Commercial space trailblazer

David W. Thompson, co-founder and former chief executive of Orbital Sciences Corp.

Positions: Board member for the Aerospace Corp., Caltech, Carnegie Institution for Science and the Hertz Foundation. 2014-2018, president and CEO of Orbital ATK, the company created from the merger of Orbital Sciences Corp. with the aerospace and defense groups of Alliant Techsystems. 2009-2010, president of AIAA; honorary fellow since 1992. 1982-2014, president, chairman and CEO of Orbital Sciences Corp. 1981-1982, special assistant to the president of Hughes Aircraft Co.’s Missile Systems Group, assessing and planning discretionary research investments for tactical missile development. 1978-1979, engineer at NASA’s Marshall Space Flight Center in Alabama, overseeing industrial studies and limited prototyping of liquid fuel engines to replace the space shuttle main engines.
Notable: Hired engineer Antonio Elias, who convinced him that satellites could be launched from a rocket released from an aircraft, which became the company’s Pegasus fleet; other products and services created under his leadership include the OrbComm communication satellites and Cygnus cargo spacecraft. Led the $9.2 billion sale of Orbital ATK to Northrop Grumman. In 2018, chaired the first independent review of NASA’s Mars Sample Return mission.
Age: Great Falls, Virginia
Residence: 70
Education: Bachelor of Science in aeronautics and astronautics, MIT, 1976; Master of Science in aeronautics, California Institute of Technology, 1978; Master of Business Administration, Harvard Business School, 1981.

Texas oil is rarely associated with space — yet without the early and persistent financial support of Texas oilmen, Dave Thompson might never have achieved his dream of establishing a space business. Likewise, without the persistence and willingness of Thompson and his co-founders to adapt in response to shifting market and customer demand, Orbital Sciences Corp. (later Orbital ATK) might never have been sold to Northrop Grumman for $9.2 billion in 2018. “I always start with revenue because if you’re running a business and you don’t have revenue, you really don’t have anything but an idea,” Thompson says. Indeed, before its acquisition, Orbital reinvented itself multiple times in 36 years, becoming a spacecraft builder, constellation operator and launch vehicle provider, among other roles. Yet Thompson says there were plenty of “crises,” especially during the early years that could have prompted the company to fold. I visited Thompson at his Virginia home to discuss those crises, his assessment of today’s commercial space market and more. 

Q: Describe your early years in the industry and how you found your focus.

A: While an undergraduate at MIT, I interned for three summers at NASA’s Langley Research Center in Hampton, Virginia. After my senior year, I had my most enjoyable summer working at NASA’s Jet Propulsion Lab in California in the summer of 1976 when the two Vikings were landing on Mars. JPL really balanced my, until then, nearly complete focus on human space. It brought in the robotic side of the equation, which I actually ended up spending probably more time on than on the human side. The mid- to late ’70s was a slow time for human spaceflight. Apollo had done its thing, the space shuttle was in development. Meanwhile, there was some exciting stuff on the robotic side, like the Voyager probes were launched. I was looking for something that was moving a little faster. In the tech world, Apple had been founded and Microsoft had gotten started. The pace was really fast and exciting, and I thought, “Why can’t we do that in a space business?” So I went to business school with the idea of eventually starting a space company. I didn’t know what sort, but I did meet two other students who thought that would be a really fun thing to do.

He’s referring to Bruce Ferguson and Scott Webster, with whom he later founded Orbital Sciences. — JC

In 1980-81, we got a little grant from NASA to do a study. The space shuttle was almost ready to go, and our charge was to help them better understand the business opportunities for materials processing in space. We knocked on a bunch of doors, and it was a lukewarm response. We reported to NASA, “Don’t count on this happening in a big way in the near term.” Here we are, what, 45 years later? It’s still not much.

Q: When did the idea for Orbital come?

A: The second half of our second year of our MBA program, while I was waiting for a delayed flight out of Boston’s Logan Airport on a snowy day. There were two transfer vehicles or upper stages either available or in development for the space shuttle: the McDonnell Douglas payload assist module and a mostly Defense Department-funded vehicle called the [Inertial] Upper Stage.

