There’s a new launch startup in the mix called Astra, which has been operating in semi-stealth mode for the past three years, building its rockets just a stone’s throw from the heart of startup central in Alameda County, Calif. Astra’s approach isn’t exactly a secret — its founders didn’t set out to hide anything and industry observers have followed its progress — but CEO Chris Kemp says he’s not particularly bothered about flying under the radar, so to speak.
Yes, the company had a somewhat splashy mainstream public premiere via a Bloomberg Businessweek profile on Monday, but that was more by virtue of writer Ashlee Vance’s keen interest in the emerging space economy than a desire for publicity on the part of Kemp or his cohorts. In fact, the CEO admitted to me that were it not for Vance’s desire to expound on the company’s efforts and a forthcoming attempt at winning a $12 million DARPA prize for responsive rocketry, Astra would still be content to continue to operate essentially undercover.
That’s just one way in which Astra differs from other space startups, which typically issue press releases and coordinate media events around each and every launch. Kemp, a former NASA CTO, and Adam London, an aerospace engineer who previously founded rocket miniaturization startup Ventions, designed their rocket startup from the ground up in a way that’s quite different from companies like SpaceX, Blue Origin and Rocket Lab.
“I’ve never been to one of our launches,” Kemp told me, referring to two test launches that Astra flew previously, both of which technically failed shortly into their flights. “Because I don’t think the CEO, or frankly any of our employees, should be anywhere near the rocket when it launches; we should automate everything. As much as possible, let’s put the rockets where they need to launch from, which might be an island on the equator, and it might be way up north near the poles, but let’s not add cost by putting a huge spaceport with fixed fortification in a very expensive place where it’s very hard to get to.”
As mentioned, both of Astra’s 2018 test flights resulted in vehicles that broke up shortly after launch — and Kemp says that he fully expects another test launch set for February 21, which will take off from Kodiak Island in Alaska, could likely also result in failure. Success or failure with test launches is immaterial to the company’s ultimate success, and is built into its existing financial model, Kemp says. The company has accumulated more than $100 million in funding and Kemp says he’s not worried at all about the company’s prospects in that regard.
Even once Astra is past the testing phase and into regular operations, he says people shouldn’t expect its launch reliability to resemble the targets others in the industry have sought.
“I think our objective is to never be 100%, or 99% reliable,” he said. “We would be very happy if we stabilized around 95% reliable at lower costs, because at the end of the day, we’re flying disposable satellites. If you’re flying disposable satellites, if you’re never flying people, if you’re never flying anywhere near people, then it’s about what are the overall economics, the cost-adjusted economics, of getting things into space?”
“When you have satellite factories like you’re seeing Planet and OneWeb and Amazon and others build — we’re talking about a satellite factory the size of a football stadium — they don’t care if a couple of satellites don’t make it, what they care about is if all of the satellites are on the same rocket,” Kemp continued.
That’s the key to Astra’s model: defraying risk by distributing constellation satellites among multiple, much cheaper rockets instead of piling a huge amount of them on one big, expensive launch vehicle. Kemp notes that when he was working at his previous company, OpenStack, he was also advising his former NASA friend who was busy building Earth-observation satellite company Planet. Planet lost 26 satellites when an Antares rocket crashed in 2014, followed by the loss of eight more spacecraft during a failed SpaceX mission just a few months later. A conversation with Planet co-founder and CEO Will Marshall following those losses about whether he’d rather split up smaller batches of satellites across more rockets was “formative,” Kemp said.
He points out that even if a launch service is 99% reliable, if a large percentage of your satellites are on the 1% of rockets that fail, then it’s going to be a tremendous cost to swallow, especially for a startup. “Many satellite companies have failed because their satellites were aboard rockets that failed,” he added.
Astra’s business model doesn’t rely on perfect performance, just a risk tolerance on behalf of clients that fits the economics of constellations of cheap, extensive small satellites. Nor is Kemp’s startup all that interested in driving down prices through niceties like rocket reusability, since its vehicle design and manufacturing plans are all about reducing costs through economies of scale. Astra hopes to quickly ramp to production rates of hundreds of rockets per year, and aims for profitable launches at a cost to clients of about $2.5 million per launch — scaling back to roughly $1 million per launch as production increases and those economies really start to come online.
Kemp says that Astra’s rocket currently uses the same material, in roughly the same quantity, as a Cessna aircraft, but with less complexity and without the overhead associated with creating a vehicle that’s rated for transporting people. That means, according to Kemp, that it has the technology to produce a “few hundred per year for $30,000 each,” which means that any recovery attempt like the one Rocket Lab is attempting with a parachute system and helicopter mid-air recovery starts to at minimum double its costs. At that point, there’s no reason not to just manufacture another vehicle instead, which is exactly the plan.
Astra’s approach to launch is considerably different from its competitors, and the company is also different in terms of where and how it operates; Kemp notes that no other space company is building rockets in the Bay Area. Aside from its physical location, it also has other Silicon Valley roots, including daily 10-minute standups Kemp conducts with his executive team; scrum planning with cascading reporting via regular meeting; and velocity measurement, like you’d see in a software startup. The company has also raised funding like a software startup, he points out, in a very different way relative to its closest competitors.
“We very consciously built this company to run the way the companies run here in the Bay Area, and we’ve funded it the way companies are funded in the Bay Area,” Kemp said. “That isn’t how Blue Origin was funded, that isn’t how Rocket Lab was funded, not how SpaceX is funded and I think that it’s deeply ingrained in our culture that we owe our investors this launch.”
In fact, Kemp says Astra’s next launch is just one of a campaign of launches, which is why they named this rocket “One of Three.” He says there will be a two, and a three, and that this first one, “statistically speaking,” probably won’t work as intended. That’s not important, he stresses, because of the way the company has been designed from the beginning.
“I don’t care about this first launch at all; it doesn’t need to work,” he told me. “Nothing about my business plan requires this rocket to work, because if it did, I’d be back worried about funding and worried about what would they say about the company in TechCrunch. And I’m just not.”