, near Rome, was talking about cost when he added: “Launch is the biggest barrier to innovation.”

As for spacecraft, that innovation has been dramatic. The breakthrough concept dates to the late 1970s, when researchers at the University of Surrey, led by Martin Sweeting, thought they could slash the cost and timescale of building a satellite by eschewing custom-made electronics for commercial off-the-shelf components. Launched in 1981, their experimental UoSat-1 weighed 54kg (119lb) and survived in low-Earth orbit for eight years. Sweeting – now Sir Martin – went on to found Surrey Satellite Technology (SSTL), which was bought by Airbus in 2009. SSTL spacecraft about the size of a home washing machine and weighing less than 300kg have brought affordable civil EO benefits to such unlikely clients as the government of Nigeria.

In Frascati, Schingler praised SSTL and others for bringing satellites down to a size and mass that makes it possible to fly them as secondary payloads on big rockets. Planet, which he co-founded in 2010, is taking that principle to an extreme. Its business is to sell EO data which it gathers itself, and on its website the company claims to be “designing, building and launching satellites faster than any company or government in history”. That claim is reasonable; when the Indian space agency’s PSLV launcher orbited a record 104 satellites – one Earth imaging primary payload and 103 secondary cubesats – from a single launch in February, 88 of those cubesats were a “flock” of Planet’s 5kg Dove spacecraft. Doves, says the website, act together “like a line-scanner for the planet”.

However, he told the Frascati gathering, it was not clear that this secondary payloads trend would be truly transformative. We were not, he said, doing enough with “nanolaunchers”.

Perhaps not. But however we want to define “nano”, there are some projects in small launchers which hold out the promise of being at least partly transformative – if not in cost then at least in giving small satellite makers the control over their timetables that comes with being a primary payload.


Last month, indeed, one of them flew for the first time. Rocket Lab, a Los Angeles-New Zealand venture that has been working on its Electron vertical launcher since 2014, chalked up a partial success. Details remain to be revealed as flight data is still being analysed, but the 17m tall, two-stage rocket launched successfully from its Mahia, New Zealand launch complex and “reached space” but failed to reach orbit, implying a maximum altitude of 100km or more. There was no payload on this test flight, but the company bills Electron as designed for a nominal payload of 150kg, maximum 225kg, to a 500km Sun-synchronous orbit – designed, essentially, for small EO or scientific missions.

Rocket Lab founder and chief executive Peter Beck says: “We’re one of a few companies to ever develop a rocket from scratch and we did it in under four years. We’ve worked tirelessly to get to this point. We’ve developed everything in house, built the world’s first private orbital launch range and we’ve done it with a small team.

“It was a great flight. We had a great first-stage burn, stage separation, second-stage ignition and fairing separation. We didn’t quite reach orbit and we’ll be investigating why. However, reaching space in our first test puts us in an incredibly strong position to accelerate the commercial phase of our programme, deliver our customers to orbit and make space open for business.”

A second launcher is in the factory and could launch “in the coming months”; a third test is also scheduled for this year.

Rocket Lab is nothing if not technically bold. Its Rutherford engine is a “state-of-the-art oxygen and kerosene pump-fed engine specifically designed from scratch in New Zealand for Electron, using an entirely new propulsion cycle. Its unique high-performance electric propellant pumps reduce mass and replace hardware with software.”

Rutherford is “the first engine of its kind to use 3D printing for all primary components”. Critically, “the production-focused design allows Electron launch vehicles to be built and satellites launched at an unprecedented frequency”. Unprecedented, indeed; Rocket Lab expects “at full production… to launch more than 50 times a year”. It adds that it is regulated to launch up to 120 times a year, compared with “22 launches last year from the United States, and 82 internationally”.

Electron has already signed customers, too – including NASA and Planet.

Electron first launch c Rocket Lab

Electron ready for launch, Mahia, New Zealand

Rocket Lab


Taking a different approach but also bearing down on a first flight towards the end of this year or in early 2018 is Virgin Galactic, through its new Virgin Orbit subsidiary. Work continues at its Long Beach, California facility on its LauncherOne air-launched rocket, to be carried aloft by a modified Boeing 747. LauncherOne – with its in-house-designed liquid oxygen-kerosene engine called Newton – is being designed to put satellites of up 300kg into Sun-synchronous polar orbits, or of up to 400-450kg into equatorial orbits, and to do it for $10 million per flight.

