This isn't about Starfactory, the actual factory making rockets. It's about the whole of Starbase, a goddamned town/city under construction.
That a massive construction site with hundreds of welders and framers and plumbers and concrete workers, and linemen and ditch diggers, and survey workers and architects, and smiths and truck drivers and landscapers, and janitors and carpet layers and warehouse workers, and sewer workers, masons, electricians, and heavy equipment operators (massive cranes, bulldozers, man lifts, graders, boring rigs, excavators, etc.) most of them working outdoors in the TX weather, building dozen of homes, a rec center, factories, launch facilities, office buildings, apartment complexes, not to mention the largest ever space vehicles in volume numbers, has a higher injury rate than some ULA office workers sitting behind desks or in clean rooms building and launching a couple of classic rockets a year?
Yes. When ULA is building a rocket, they're in a many years old, long completed clean room assembly line with a few machinist stations. When they're launching, they have a tiny pad staff and medium office staff. Comparing those to the construction site that is Starbase, is apples to oranges.
Maybe they can normalize for number of launches. I'd be curious if the number/rate is actually lower when adjusted for activity level.
I just skimmed the article, but it looks like the rate is based on headcount, not productivity:
> TRIR topped out at 4.27 injuries per 100 workers in 2024, when it employed an average of 2,690 workers, according to the data submitted to OSHA. Injured Starbase employees were unable to perform their normal job duties for a total of 3,558 restricted-duty days, plus 656 lost-time days where injuries made them unable to work at all.
> Starbase is classified by the U.S. government as a space vehicle manufacturing operation. The injury rate in this sector has fallen dramatically since 1994, dropping from 4.2 injuries per 100 workers to 0.7 injuries per 100 workers in 2023, according to historical data from the Bureau of Labor Statistics. (BLS calculates these rates through its annual company surveys, which asks for the same information found in OSHA’s worker injury forms.) But despite major changes in safety processes across the industry, Starbase is closer to the rates of 30 years ago.
It's not even about the rockets. This is a city being built. Hundreds of homes and apartments and a massive office building and a rec center and roads and roundabouts and launch pads and tank farms and inventory buildings, and yes, a rocket factory that makes and launches massive prototypes.
Compare Starbase not to other rocket facilities but to large-scale construction projects. Then we'll have apples to apples injury rates.
If you want to compare rocket factories, then only count injuries in the Starfactory itself, or factory+launch staff, or whatever. But comparing a city under construction with a long-completed launch facility and the offices used by the launch staff is disingenuous at best.
Smells like fanboy in here. The article mentions nothing about construction operations which makes sense since that is best contracted out. I can’t imagine Musk tolerating the hiring and layoffs of construction crews. The numbers in the article are based on SpaceX’s actual manufacturing facilities and operations. The highest injury rate seems to be in booster recovery operations with other facilities bringing the average down.
> OSHA uses a standardized safety metric called Total Recordable Incident Rate (TRIR) to measure a company’s safety record and compare it to industry peers, like Blue Origin and United Launch Alliance.
> Of the 14 OSHA inspections at SpaceX facilities over the past four years, six involved accidents and injuries at Starbase. That includes a partial finger amputation in 2021 and a crane collapse in June 2025. The latter inspection is still ongoing. Investigations by other news outlets including Reuters have uncovered hundreds of previously unreported worker injuries, including crushed limbs and one fatality.
It's not about the rockets built. It's about the city being built. It's the welders and the framers and the roofers and the concrete workers and the jackhammer guys and the linemen and the plumbers, etc. that are leading the injuries, not the guys on the line building the actual ships and boosters. Starbase is a city, not just a factory.
i worked construction for a summer (roofing) and a lot of guys were drunk or on something fun.
i imagine elon is pushing them to build much faster than normal and they'll try.. but you also can't be on anything working a fast schedule. i think zyns would be the only thing safe, adderall would leave you too dehydrated, painkillers sure but you also get clumsier ime, even advil had me feeling off.
construction is hard! i have no idea how the Japanese have such efficiency. they work fast and make it look relaxed.
> construction is hard! i have no idea how the Japanese have such efficiency. they work fast and make it look relaxed.
I have no special insight, but I would speculate that it’s a byproduct of their culture of keeping everybody on the same page and measuring twice (or thrice) before cutting. In the US, we tend to try to speed up by making speed the goal which counterintuitively slows things down due to corners cut, i’s and t’s not dotted and crossed, poor coordination, etc.