He’s referring to rockets that would be released from the payload bay of a shuttle orbiter to carry a satellite or other payload to a higher orbit. — JC

NASA planned for a third, high-end vehicle: an Atlas Centaur upper stage modified to fly in the shuttle for big satellites and high-energy missions. My idea was we’ll do it commercially, quicker and cheaper than the big guys. All NASA had to do was help us a little bit and, if we met certain milestones, agree to buy our product. Bruce Ferguson wanted us to start the company right away to implement this idea, but I didn’t think we were ready for that. So we took respectable jobs, stayed in touch and hammered out the very rough outlines of what our business might look like. We reconnected in October 1981 when our NASA report on materials processing won an award from the Space Foundation in Houston. We were taken to this very nice dinner at the River Oaks Country Club, and sitting next to me, by happenstance, was an oil man named Fred Alcorn. I was telling him about our business plan, and at the end of the dinner, he said, “If you ever get serious about doing this company, come on down and let’s talk. I might be able to provide you guys with some walking-around money,” which was seed capital. We incorporated the company a year later on April 2, 1982. Fred and one of his colleagues provided the first external financing, and in summer 1983, three venture capital firms together put in a first round of more serious money. However, it became clear somewhere between the Alcorn walking-around money and the first venture capital that our plan wasn’t going to happen. NASA was not at all keen — for understandable reasons — on having these three kids do this important element of the whole shuttle infrastructure. Instead, they wanted us to focus on this gap: The IUS does two maneuvers to get from low-Earth orbit to geosynchronous orbit. Most commercial satellites don’t want the second maneuver, so this IUS provides functionality they don’t need. So a transfer orbit stage would be our first product. Instead of a $500 million development program with a $50 or $75 million recurring cost price tag, this was going to be a $50 million development program with a $10 million or so recurring price tag. It was now up to us to raise the funds to actually carry out the development program.

Q: Let’s go over the five crises that you’ve said Orbital faced early on.

A: NASA was still planning to do the transfer orbit stage as a traditional government-funded R&D program. They said, “We’ll hold up that for a couple months because we’ve heard your story, but you have to convince us.” So the first crisis was how do we in six weeks go up in terms of financial resources by a factor of 10 and line up one or more respectable aerospace integrators to commit to what they had to do? The second crisis was a year later. We had a $50 million financing problem, because you can’t get there from where we were with traditional venture capital. Bruce Ferguson through his legal work learned about how early biotech companies like Amgen were structuring their R&D investments as limited partnerships. Nobody had ever done this for the space industry. We traveled the whole country for months and months, selling units for $50,000 increments apiece to high-net-worth investors from a whole variety of backgrounds: doctors, dentists, professional baseball players. It was structured so that limited partners — all these investors — would be able to deduct R&D expenses from their income taxes as we incurred the R&D cost. But it had to be funded in the particular tax year, which was 1983. It was pretty clear around the first of December that we were going to be short $20 million, which meant starting all over next year. So we had to convince the investment bank running this to potentially put as much as $20 million of its capital on the line if we couldn’t finish the fundraising in the early part of the new year — which we finally did. With that money, our subcontractors lined up and a little bigger staff, come early 1984, we were pretty much off to the races developing the transfer stage. By early 1986, we had the prototype designed and were about ready to go into full testing. Ford Aerospace had signed up for three launches. NASA hadn’t quite signed on the dotted line, but it was pretty clear they were going to buy a few. 

Q: The third crisis was the January 1986 Challenger tragedy?

A: It was clear from that point forward that shuttles weren’t going to be launched once a month or even going to be as routine. They weren’t going to be as inexpensive because NASA was beginning to phase out the subsidy in a pretty big way. And they generally weren’t going to be available for nongovernment customers, because the national policy became to rejuvenate the expendable vehicle business so the U.S. wouldn’t be dependent on European vehicles if the shuttle wasn’t available. It was terrible for the country and life-threatening for our little venture, but the silver lining was that it prompted us to rethink the whole strategy. Instead of plugging holes in the shuttle infrastructure, we’re going to backward integrate ourselves from the very beginning; we’re going to expand and develop a really great engineering team; we’re going to acquire facilities and build our own products. The question was: what products? A couple months after Challenger, I called one of our board members, MIT professor Jack L. Kerrebrock, and asked his recommendation for the brightest systems engineer he’d seen. He told me Antonio Elias, who is the best person I ever hired. After about six months, he came up with the idea for the Pegasus. My first reaction was, “No, that’s no good,” but we did some internal work that convinced me otherwise. We know who the customers would likely be, who the competition would be, a pretty good idea of what it would take in time and money. As for the anchor customer, DARPA said, “If you guys can actually develop this thing, we’ll buy the first six launches.” All those pieces kind of miraculously fell in place. From the day he thought of it to the first flight was about a week less than three years.

