SpaceX Discussion Thread

[joesxm3]

We have the IPO discussion threads but I would like to start a thread like the self driving or electric vehicle threads to discuss some more technical info about SpaceX, such as advantages, or product plans etc.

I put this in Other Topics, but moderators please move if a better place is found.

 

[joesxm3]

I will start off with discussion of a new aspect that I researched today.

This week SpaceX filed a patent for StarFall, a reentry pod with a heat shield and parachute. It is relatively large, 10 feet by 2.5 feet and can carry 1000 kilograms. Starfall can be used to deliver product from an orbital factory.

I had a discussion with Grok asking if it was feasible for a starship to be configured as a factory and to stay in orbit for a long time, potentially without any permanent human crew. The answer was 1 to 5 years.

Starships can dock. So resupply ships could deliver raw materials to the factory or bring up a temporary human maintenance team.

What would makes sense to manufacture in orbit rather than on Earth? One thing is silicon crystals to be used as wafers for chip fabs. Gee, Tesla and SpaceX plan to build there own chip fab.

The main benefit to having a silicon foundry in orbit is that the quality of the crystals is much higher. This would increase yield, but Grok says this is secondary. The higher quality wafers are a real advantage in making faster, more powerful chips that make less heat and are more tolerant of radiation. Gee, I wonder if SpaceX could use this type of chip in their AI satellites?

I asked if Optimus robots could run this foundry, but Grok said that it would be much more feasible to have special purpose automation for most tasks, much like Tesla does for cars. Optimus could have a role but more as a care taker than a factory machine.

Grok thought there is a high probability that a pilot project could be running by 2028 to 2030, and a very high probability that a hybrid operation with occasional human visitors by 2030 to 2032.

It seems that having a monopoly on this better quality silicon might be a huge advantage for SpaceX.

[eta] I should give credit to this article on X which inspired me.

https://twitter.com/x/status/2067295318203351365

 

[SnowballCamper]

When will Optimus (or any other) robots be able to clean my house, cook meals, do grocery shopping, wash, dry, fold, and put away my clothes, maintain my house, mow my grass, and prune my shrubs?
All the other stuff is neat, but it doesn't make the average person's life better IMHO.

OMG, I got the hot topic popup... what's up with that?

 

[audreyh1]

When will Optimus (or any other) robots be able to clean my house, cook meals, do grocery shopping, wash, dry, fold, and put away my clothes, maintain my house, mow my grass, and prune my shrubs?
All the other stuff is neat, but it doesn't make the average person's life better IMHO.

OMG, I got the hot topic popup... what's up with that?

Optimus is a Tesla product, not part of SpaceX (yet).

 

[joesxm3]

When will Optimus (or any other) robots be able to clean my house, cook meals, do grocery shopping, wash, dry, fold, and put away my clothes, maintain my house, mow my grass, and prune my shrubs?
All the other stuff is neat, but it doesn't make the average person's life better IMHO.

OMG, I got the hot topic popup... what's up with that?

I think it will be a long time before Optimus is sold to people as a household assistant. Maybe one of the other companies may focus on that.

I think that it is much more likely that Optimus will be used internally by Tesla and SpaceX at first. After that maybe offered to the Tesla supply chain, or maybe paired with a self-driven van for package delivery.

Working in a home is a very complex situation that presents a lot of liability issues.

Elon has made it clear that Optimus will be the go to item for getting the Moon and Mars ready for humans.

Well maybe not the Moon so much but it certainly makes sense to sent Optimus for the initial one-way trips to Mars.

 

[joesxm3]

I see that this thread has already attracted a hot topic flag.

I submit as evidence of its relationship to FIRE that it is an educational resource to help people understand SpaceX in order to support their evaluation of SPCX and TSLA stock as potential investments.

I at first was going to put this in the active investing section. If the moderators feel it better there, please move.

I also submit that this thread is intended to also be similar to the "Space the Final Frontier" thread but more narrowly focused.

Posts that are confrontational or driven by Elon Derangement Syndrome are not welcome.

Posts that disagree and present supporting facts are welcome.

Thanks.

 

[SnowballCamper]

Yes, I'm not following things that closely, sorry I hijacked this with Optimus. I have friends that think robots can do the things I described. I do get the complexity involved with such things as household chores...and thats my point about AI hype... it is mostly hype to me. However, I do concede that the LLMs are impressive in terms of having a conversation with a computer.

It would be amazing if the AI infrastructure ran off solar energy harvested by orbiting satellites that enabled communication everywhere on the surface of the earth. This seems to me the main proposition of SpaceX. This is great for those that get value from the current state of AI (mostly LLMs).

But until the AI makes the average persons life materially better...?? I think that question is not answered yet, nor even the path to get there.

