Founding and Selling A Space Company for $145M

This week’s SpaceDotBiz highlights a key figure in one of the major space acquisition events of the last decade. I interviewed Mike Cassidy, the founding CEO of Apollo Fusion. Apollo builds electric in-space propulsion systems, specifically Hall Effect Thrusters. In 2021, Apollo was acquired by Astra Space for $145M. These exit events are critical to the space ecosystem, they provide liquidity to early investors and keep the funding cycle going for the industry.

Mike has had a storied and extremely interesting career as an entrepreneur, including as a multi-time founder with multiple exit events under his belt. In addition, he’s been an entrepreneur-in-residence at Benchmark Capital and a Vice President of GoogleX, Google’s Moonshot Factory. Mike also gives back a ton to the aerospace community and is a mentor to hundreds of startup founders.

Mike and I discussed so much and the interview was so thorough that I’ve had to split it into two parts. This week’s newsletter contains part one below. Subscribe to SpaceDotBiz to make sure you get part two in your inbox next week!

Now lets dive in!

Apollo Fusion has an interesting founding story that wasn’t originally focused on spacecraft propulsion. How did the company get started?

Well, it's funny, we were actually called Apollo Fusion because we were trying to solve nuclear fusion. The short story is, it didn't work. We didn't get to nuclear fusion breakeven, or else you would've seen us win a Nobel Prize and we would've solved global warming. So we had to pivot. In fact, four out of five of the startups I’ve built have had to pivot. We were lucky in that we had somewhat randomly picked the word Apollo when we were forming the name Apollo Fusion, just because we thought Apollo was a cool name. So when we pivoted to making ion thrusters, it was serendipitous because we didn’t have to change the name. As a result, we left our name as Apollo Fusion.

What drew you ultimately to the market of selling Hall Thrusters to satellite developers?

Firstly, I was originally an aerospace engineer. So I was familiar with, at a high level, spacecraft propulsion systems. Second, our approach to building fusion was actually a very similar plasma physics approach to developing hall thrusters. So the equipment was very similar, and the team was very similar. We had a couple of PhDs in nuclear engineering, and instead of being like, “All right, we're not building nuclear fusion anymore, we’re all fired,” we instead decided “We're building something that requires highly technical plasma physics expertise.” That made the transition and pivot smoother.

The third piece was that the space industry was booming. With SpaceX, OneWeb, and Telesat, and all the other constellations, the projected numbers of satellites are obviously huge. Since Sputnik, something like 6,000 satellites have been launched. In the next three to five years, something like 20,000 satellites are potentially going to be launched. So three times as many satellites will go up over the coming few years as have been launched in the past 60 years. All of that led us to pivot towards the hall thruster product.

I know it wasn’t an easy process in the early days of getting Apollo Fusion off the ground. Can you share a little about the challenges you faced and the creative ways you were able to mature the technology?

We definitely had some challenges. For example, we tried a number of different propellants at first. One of the ones we looked closely at was Mercury. Mercury is potentially scary from an environmental hazard perspective. Although, we were arguing that the exhaust velocity of a Hall thruster is something like 10,000 meters per second. The escape velocity in orbit is something like 7,500 meters per second. So in some ways, we're ridding the planet of Mercury, but anyway that didn't fly. We never got FCC approval to do a demo with that propellant. We built an entire satellite for that demo and we never got approval to launch it. So we wasted a lot of time and money there.

Another big challenge we had was that we had a lot of customer interest, but no one wanted to be our first customer until we had flight heritage. But then it’s hard to get flight heritage if no one wants to be your first customer. It created this chicken or egg problem. Eventually, we did a great deal with Spaceflight Inc. They had a space tug called Sherpa that takes you from where your rocket drops you off, out to whatever altitude and inclination you want to reach. They were very interested in having an electric propulsion system and they were interested in us as well, but they also had the same flight heritage challenge that we had. So we found a way to work with Spaceflight where we said, “Okay we're not gonna charge you for this thruster, and you don't charge us for doing a demo flight, and we will share the results.”

They found a way to put us on one of their Sherpa missions where the Sherpa was deploying satellites for other customers. After the Sherpa deployed its three customer satellite payloads, it then fired our thruster. As a result, we were able to test our thruster without any risk to their customers. If our propulsion system had failed, they would have already deployed their customer satellites and had a successful mission. Fortunately, it all worked great. We tried to fire our propulsion system 700 times and it fired successfully all 700 times. That unlocked a whole bunch of business for us.

Other struggles we faced were with messed up supply chains due to the pandemic. The pandemic also had an impact on access to radiation testing facilities. As a result, we came up with some very creative ways of doing radiation testing and very creative ways of reducing the requirements for rad-hard parts.