> Starbase, which plays a central role in SpaceX CEO Elon Musk’s mission to make life multi-planetary, is an outlier in the company and across the industry as a whole. Its TRIR topped out at 4.27 injuries per 100 workers in 2024, when it employed an average of 2,690 workers, according to the data submitted to OSHA. Injured Starbase employees were unable to perform their normal job duties for a total of 3,558 restricted-duty days, plus 656 lost-time days where injuries made them unable to work at all.
> Starbase ... is an outlier in the company and across the industry as a whole. Its TRIR topped out at 4.27 injuries per 100 workers in 2024, when it employed an average of 2,690 workers, according to the data submitted to OSHA. Injured Starbase employees were unable to perform their normal job duties for a total of 3,558 restricted-duty days, plus 656 lost-time days where injuries made them unable to work at all.
Looks to me like Starbase is far safer than automobile manufacturing. Let alone hospital nursing care.
So - other than "We <3 Elon Bashing", what's the point of the article? Nobody who's been inside a real factory would be surprised that busy Starbase has higher injuries rates than its "Slowly Going Nowhere, Ferociously" competition.
Yes, a higher injury rate may reflect corners being cut in the name of speed. But I'm sure the most major contribution to the numbers (besides people actually reporting their injuries) is a result of people working then anything else. Of course injuries are going to be much lower if you haven't shipped anything in a decade plus.
> But I'm sure the most major contribution to the numbers (besides people actually reporting their injuries) is a result of people working then anything else.
It's a rate per person, so no.
There's an implicit "working on rockets" in the OP's comment. ULA did 4 launches in 2024 with 2700 employees. SpaceX did 138 launches in 2024 with less than 5 times as many employees. (They had 13,000 in 2023. It's grown a lot since then, but hasn't doubled, I don't think).
ULA employees spend a lot higher percentage of their time on paperwork than SpaceX employees do. The injury rate while doing paperwork should be essentially zero.
SpaceX does 5-10 times as many launches per employee, and has an injury rate 6 times as high per employee. So the injury rate per rocket launch is comparable.
> Among the more baffling details in the report are several sections about how Elon Musk’s personal tastes appear to have affected the factory’s safety for the worse, “his preferences … were well known and led to cutting back on those standard safety signals.” Musk, apparently, really hates the color yellow. So instead of using the aforementioned hue, lane lines on the factory floor are painted in shades of gray. (Tesla denies this and sent Reveal photos of “rails and posts” painted yellow in the factory.) He also is not into having “too many signs” or the beeping sound forklifts make in reverse.
Rate per person does not counter neuroelectrons hypothesis.
If you have 100 people standing around for a year, they probably have lower injuries than the same 100 people who are using heavy machinery over the course of a year. And it's not like most corporations are efficient.
SpaceX may have less redundancy in their workforce causing the injury rate to go up since more people are working more often.
Or SpaceX is overworking people in unsafe conditions and causing way more injuries.
Either way, rate per person does not negate people working vs. Not.
It's because Starbase workers are roofers, plumbers, and linemen working in the TX weather, not just engineers or assembly line workers in air conditioned clean rooms and offices.
This isn't about Starfactory, the actual factory making rockets. It's about the whole of Starbase, a goddamned town/city under construction.
That a massive construction site with hundreds of welders and framers and plumbers and concrete workers, and linemen and ditch diggers, and survey workers and architects, and smiths and truck drivers and landscapers, and janitors and carpet layers and warehouse workers, and sewer workers, masons, electricians, and heavy equipment operators (massive cranes, bulldozers, man lifts, graders, boring rigs, excavators, etc.) most of them working outdoors in the TX weather, building dozen of homes, a rec center, factories, launch facilities, office buildings, apartment complexes, not to mention the largest ever space vehicles in volume numbers, has a higher injury rate than some ULA office workers sitting behind desks or in clean rooms building and launching a couple of classic rockets a year?
Surprise, surprise.
> some UULA office workers sitting behind desks
You believe OSHA's categorization of aerospace "manufacturing" facilities is wrong? Why?
Yes. When ULA is building a rocket, they're in a many years old, long completed clean room assembly line with a few machinist stations. When they're launching, they have a tiny pad staff and medium office staff. Comparing those to the construction site that is Starbase, is apples to oranges.
As of mid-May the population was around 500, including around 120 kids.
It voted itself a city for bureaucratic reasons, but a teeming metropolis it is not.
Plus the stats-in-question are already scaled by hours worked. (Which also renders the number of children moot, unless something illegal is going on.)
There are far more than 500 people working in Starbase. Most of them do not live there so it's kind of pointless to mention that.