Virgin Galactic is contracted to 39 flights – and 100 options – for UK-based OneWeb, which plans to orbit 900 Airbus Defence & Space-built microsatellites of less than 150kg each, to provide affordable broadband internet to rural areas around the world from 2019. The bulk of OneWeb’s lifting will be done by heavy Soyuz rockets from ESA’s spaceport in Kourou, French Guiana; Virgin’s appeal to the project is its proposed ability to make quick, tactical launches to fill gaps in the constellation left by, say, an in-orbit failure. The same sort of on-demand service could well appeal to a company like Planet.

Virgin’s setting of a $10 million price tag looks significant. Rocket Lab is not yet indicating its target price, but to match that cost per kilogram of about $25,000 it would have to price a flight at about $5 million.

Working to a similar cost matrix is a UK start-up called Orbital Access. Based at Prestwick airport near Glasgow, Orbital is entering the design phase of a programme that hopes to air-launch, from a McDonnell Douglas DC-10 or MD-11, a reusable rocket capable of putting up to 500kg into a 600km Sun-synchronous orbit. As founder Stuart McIntyre describes it, the “sweet spot” mission would be to fly three 150kg/1m3 satellites – units in the OneWeb category. By going a bit larger than Virgin Galactic (and also the venerable Pegasus air-launched system from Orbital ATK in the USA), McIntyre plans to be capable of a wider range of missions.


As a commercial venture that can recover its costs and return a profit, McIntyre stresses the need for a high flight rate and economies of scale. So far, about £500,000 ($643,000) has been spent developing the concept, half from private investors and half in matching funds from the UK Space Agency. Now, McIntyre is raising funds for a €2.5 million ($2.8 million) preliminary design phase, after which he reckons he will need €50 million for design; ultimately the project could cost “an easy $500 million”.

Commercially, then, the concept is built around reusability and a business that could be deployed to any suitable runway take-off spaceport. Prestwick airport is an ideal starting point, he says, because 80 years of local aerospace manufacture has created a skills base; the UK government is also keen to encourage development of a spaceport for air launches, and Prestwick is but a short flight from open skies above the Atlantic.

The rocket is to be built from off-the-shelf motors and, rather than being stacked with the payload at the top, will feature a central payload “cartridge”, with avionics and other systems in the nose cone. After release from the carrier aircraft, the rocket will fly to 90,000ft and Mach 8 or 9, when the cartridge will open to release the payload with orbital motor. Then, the cartridge closes and the vehicle is recovered following fly-back.

The business plan also calls for offering in-atmosphere microgravity services via parabolic flight paths. McIntyre hopes to start those services as early as late 2017 in turboprop aircraft and, maybe in late 2019 or 2020, in either a Dornier 328Jet or Avro RJ (with some connection to BAE Systems via the Prestwick location).

McIntyre’s launch cost target is $30,000/kg, implying a per-flight ticket of $15 million.


None of these projects, of course, has demonstrated anything approaching the reliability of established, big, agency rockets. Take ESA’s Vega small launcher, for example. In March, it orbited the European Commission’s Sentinel-2B EO mission – 1,130kg to 786km – and it put it in a near-perfect circular orbit, continuing a perfect reliability record over nine flights. The cost was €35 million, which works out to about $35,000/kg; roughly the same as its main rival for these EO or scientific missions, India’s PSLV. With a smaller main payload, either launcher has plenty of room for secondary rides, as PSLV demonstrated in February.

Moreover, it is not clear yet that any of the new generation of “nanolaunchers” will match a vehicle like Vega in being able to deploy, accurately, the multiple spacecraft that keep prices down by flying with a full load. In short, while the new small launchers may be useful for some missions, none of them promises to revolutionise access to space, because their cost per kilogram is essentially the same as the big launchers.

In Frascati, SSTL’s director of EO and science, Luis Gomes, gave FlightGlobal the example of the in-development Carbonite satellite platform. SSTL’s goal is 50kg in orbit for less than $5 million; such a satellite is not as capable as a big “agency” spacecraft, but will do much of the job for a fraction of the cost.

Unfortunately, he says, at current costs per kilogram, the implied $1.5 million launch bill still threatens to undermine the economics for many customers, who would otherwise benefit from a mission.

Vega VV09 march 2017 c Stephane Corvaja/ESA/SIPA/R

Vega flight VV09, March 2017

ESA/Stephane Corvaja

Source: Flight International