That discussion around the propellant choice, it brings to mind the recent discussion around SpaceX moving to an argon thruster. Do you have any thoughts on that propellant choice and using that?

SpaceX is very, very smart, so I think they know what they're doing. From my basic understanding, it requires a lot more power to use the argon thruster. So Apollo Fusion was targeting, say, 250-kilogram type satellites that might have, well, 400, 600, 800 watts of power or something like that. Initially, SpaceX was in that power and mass range. However, SpaceX really stepped up the mass on their next-generation Starlink. I don't recall exactly, but I think it's around four or five times the mass of their first-generation satellites. I also think they have way more power now, around several thousand watts of power. When you have that amount of power, I think argon can make sense. The specs I saw looked pretty appealing in terms of the Isp and thrust levels. It just takes a lot of power. Ultimately, I'm sure they know what they're doing.

It’s a challenge in the space world to build products that can meet the needs of both commercial and DoD customers and that’s something that Apollo Fusion has been able to accomplish. Was there a particular strategy to building an organization that could work with both of those types of customers? Were you concerned about getting too focused on one at the expense of the other?

We were not worried about focusing on one or the other. The key for our strategy from the beginning was being focused on building thrusters at large scale. So we didn't want to be a propulsion company that made three thrusters per year. We wanted to make products for satellite constellations consisting of 100 or 300 satellites each. That drove our manufacturing strategy.

From the beginning, we knew that we weren’t going to build a facility that could make 300 thrusters a year. We knew we would work with a manufacturing partner that could build 300 thrusters a year. So we found a great one that had already built aerospace hardware at that kind of volume. A lot of our customers loved the fact that we were a fast-moving scrappy R&D type startup, but that our flight hardware was made by an ISO-9001 and AS9100 certified manufacturer.

That also drove our strategy for taking the time during the design process to have design revs that were focused on design-for-manufacturing. I think the rev that we eventually flew successfully was version 1.8, but I think version 1.6 was a designed-for-manufacturing rev. Nothing critical changed. The performance didn't change, none of the specs changed. But we talked to our suppliers and to our machine shops and asked “How can we make this easier for you to make?” They would tell us “Well if you change the interior diameter to be this, then this tool can get in.” We did similar things for assembly as well. That doesn't make any sense if you're only planning to build three thrusters a year, but if you're going to build 300, then it’s important.

That made it easier to win both commercial deals and government customer deals because both customers were targeting large constellations. Not only commercial, as I mentioned before, but also the government. For example, the SDA, Space Development Agency, is building a ton of satellites. That ability to produce at large scales was a good fit for our strategy.

In the summer of 2021, Apollo Fusion was acquired by Astra Space. There haven’t been a ton of M&A deals in the “New Space” world. In a previous interview, I asked Dan Berkenstock, who had founded Skybox, about being acquired by Google and he shared about how that acquisition came to be. Without going into anything you can’t discuss, how did that acquisition come to be and what is that process like?

Astra has a small launch vehicle that's bringing a payload to a low Earth orbit. It’s kind of a natural extension to say, “If we had a very high-performance ion thruster, then we could take someone's payload from LEO to a higher orbit.” For the awareness of your readers, the thrust on our system is tiny, around 25 milliNewtons, which is about the weight of a paperclip. So you can't do anything with that unless you're in space. Once you're in space, it's fine because there's nothing resisting it. You can accelerate at 0.001 meters per second squared for a hundred thousand seconds and actually have some reasonable velocity. So it's a good fit to combine a small launch vehicle with a highly efficient ion thruster.

In addition, there are of course other factors that come into play. From our perspective, it was kind of a local optimum in timing in terms of valuation. The market was white hot at the time. There was a lot of venture investment going into space, and SPACs were frequent. In early 2021, the SPACs were obviously bringing significant capital to these newly public companies that now had cash for acquisitions. In addition, we had a very good pipeline with a lot of customers lined up, provided we could have a successful demonstration flight. In addition, Astra was able to move faster than other potential acquirers and close the deal.

The way the timing worked out was also fun in that I actually got to be there when Astra rang the NASDAQ bell after going public on July 1st. We officially closed the Apollo acquisition 15 minutes later because Astra needed to close the deal to have the cash to buy us, and we wanted to have a publicly traded acquirer.

That’s it for part one of this two-part interview with Mike Cassidy. In part two, we discuss some really juicy topics, including how he tactically pivoted his startup, his time as Project Leader for Project Loon within Google X, advice he’d give to startup founders, and more. Make sure to subscribe to SpaceDotBiz so that part two hits your inbox next week!