[Note: I started this blog post in June of 2023, and was hoping I would never have to finish it. But given the recent decision to return the CFT crew members on Dragon Crew-9 due to technical issues with its thrusters, I figured it was worth finishing this post and submitting it.
Also, in order to increase my odds of finishing this post, I’m going to split it up into three parts — the first one on background and why competition matters, the second one on why almost everyone would be better off if NASA pulled the plug and recompeted, the COTS precedent for doing so, and whether we’re too late, and the final on some other considerations.]
In spite of requesting and winning nearly double the commercial crew development funding that SpaceX received, Boeing’s CST-100 Starliner capsule is running years behind schedule, and has been plagued with a series of technical problems. Even if Boeing could get Starliner into operational service, it might actually be better for NASA, the commercial space industry, and even Boeing, for NASA to cancel it now and use the remaining funding to find a replacement vendor.
How We Got Here
Starting nearly 20 years ago, right at the very start of the execution of President George W. Bush’s Vision for Space Exploration, NASA realized that the only way it could afford human spaceflight beyond LEO was if it could offload LEO crew and cargo logistics to commercial providers1. NASA started this process with commercial cargo delivery through the Commercial Orbital Transportation Services program in 2006, but the intention was always to leverage cargo solutions to move into commercial crew solutions2. After a few initial phases with a larger number of providers, in 2014 NASA downselected to two CCtCap3 developers — Boeing and SpaceX. Boeing’s contract was worth up to $4.2B and SpaceX’s contract was worth up to $2.6B4.
Having a traditional aerospace contractor as one of the winners may have been important politically to NASA’s fight to keep the effort funded5. However, the benefit of having a politically-connected traditional aerospace prime as a competitor didn’t come without its costs. During this time Boeing lobbied Congress pretty successfully to drive a “safety first” approach to Commercial Crew development6, that I think they felt would play better to their strengths as a large company that was used to dealing with complex regulatory bureaucracies. They used this “safety first” approach, as an argument to try to get Congress to cancel funding for SpaceX’s development, and downselect to just Boeing7. I’m really glad they weren’t successful with that lobbying!
After that though, Boeing started having a long string of problems with space and aviation programs. For instance, their DARPA ALASA8 and XS-1 9programs both cratered. SLS was running years behind schedule10. On the commercial aircraft side, Boeing’s 737 Max program ended up being grounded after two Max aircraft crashed, killing everyone on board. On the Commercial Crew front, after years of delays, a software glitch on Starliner’s first Orbital Flight Test prevented it from even attempting to dock with ISS11. NASA provided some funding for a follow-on demo12 that flew in 2022, and did dock with the ISS, but it also had some serious issues, though not bad enough to prevent docking with the ISS this time13. After these uncrewed demos, Boeing still needed to fly a Crew Flight Test with two NASA astronauts in 2023, but this third demo mission was delayed for almost a whole year due to additional issues discovered with parachutes and wiring harnesses14. When it did finally fly in June of 202415, problems with the thrusters and seals almost precluded them from docking with the ISS, and finally led to NASA announcing on 24 Aug 2024 that the two astronauts who flew up on the CFT Starliner would be coming down on a SpaceX Dragon flight in Feb 2025, and the Starliner would be returned uncrewed. And that because the Starliner hasn’t met all of its test objectives, that it would not be considered certified for crew launches yet16, potentially requiring a fourth test flight sometime in the future.
Though I think that at this point, it might be best for everyone, including Boeing, if either Boeing or NASA decided to pull the plug on Starliner, and NASA starts funding someone else to develop a replacement.
Why Competition Is So Critical
Before getting into why I think that’s the case, I’d like to first point out why it’s so critical for NASA and the larger commercial space market that we get at least one viable, economically competitive17 alternative to Dragon for crew launch.
While you can get redundancy or “assured access” from even an economically uncompetitive alternative, like Starliner, without the alternative being economically competitive, you miss out on the continuous economic improvement that can come from true competition. Real competition tends to force competitors to lower prices to maintain market share. Without competition, there’s rarely incentive for a monopoly provider to pass cost savings on to customers18. In an economically competitive market, you have to invest in ongoing future cost and/or performance improvements. If you don’t, and your competitors do, they can lower their prices or offer an improved product/service and take more of the market from you.