 

[Route246]

I read an article yesterday about NV's Gr00t, the humanoid robot reference design which may have something to say in this matter. The model is Google Android where a standard S/W stack is provided with a reference design that can be built by anyone and allows a team to focus on the application layer (just like an Android phone maker) and have their application run on a reference design (Android-compatible phone) where time and effort doesn't need to be expended on the H/W stack and fussing over designing a phone from scratch. Gr00t seems to be aimed at startups who want to provide a better robot S/W stack while not worrying about the robot's hardware because that is all spec'ed and designed already.

 

[joesxm3]

I read an article yesterday about NV's Gr00t, the humanoid robot reference design which may have something to say in this matter. The model is Google Android where a standard S/W stack is provided with a reference design that can be built by anyone and allows a team to focus on the application layer (just like an Android phone maker) and have their application run on a reference design (Android-compatible phone) where time and effort doesn't need to be expended on the H/W stack and fussing over designing a phone from scratch. Gr00t seems to be aimed at startups who want to provide a better robot S/W stack while not worrying about the robot's hardware because that is all spec'ed and designed already.

I had to have Grok confirm that NV was Nvidia. Grok added that this is indeed related to my previous discussion about space factories and thought that a standard like this could move things forward.

Nvidia seems to be doing a lot of this "building the framework" stuff, cars CUDA for AI etc.

The reference design still needs to be built, but it would give a start up a step to stand on. I have to think, though, that when it comes to building a factory to make thousands of these robots, Elon's focus on "building the machine that makes the machine" gives Tesla and SpaceX an advantage in ramping to scale.

Thanks for pointing me to this Nvidia info. I had not heard of that.

 

[Route246]

I had to have Grok confirm that NV was Nvidia. Grok added that this is indeed related to my previous discussion about space factories and thought that a standard like this could move things forward.

Nvidia seems to be doing a lot of this "building the framework" stuff, cars CUDA for AI etc.

The reference design still needs to be built, but it would give a start up a step to stand on. I have to think, though, that when it comes to building a factory to make thousands of these robots, Elon's focus on "building the machine that makes the machine" gives Tesla and SpaceX an advantage in ramping to scale.

Thanks for pointing me to this Nvidia info. I had not heard of that.

I think one more thing that people are not focused on is Spacex is going to buy Cursor for 60B. I know a little about Cursor as I was using it (the paid enterprise version) quite a bit before I retired last October. The first thing I did was sanity checked my entire toolkit codebase. This included bash, python, C and a little bit of this or that (assembler drivers, etc.). I asked it to stylize my scripts and code, adding relevant comments, etc. and it did a really good job of this. It also found some issues in my code and fixed them while providing a test case to prove the concept. It was impressive but coding is not my primary role, I just do it to maintain an efficiency toolkit (saves tons of time through automating mundane tasks, etc.). The place where it really excelled was coding and cleaning up some of my very complex regex (regular expressions). Some of my regex was so complicated I didn't even remember what it was doing (LOL). Cursor did a good job of explaining and adding good comments. Very impressive stuff. The rub is, Cursor was, at the time, a really good coding assistant engine.

Hardcore agentic tasks were still not being mainstreamed at the time of my retirement. Fast-forward and I'm told that Claude runs circles around Cursor (a few ex-colleagues' opinion, not mine) coding agentic tasks but it is expensive and burns through tokens like a drunk at Mardi Gras. You may have read that some S/W development groups are now under strict token budgets. I think this will normalize as alternatives come into being (remember, excess profits breeds excess competition).

One article I read about the acquisition is that it is a tacit admission by Musk that xAI has not been able to compete. It is no wonder that he is renting out his excess capacity.

60B is a lot to pay for Cursor but it may be OK, given the valuation of Spacex.

 

[Alan]

I at first was going to put this in the active investing section. If the moderators feel it better there, please move.

Thanks.

We agree and have moved the thread to Active investment

 

[Aerides]

I bought a handful of spacex shares enough to have some interest, but I do not remotely believe in the valuations coming to reality in the next decade.

I think it would be helpful to keep what "Grok" has to say from this thread. Musk's AI is part of spacex, and I would not think it unbiased. As always, I'd rather hear what other members think and are analyzing.

 

[joesxm3]

I bought a handful of spacex shares enough to have some interest, but I do not remotely believe in the valuations coming to reality in the next decade.

I think it would be helpful to keep what "Grok" has to say from this thread. Musk's AI is part of spacex, and I would not think it unbiased. As always, I'd rather hear what other members think and are analyzing.

Is this a site team directive, or just your opinion? If it is an official directive, I will follow the ruling.

However, Grok is my go to LLM

I don't think Grok is intentionally biased in favor of SpaceX. This past week Grok has been advising me to reduce my holdings of TSLA stock because of concentration risk and helping to plan a schedule to sell my BPTRX and SPCX shares.