Q: What were your initial concerns?

A: Since the 1970s, there have always been multiple small launch vehicle programs; they kind of come with the blooming of flowers in the spring. But Antonio convinced me that we had some advantages. In retrospect, one was that we were more knowledgeable about what could be done and what the options for doing it would be in terms of raising private capital, because we had done it already for our transfer vehicle.

Q: The fourth crisis was financial too?

A: We’re now in the spring of 1988. The missing link is how we are going to come up with, coincidentally, another $50 million development program. We tried, again, the industrial partner route. It worked this time: Hercules Aerospace became our propulsion supplier. We needed Hercules to cover all their development costs on the rocket motors and to accept a joint venture arrangement on the Pegasus rocket. We could partly develop our own engineering and manufacturing and assembly capability. But we found this small private company in the Phoenix area that we really wanted to acquire.

He’s referring to Space Data Corp., an Arizona firm that produced suborbital sounding rockets. Its acquisition allowed Orbital to enter the sounding rocket market. — JC

We needed them to agree to be acquired, and we needed some investor to put in $20 million for our part of the development program. It all came together.

Q: Tell me about the fifth crisis, around Orbital going public.

A: In the run-up to the first Pegasus launch, I’d gotten a call from a reporter of a well-respected national newspaper.

He’s referring to Bob Davis of the Wall Street Journal. — JC

He said, “You’re doing some really cool stuff, and I would like the inside scoop.” I foolishly agreed to let him come spend some time with us. By early 1990, we were really, really short on money. We weren’t quite through the Pegasus development program, plus we were now building these transfer orbit stages. We thought it might be the time to go public, but the underwriters wanted to wait until after the first Pegasus launch. We convinced them to go ahead anyway. The January launch date became February, March and finally slipped into early April. But we were ready in late March to do the initial public offering and feeling pretty good about it. What happened was a little different: The day the stock is to start trading on the NASDAQ exchange, we are the front-page article, which is basically saying “Rockets don’t always work. And these guys, they haven’t been doing this for 30 years.”

The article was “Start-Up Firm Faces Big Risks in Launching Rocket From a Plane: Orbital Sciences Might Offer Cheap Way for Companies To Put Satellites in Orbit” — JC

The underwriter calls and says, “We think we ought to postpone. Go launch the first Pegasus, then come back.” That’s what we did. We were on fumes by the time that was over, but it worked.

Q: How did Orbital branch out after the success of Pegasus?

A: We completed the initial public offering at a little higher price than previously priced, so we now had a little more financial flexibility. It was time to start thinking about what to do with another idea that Antonio had come up with earlier: A network of 40- kilogram small satellites in LEO, forming what today would be described as a machine-to-machine data communication system. Not voice, not high bandwidth — all you really wanted to do was let something like a trucking company receive signals to track where the trucks were. We put together a little team, and it became Orbcomm — Orbital Communications. We launched the very first satellite, which was really just a spectrum test, in 1991. We launched the two more serious prototypes in 1992, then the first operational satellites in 1995. Over the following four years, we launched three dozen of these satellites all on Pegasus, in most cases stacked up inside the payload fairing like Oreo cookies. When they reached orbit, they would deploy and arrays would come out, big antennas would unfold. But by the end of the decade, we had to decide what business were we really going to be in: a hardware developer and builder or a space system operator? We concluded, like everybody else, that you could be one or the other, but you couldn’t be both. The exception was Hughes Aircraft Co. and what became DirecTV. That was and — I think for some time to come — will be the one of two outstanding successes in commercial space. The other is SpaceX. It kind of pains me to say this, but I’m very impressed with what they’ve done. Their technical work is really impressive.

Q: On paper, air launch seems to make a lot of sense for reusability, but if SpaceX is any indication, chemical rockets are the wave of the future.

A: It certainly will depend on long-term sustainable levels of demand. You can say, “Well, it takes high flight rates to achieve and sustain cost at this level,” but you need both the supply and the demand to be in sync. Right now, on the SpaceX side, it seems like anecdotally maybe two-thirds of the launches are essentially for their own account: Starlink.