Why does this matter for commercial crew deliveries? Because right now NASA is trying to transition from a NASA owned and operated space station to being able to purchase time on Commercial LEO Destinations (CLD), as one customer among many. If they can succeed at doing this, they can potentially shift from spending ~$4B/yr on ISS related expenses to maybe $1B/yr in CLD service orders, freeing up almost $3B/yr for beyond LEO human exploration. But they can only do this if the CLD developers are commercially successful. And the problem is that at the current $50-80M/seat crew delivery prices, and $50-75k/kg round trip cargo prices, the economics of commercial LEO destinations are challenging. I think the crew/cargo logistics prices are just barely low enough where there’s a fighting chance of being able to close the case for one or maybe two CLD facilities, but one of the best ways to improve the economic prospects of Commercial LEO Destinations would be to drive down the cost of crew and cargo logistics. And the only way that’s going to happen is if SpaceX has competitors who are price competitive. And that is almost certainly not going to happen soon in a world where Starliner is the only funded competitor to SpaceX.
Finally, even if you don’t believe in the ability of real competition to drive down prices over time, it’s hard to assure access if only one of the providers is economically viable19. One of the real risks NASA has is that with Boeing losing so much money with Starliner, and being so expensive that they’re unlikely to win commercial crew launch orders competing against SpaceX, is that they’re going to eventually decide to pull the plug on Starliner. Even if they were to fly out their existing contract, if they back out after that, NASA is still stuck in a world where they want to keep flying to commercial LEO destinations, but there is US crew launch monopoly. Does anyone seriously think Boeing is likely to keep flying Starliner post ISS? Additionally, if you have a provider that isn’t competitive, they’re likely to fly far less frequently, which strongly correlates with lower system reliability. What happens if a future Starliner has an issue requiring their crew to be returned on another vehicle? If you have a dead-end program, that has no future after the existing flights, are you really going to be able to retain the best people for that program? For true dissimilar redundancy, you need a second provider that has a realistic chance of winning commercial business, and has a real hope for long-term existence. Otherwise, you’re just fooling yourself about how much redundancy you truly have.
NASA doesn’t have to deal with a self-imposed US crew launch monopoly, or the risks of having a backup plan that’s being kept alive on life-support. But fixing that would require taking the bold action of pulling the plug on Starliner now, and redirecting what funding would’ve been left for Starliner flights20 to funding a recompeted replacement.
In my next post, I’ll go into why I think almost everyone would be better off if NASA did this, the valuable precedent NASA set by doing something similar in the COTS program, and whether or not it’s too late to do so.
Jonathan Goff
Latest posts by Jonathan Goff (see all)
- RLV Markets VII: Four Models for Commercializing Reusable Launch - July 25, 2026
- Happenings Elsewhere In the Goffosphere - June 6, 2026
- Starliner Reponendum Est? (Part II): Why Boeing’s Troubled Capsule Still Needs to be Replaced - March 4, 2026
- See for instance Chris Shank’s comments at the 2005 Return to the Moon Conference: “[I]f we didn’t take a firm fixed-price approach towards our acquisition practices on how we’re going to provide ISS crew and cargo, we could not afford to move on to the Moon.” Also, the original VSE document stated that NASA should “Pursue commercial opportunities for providing transportation and other services supporting the International Space Station and exploration missions beyond low Earth orbit.”
- All COTS proposers were invited to submit, as part of their COTS proposal, a priced option for evolving their vehicle into a crew version under “Capability D”. SpaceX had initially proposed a Capability D solution for $300M and 2yrs of duration. Though it should be noted that this was assuming an approach with dramatically less NASA involvement than the eventual CCtCap approach. I’m not totally convinced that NASA’s selected approach was better.
- see the Commercial Crew Transportation Capability section of the Wikipedia page on the Commercial Crew Development program
- Each contract was guaranteed at least two crew missions, and up to six, hence why the awards were given as the max contract value.
- During the mid-2020s, there was an ongoing running battle between NASA and Congress over SLS/Orion and Commercial Crew, with the President’s budget request typically asking for less money for SLS/Orion, and more for Commercial Crew, but with Congress then giving more money to SLS/Orion, and slashing the budget for commercial crew
- I think calling it “safety theater first” is probably more accurate.