Thinking about your comment I decided to ask Google Gemini whether the orbital factory was plausible, the projected timeline for pilot, hybrid automation with some human involvement and full automation. Gemini agreed that it is plausible and would be a significant advantage for SpaceX, but was not as optimistic on the timelines.

Gemini timelines. Pilot 2029-2032. Hybrid 2033-2037. Fully automated 2038.

Your post led me down an interesting rabbit hole. I took the Gemini response and asked Grok to evaluate since it was more pessimistic. Grok defended its answer but was willing to revise its timelines by a couple years. Grok said that it basically agreed with Gemini's conclusion that orbital silicon wafer factories were possible, likely and would create a significant advantage for Tesla and SpaceX.

Next, I decide to ask Grok if it gives answers that are biased in favor of Elon, Tesla and SpaceX. I fully expected it to deny being biased.

Instead Grok said that based on its training data and the fact that a lot of its information comes from X, that it is likely that its answers are 10% to 20% more favorable than answers from other LLM's. It suggested prompts I could prefix my questions with to instruct it to tone down any bias it might have.

 

[joesxm3]

Route246,

Thank you for sharing your personal experience with Cursor.

I worked as a programmer but will have been retired for ten years this August. I tried to have Grok generate some code to send and receive TCP packets to test my network setup at home, but have not had any hands on experience with Claude or Cursor. I have one friend who uses Claude to automate his computer security work.

Looking at the IPO prospectus, the claimed total addressable market was $28.5 Trillion. Space $370 Billion; Starlink $1.6 Trillion; AI $26.5 Trillion (Infrastructure $2.4 T, Consumer and Advertising $1.36T and Enterprise Applications $22.7T). SpaceX was intentionally vague on what market share of each category they expect to take.

So, if SpaceX dominates its main categories, that is only about $3T TAM. AI is the huge wild card, so monitoring and understanding the AI part, Cursor, Macrohard *digital Optimus" etc. will be critical.

I keep wanting to get moving on learning to use AI for more than just a research assistant, but I get overwhelmed by last century tasks like mowing the lawn and splitting wood.

 

[Aerides]

Is this a site team directive, or just your opinion? If it is an official directive, I will follow the ruling.

However, Grok is my go to LLM

It's my personal request. I don't think LLM's have a place in most discussions here (which IS site policy) unless they are used sparingly.

Grok bias may or may not be (though I'm sure examples are easy to find), but given it is part of the very stock we're hoping to discuss, there are probably better ways to have healthy inputs from the forum members since that is why we are here instead of the machines.

 

[Chuckanut]

I also submit that this thread is intended to also be similar to the "Space the Final Frontier" thread but more narrowly focused.

Actually, this thread is not that similar. “Space the Final Frontier” was and is for everything outer-space. Personally, I would prefer that it stay that way. That would include Astronomy, various discoveries about the Cosmos, planetary discoveries, how to best use a telescope, etc.

The Final Frontier seems to be SpaceX centric, because SpaceX is easily the #1 leader in launching things into space. Nobody else comes close at this time. Hopefully, that will change in the not too distant future. We need competition and variety in the area of space exploration.

 

[Route246]

Route246,

Thank you for sharing your personal experience with Cursor.

I worked as a programmer but will have been retired for ten years this August. I tried to have Grok generate some code to send and receive TCP packets to test my network setup at home, but have not had any hands on experience with Claude or Cursor. I have one friend who uses Claude to automate his computer security work.

Looking at the IPO prospectus, the claimed total addressable market was $28.5 Trillion. Space $370 Billion; Starlink $1.6 Trillion; AI $26.5 Trillion (Infrastructure $2.4 T, Consumer and Advertising $1.36T and Enterprise Applications $22.7T). SpaceX was intentionally vague on what market share of each category they expect to take.

So, if SpaceX dominates its main categories, that is only about $3T TAM. AI is the huge wild card, so monitoring and understanding the AI part, Cursor, Macrohard *digital Optimus" etc. will be critical.

I keep wanting to get moving on learning to use AI for more than just a research assistant, but I get overwhelmed by last century tasks like mowing the lawn and splitting wood.

I detached in October 2026 and walked away from the work grind cold turkey. I stated it here a few times but my break was clean without any negative withdrawal symptoms.

The timelines are so compressed now and the LLM of choice might change based on a new model being introduced that does something better overnight. Competition is a wonderful thing. My only "regret" at this time is I missed out on having access to an enterprise Claude environment. I believe the retail price is $200/month or something like that. My previous Megacorp colleagues have unlimited access to it now (at least for now) and they rave about it. I'm still playing with code but it is my own personal toolkit at home, not the toolkit that I used and shared at Megacorp. I've been so occupied with my personal projects (primary one a major remodel but restarting ham radio, working on a car project and planning my new workshop and man cave) that my coding is truly winding down but winding down at a perfect time because coding assistants are coming into play at the perfect time, allowing me to spend more effort on perfecting the UI and application layer and spending less time on grinding debugged code. For that, I feel blessed and am grateful but I'm on the fringe and it is getting more and more difficult to stay current.