SpaceX conducted 98 Falcon launches in 2023, 64 of which were for Starlink. — JC

At some point, they’re going to not want to be continually launching Starlinks if they want to flip around from being a consumer of cash to being a producer of cash. And eventually, when they’re in some stable situation, you want those satellites to last a long time so you don’t have to launch constantly, even though it’s pretty cheap to build them and pretty cheap to launch them. It’ll be interesting to see how that all plays out.

Q: What would you tell students or young professionals who want to be space entrepreneurs today?

A: In 2019, a year after we sold Orbital, I was a visiting professor at MIT and taught a course on business principles in aerospace industry. There was a subset of students that came in thinking, “Come next spring, we’re going to graduate and two or three weeks after that, we’re going to be running our own mega company.” I always start with revenue because if you’re running a business and you don’t have revenue, you really don’t have anything but an idea. Toward the end of the course, we talked about new ventures. By that time, I’ve hopefully given the students most of the fundamentals they need to do a little bit more of a clear-eyed assessment of things like new business formation.

Q: What prompted the decision to sell Orbital?

A: It was a tough one. From the perspective of customers, most employees and shareholders — the three constituencies that I was responsible for — it’s been a great success. The combination of our space business and Northrop space business has continued now five, almost six years later to grow like crazy. In 2018, the combined space businesses were about $7-ish billion dollars. Last year, they were $14 billion. So customers must be saying, “There’s something good about this we like,” because there has been general market growth, but overall space business of rockets and satellites hasn’t doubled in the last five years. Some of that’s been rising demand — some from NASA, a lot from the military.

Q: What are the big lessons you see from the International Space Station?

A: One of the really positive things that’s come out in the last decade or so has been those Commercial Cargo and Commercial Crew initiatives. It’s demonstrated to NASA and to Congress and the executive branch that if you pick the right problem and you maintain a reasonable degree of competition — not too little, not too much — and you try to get both NASA and the private company they’re working with more or less on the same wavelength, you can do some pretty good stuff. Now, will they push it too hard in the lunar or cislunar era? I don’t know. For ISS, I think the thing that’s still missing is some kind of fundamental scientific breakthrough on the station. I just don’t know that they’ve ever gotten it. This goes all the way back to when we were trying to cook up some argument for nonspace companies to fly materials processing experiments on the shuttle.

Q: When you chaired the first independent review of Mars Sample Return, the projected cost was $4.5 billion. The latest estimate is $11 billion. What are your thoughts on the program now?

A: I thought we did a pretty good job, but it was real early in the program — in terms of time or money, at the 5% complete point. The second Independent Review Board last fall had the benefit of three and a half years of actual experience. I’m quite worried about [the mission]. The cost increase has now forced particularly JPL, the overall mission architect, to think creatively about how the mission approach might be changed to keep the cost under control, at a level where it’s not doing major damage to the other science programs that are competing for the same money. One option is to do what the James Webb Space Telescope did, which is to say, “The program may take a couple of years longer, but we’re going to finish it at a flat annual cost.” That lets NASA plan for the next five years or seven years because the budget line won’t go up. I’d make the argument that Mars Sample Return is far and away the most likely space mission over the next 20 years or so that will either answer or leave us in suspense as to whether life might have emerged on another planet in the solar system. I think that’s worth 3% per year of NASA’s budget.

Q: Was space sustainability a concern for Orbital when it began launching satellites, as it is for today’s operators?

A: That’s much more recent. When we were almost 30 years ago conceiving, designing and beginning to deploy the Orbcomm network, we were building into the design the ability to satisfy the unwritten but informal understanding that any satellite being launched into LEO would have the means of deorbiting itself within 25 years. Today, the numbers [of satellites being launched] are so much bigger. The three constellations deployed in the late 1990s — Orbcomm, Iridium, Globalstar — collectively numbered 150 satellites. These days, two launches can deploy that many, and you got launches every other week. The concern is that you get one of these unintended collisions of some kind and it just propagates. I’m just not knowledgeable enough to independently tell you that it is getting enough attention, but it’s certainly a changed environment. On top of that, of course, you’ve got military action. Hopefully we won’t end up with Earth looking like a baby version of Saturn with a ring of debris.

About Jonathan Coopersmith

Jonathan is an historian of technology and former professor at Texas A&M University in College Station who has written about the failures of 20th century space commercialization. He has a doctorate in history from the University of Oxford.

Commercial space trailblazer