- While that was pretty obvious to anyone following space policy during the 2010s, I heard from a DC friend of mine that at the peak of their lobbying hubris, Boeing was trying to also get Congress to cancel Orion, and basically switch over to using a Starliner derivative for lunar missions. Basically they wanted to get NASA to sole-source human spaceflight to Boeing. Given how successful they had been with getting Congress to force NASA to develop SLS and give Boeing the contract without even requiring competition, I can see why they thought they had a shot at pulling it off…
- A DARPA program to develop an air-launched rocket using mixed monopropellants that Boeing primed.
- Another DARPA program, this one to develop a semi-reusable smallsat/hypersonics test launcher, that Boeing won the downselect on, and also ran into the ground.
- Ultimately taking over twice the amount of time and budget as was originally claimed by Congress in their 2010 NASA Authorization discussions.
- See the Wikipedia entry on Boeing’s OFT-1 mission.
- I’m not positive NASA provided funding, I thought I heard this was the case, but may have misremembered. I’ll update this if it turns out I was mistaken.
- See the Wikipedia entry on Boeing’s OFT-2 mission.
- See the Wikipedia entry on Boeing’s Crewed Flight Test.
- A year after I started writing this post that I was really hoping I wouldn’t have to finish.
- See this article from NASA’s Commercial Crew Program
- By economically competitive, I mean a crew launch option that could either come close to meeting or even beating SpaceX’s current crew delivery ticket price per seat. Note that this doesn’t require them to beat SpaceX’s internal costs, just the price they feel the market is willing to bear without any real compeitors.
- I know many people talk about how much reuse has allowed SpaceX to drive down costs, but if you look at the price a customer has to pay for a Falcon 9, it actually hasn’t gone down at all in the past 10 years. Even on a $/kg basis, it hasn’t really dropped by very much. See the table in the Historical Falcon 9 prices thread on Reddit for example, amended by the recent price increase to $67M for a reusable Falcon 9. $67M for a 15mT to LEO reusable Falcon 9 (its payload to an ISS like orbit) is ~$4.5k/kg, which is almost exactly where they started in 2010. With Transporter, they’ve lowered the entry cost for getting a rideshare launch, and that makes a difference to many customers, but most of the cost savings from reuse haven’t shown up in the form of price savings for customers yet, and likely won’t until SpaceX has a competitor that can undercut SpaceX’s current launch prices.
- This was a point Scott Pace brought up when I was discussing my idea for this article with him a year ago at the Matt Isakowitz Fellowship Program dinner
- Likely minus the cost of replacing those planned Starliner flights with additional Dragon flights
There’s a lot of guessing on F9 cost for SpaceX. Anywhere from $15 to $35 million. I think the $35 million was pretty much a wild guess to fill in a blank. Including all recovery, normal maintenance and recurring launch costs, $25 million seems in the ballpark.
In any case, the fact that they could probably cut today’s customer case in half and still make money makes real competition hard to see.
And that’s with today’s situation. If Shipzilla works out, everything is out the window.
I have to wonder if this new trunk they’re building to splash the Station will be the basis of something more.
I have this sense of HUH,WHAT about the whole thing. That sense is mainly because I don’t have a feel for why the capsule should be a decade and billions in the making and still not get there. I vaguely remember being more annoyed by the SpaceX $2.6B than the Boeing $4.2B believing that it was a fairly straightforward upgrade to the cargo capsules already developed. So much for my beliefs.
I have similar mental disconnects on people saying the vertical recovery on Falcon competitors is too difficult. That disconnect is fueled by the regular landings by Masten, Armadillo, and Unreasonable a couple of decades back. (at several orders of magnitude less money) Mercury, Gemini, and Apollo were all done before most people reading this were born and I foolishly thought it was a fairly straightforward task for both companies.
It seems a fundamental problem with Starliner is it takes to long to build one. Sort of like SLS, taking 2 years to have the rocket ready to launch.
In a sense, NASA has two other launch providers to replace SLS. The plan is to keep SLS, but at some point, maybe coming within a year or two, NASA can decide it’s not needed to land crew on the Moon.
So, maybe keep Starliner, but press for faster construction of Starliner, and open up for 1 or 2 others companies to get a contract to deliver crew to ISS.