BLS has the TRIR for the whole constructon industry as 2.3 per 100k, which is lower than starbase. [0]
0 - https://www.bls.gov/web/osh/table-1-industry-rates-national....
You make a good point. It should be compared to a generic construction project, where the incident rate is only roughly double the industry norm
Maybe they can normalize for number of launches. I'd be curious if the number/rate is actually lower when adjusted for activity level.
I just skimmed the article, but it looks like the rate is based on headcount, not productivity:
> TRIR topped out at 4.27 injuries per 100 workers in 2024, when it employed an average of 2,690 workers, according to the data submitted to OSHA. Injured Starbase employees were unable to perform their normal job duties for a total of 3,558 restricted-duty days, plus 656 lost-time days where injuries made them unable to work at all.
> Starbase is classified by the U.S. government as a space vehicle manufacturing operation. The injury rate in this sector has fallen dramatically since 1994, dropping from 4.2 injuries per 100 workers to 0.7 injuries per 100 workers in 2023, according to historical data from the Bureau of Labor Statistics. (BLS calculates these rates through its annual company surveys, which asks for the same information found in OSHA’s worker injury forms.) But despite major changes in safety processes across the industry, Starbase is closer to the rates of 30 years ago.
It's not even about the rockets. This is a city being built. Hundreds of homes and apartments and a massive office building and a rec center and roads and roundabouts and launch pads and tank farms and inventory buildings, and yes, a rocket factory that makes and launches massive prototypes.
Compare Starbase not to other rocket facilities but to large-scale construction projects. Then we'll have apples to apples injury rates.
If you want to compare rocket factories, then only count injuries in the Starfactory itself, or factory+launch staff, or whatever. But comparing a city under construction with a long-completed launch facility and the offices used by the launch staff is disingenuous at best.
I know SpaceX likes to make things in-house, but I really doubt that extends to residential subdivisions. That's what contractors are for.
Smells like fanboy in here. The article mentions nothing about construction operations which makes sense since that is best contracted out. I can’t imagine Musk tolerating the hiring and layoffs of construction crews. The numbers in the article are based on SpaceX’s actual manufacturing facilities and operations. The highest injury rate seems to be in booster recovery operations with other facilities bringing the average down.
> OSHA uses a standardized safety metric called Total Recordable Incident Rate (TRIR) to measure a company’s safety record and compare it to industry peers, like Blue Origin and United Launch Alliance.
> Of the 14 OSHA inspections at SpaceX facilities over the past four years, six involved accidents and injuries at Starbase. That includes a partial finger amputation in 2021 and a crane collapse in June 2025. The latter inspection is still ongoing. Investigations by other news outlets including Reuters have uncovered hundreds of previously unreported worker injuries, including crushed limbs and one fatality.
Starbase builds far more rockets than any competitor so it stands to reason the injury rate is higher.
Hard to get injured making a powerpoint presentation.
It's not about the rockets built. It's about the city being built. It's the welders and the framers and the roofers and the concrete workers and the jackhammer guys and the linemen and the plumbers, etc. that are leading the injuries, not the guys on the line building the actual ships and boosters. Starbase is a city, not just a factory.
No, because the injury RATE is the number of injuries divided by hours worked.
More rockets = more people working more hours.
Hours worked… making PowerPoints should hardly count.
Conversely, they make more rockets than any competitor, so they should be better at doing it without injury.
why would that be?
No real reason. Just pointing out that the parent argument cuts both ways.
No, that would justify a higher absolute count of injuries. If the rate is higher, it's less safe.
So SpaceX could improve their numbers by hiring a bunch of people that do nothing but make useless PowerPoints.
i worked construction for a summer (roofing) and a lot of guys were drunk or on something fun.
i imagine elon is pushing them to build much faster than normal and they'll try.. but you also can't be on anything working a fast schedule. i think zyns would be the only thing safe, adderall would leave you too dehydrated, painkillers sure but you also get clumsier ime, even advil had me feeling off.
construction is hard! i have no idea how the Japanese have such efficiency. they work fast and make it look relaxed.
> construction is hard! i have no idea how the Japanese have such efficiency. they work fast and make it look relaxed.
I have no special insight, but I would speculate that it’s a byproduct of their culture of keeping everybody on the same page and measuring twice (or thrice) before cutting. In the US, we tend to try to speed up by making speed the goal which counterintuitively slows things down due to corners cut, i’s and t’s not dotted and crossed, poor coordination, etc.
Read that as Starbucks injury rates and was intrigued.
> Starbase, which plays a central role in SpaceX CEO Elon Musk’s mission to make life multi-planetary, is an outlier in the company and across the industry as a whole. Its TRIR topped out at 4.27 injuries per 100 workers in 2024, when it employed an average of 2,690 workers, according to the data submitted to OSHA. Injured Starbase employees were unable to perform their normal job duties for a total of 3,558 restricted-duty days, plus 656 lost-time days where injuries made them unable to work at all.
Oh, "Starbase" is the name for the Boca Chica site.
Why is it a city.
> Why is it a city.
Because it was voted into being. https://www.pbs.org/newshour/politics/voters-in-texas-approv...
It has housing and a school. It's a city.
It's a bit of an off-topic, but traditional settlement hierarchy was this:
The settlement only has houses - it's a hamlet.
The settlement has a church - it's a village.
The settlement has a market - it's a town.
The settlement has a cathedral - it's a city.
Many towns, villages, hamlets, etc. also have those.
Starbase was made a city. It's not a town, village, or hamlet because it's been legally designated a city in Texas.
Yes, we agree. The parent poster seems to think city is a de facto designation for a place with housing and schools; I'm noting that it isn't.
> Starbase ... is an outlier in the company and across the industry as a whole. Its TRIR topped out at 4.27 injuries per 100 workers in 2024, when it employed an average of 2,690 workers, according to the data submitted to OSHA. Injured Starbase employees were unable to perform their normal job duties for a total of 3,558 restricted-duty days, plus 656 lost-time days where injuries made them unable to work at all.
Compare with https://www.bls.gov/web/osh/table-1-industry-rates-national....
Looks to me like Starbase is far safer than automobile manufacturing. Let alone hospital nursing care.
So - other than "We <3 Elon Bashing", what's the point of the article? Nobody who's been inside a real factory would be surprised that busy Starbase has higher injuries rates than its "Slowly Going Nowhere, Ferociously" competition.
Yes, a higher injury rate may reflect corners being cut in the name of speed. But I'm sure the most major contribution to the numbers (besides people actually reporting their injuries) is a result of people working then anything else. Of course injuries are going to be much lower if you haven't shipped anything in a decade plus.
> But I'm sure the most major contribution to the numbers (besides people actually reporting their injuries) is a result of people working then anything else. It's a rate per person, so no.
There's an implicit "working on rockets" in the OP's comment. ULA did 4 launches in 2024 with 2700 employees. SpaceX did 138 launches in 2024 with less than 5 times as many employees. (They had 13,000 in 2023. It's grown a lot since then, but hasn't doubled, I don't think).
ULA employees spend a lot higher percentage of their time on paperwork than SpaceX employees do. The injury rate while doing paperwork should be essentially zero.
SpaceX does 5-10 times as many launches per employee, and has an injury rate 6 times as high per employee. So the injury rate per rocket launch is comparable.
I mean, the CEO also reportedly dislikes yellow, warning signs, and beeping.
https://nymag.com/intelligencer/2018/04/tesla-workers-gettin...
> Among the more baffling details in the report are several sections about how Elon Musk’s personal tastes appear to have affected the factory’s safety for the worse, “his preferences … were well known and led to cutting back on those standard safety signals.” Musk, apparently, really hates the color yellow. So instead of using the aforementioned hue, lane lines on the factory floor are painted in shades of gray. (Tesla denies this and sent Reveal photos of “rails and posts” painted yellow in the factory.) He also is not into having “too many signs” or the beeping sound forklifts make in reverse.
A person familiar with Musk stated that he prefers workers at elevation use neck nooses of fall arrestment gear.
More people sitting around not doing anything can lower the numbers.
The amount of people here who didn’t read the article and don’t understand basic statistics is shocking.
Rate per person does not counter neuroelectrons hypothesis.
If you have 100 people standing around for a year, they probably have lower injuries than the same 100 people who are using heavy machinery over the course of a year. And it's not like most corporations are efficient.
SpaceX may have less redundancy in their workforce causing the injury rate to go up since more people are working more often.
Or SpaceX is overworking people in unsafe conditions and causing way more injuries.
Either way, rate per person does not negate people working vs. Not.
You can have 100 people using heavy machinery safely if you have good culture and take safety seriously.
It's because Starbase workers are roofers, plumbers, and linemen working in the TX weather, not just engineers or assembly line workers in air conditioned clean rooms and offices.
Are they direct SpaceX employees?
I know SpaceX is vertically integrated, but I wouldn't have thought that'd extend to stuff like building sheds and running toilet plumbing.
It's a mix.