Or Starliner is not a back up for Dragon Crew, as it’s ready fast enough. Or it’s possible Falcon-9 could be grounded for more than couple weeks. Of course perhaps SpaceX might offer Starship as a crew delivery to ISS, so perhaps take Starship and also see if can get another rocket company can deliver crew to ISS.
$4.5k/kg being exactly where they started in 2010 doesn’t seem entirely fair since due to inflation – $4500 then is about $6500 today according to this calculator
https://data.bls.gov/cgi-bin/cpicalc.pl?cost1=4500&year1=201007&year2=202407
So adjusted for inflation that’s about a 30% decrease in cost per kg.
Competition is important for more than cost. Since SpaceX is a private company, it’s anybody’s guess if they turn a profit or desperately need all of the money they bring in a few times a year with rounds of outside funding to keep the doors open. Should they suddenly be required to pass an audit by a large creditor, the government could be on a big hook to bail them out for unknown billions to make sure there is some way to keep rotating crews through ISS until it’s decommissioned. It’s either that or to go hat in hand to V. Putin to buy more seats on Soyuz at whatever price he thinks they can get.
The SpaceX Starship is a long way from being an operational system and how they could get it rated for launching with a crew (a NASA crew particularly) is still unknown. An abort system being important for launching from Earth. Until there is infrastructure on the moon that can be used by a crew that’s aborted from a lunar ascent, a Starship/HLS can be without a way to abort once the motor(s) is lit and they’ve lifted off the surface.
Things don’t look good for Starliner. A comprehensive redesign of a major onboard system would mean they wouldn’t have many years of servicing ISS to earn a return on the investment. I can see how cancelling out the contract might be the only way forward and just risk having a single key supplier with an expensive and politically sensitive back up in case of dire circumstances.
Masten, Armadillo and Unreasonable all went with simple, straight forward designs. There’s also the benefit of not needing the designs to take into account operations outside of the atmosphere initially while getting software sorted and hardware evolved. With a limited budget, never enough people and other constraints, feature bloat didn’t have as much of a chance to creep in. Not that there weren’t plenty of ideas for later.
It should have been easy for Starliner to be something of an iteration of Apollo hardware with materials and electronics updates. Just like when I have a photography job, I first have to execute on the brief in the contract before I can get distracted with targets of opportunity, change orders and other last minute extras. If I don’t deliver everything on the list, it might be hard to get paid. When I’m doing product development for a client, it’s the same thing. The first job is to deliver a working prototype and hold back on “improvements”. Better is the enemy of good enough.
Andrew,
A lot of common price calculations have to be discarded when you are building a highly technical piece of equipment that isn’t going to sell many copies before it needs to be replaced or seriously updated. You can sell your product/service for under the entire cost including ongoing R&D and rely on investors to keep funneling money in, but the music stops at some point.
Even if Shipzilla does work out, there would remain a big question about there ever being any ROI from it. Criticism rains down on SLS for the costs, but Starship is digging a truly gigantic hole with their development program.
Jon,
Competition is good for price, but with something like this, a difference in approach to the hardware means redundancy. If there’s a serious mishap with Dragon and let’s also say they figure out what went wrong, SpaceX might have to scrap any work in progress and after coming up with a suitable fix, start production all over. Granted, it could just be software, but it still means a grounded system with a grinning Putin moving the cost per seat sharply upwards until the fix is certified and there are craft to fly again.
For commercial LEO destinations, there has to be enough interest and reserve capital to maintain the facility. Who would have thought Boeing would be in the position they are now so even a very large enterprise might have to pull back from high cost activities while a station in orbit has recurring costs that won’t change with the level of usage. It would be preferred that there is a good size waiting list so even a 2008 level of financial upheaval doesn’t dump the cost of managing the assets into the lap of the US government just to make sure the thing stays in its orbit.
Ideally, a government contractor that is working on a R&D contract should be required to hand over all of the IP and associated tangible assets if they don’t meet the terms or decide to leave the business. This would let the government put the contract out to bid again, but with a good head start. There aren’t too many companies that have the institutional background and facilities for this sort of work. With a boost up, it could mean a new company/division can get into the business. I see these failures as firmly on the shoulders of management rather than the engineers and technicians. Working on rockets is a pretty cool job so there’s lots of enthusiasm to do it well at the lower levels of the org chart.