The fact you are staying engaged with Grok is nice to see. I'm partial to Perplexity for now and have a paid subscription so I use it as a primary chatbot. When I need clarification or conformation I engage Gemini, ChatGPT and Claude but I'm just a hacker in that regard feeling my way around. The whole hallucination thing is blown way out of proportion, IMO. What expert user in your past did not occasionally give you information that was not exactly optimal? Even the best expert users have their faults. Everyone made a big deal about hallucinations, trying to discredit LLM/chatbot but it is an evolutionary process to perfect these apps. Competition will take care of the perceived shortcoming IMO.

I see Spacex is coming back down to earth, and I view this as a healthy correction. This is how markets work. Sometimes, irrationality happens but people with money didn't get that way by making poor choices. They got that way by cutting losses and minimizing mistakes. The market is a wonderful thing. People are able to put their money where the mouth is and sometimes they get punched in the nose. Again, a wonderful thing and that is how markets remain perfect and how they self-correct towards that goal.

 

[joesxm3]

Actually, this thread is not that similar. “Space the Final Frontier” was and is for everything outer-space. Personally, I would prefer that it stay that way. That would include Astronomy, various discoveries about the Cosmos, planetary discoveries, how to best use a telescope, etc.

The Final Frontier seems to be SpaceX centric, because SpaceX is easily the #1 leader in launching things into space. Nobody else comes close at this time. Hopefully, that will change in the not too distant future. We need competition and variety in the area of space exploration.

Actually, my intent when starting this thread was more of a combination. I think most people underestimate the marvel of the change we are going through and SpaceX is driving a lot of that change.

There can be endless discussion of space factories, rail guns, armies of Optimi being the first Von Neumann probe, cities on the Moon and Mars. That was what I meant by "like Space the Final Frontier."

The pop up warns that content must be "FIRE-Related", which talking about "everything outer space" does not seem to be. So I adjusted the goal to include SpaceX as an investment.

I may be wrong, but think a lot of the irrational excitement and FOMO around SPCX is not just hunting the quick buck but more related to the belief that SPCX is today's Dutch East India Company.

I read that thousands of employees bought more stock in the IPO and that they are not likely to jump out as soon as the lockup expires. I don't know if that is just because they figure they will get more money by holding or whether they are truly committed to the long term goal of making mankind multi-planetary.

I may be a victim of propaganda, but I think that Elon is systematically working on a step by step plan to achieve those goals. I think he sees accumulating so much wealth as a way to have the ability to direct the direction of things. People on this forum already have trouble spending enough of their wealth, so Elon long ago passed the point where more money affects his standard of living. Maybe he is on a power trip, but I would bet more on the idea that he is a nerd making his childhood fantasies a reality.

But bringing thoughts back to investing, I think that the "dreamer effect" and the "hero worship effect" is a major factor in the stock price. I agree that there is no way to justify the current valuation even using extremely optimistic forward projections.

The hype seems to be wearing off a little in the past two days. Once things normalize to a still overvalued by normal measures levels I plan to slowly DCA into a larger position. But I realize that the time frame is more suited to buying some hold forever shares for my 13 year old cousin.

 

[Aerides]

I may be a victim of propaganda, but I think that Elon is systematically working on a step by step plan to achieve those goals. I think he sees accumulating so much wealth as a way to have the ability to direct the direction of things. People on this forum already have trouble spending enough of their wealth, so Elon long ago passed the point where more money affects his standard of living. Maybe he is on a power trip, but I would bet more on the idea that he is a nerd making his childhood fantasies a reality.

In post 6 you have asked forum members not to be confrontational re Elon. It might then be helpful if you not continue to share your positive opinions about him because that invites the alternate more critical view and then things will go south pretty quickly. Been there, seen that, as have all of us.

Whether you'd consider my opinion of him confrontational or deranged is beside the point: any time we've had Elon be the topic in this forum it is a challenge because he is a divisive and polarizing figure (and that started years before his politics became evident).

I realize it's difficult to extract an oversized personality from this topic - the elephant in the room as it were - but I'm sure we can manage to do so for the sake of the discussion.

 

[joesxm3]

Good point.

It is hard to separate SpaceX activities from Elon's vision (not to mention the key man risk to the stock), but I agree that I don't want to provoke people into the type of discussion we do not want on a friendly thread.

Thanks for pointing it out.

 

[joesxm3]

I just saw that SpaceX receieved a Baa1 investment grade bond rating from Moody's. I also saw that there is a $20B or so bridge loan from xAI that they are going to roll over into s bond issue.

https://twitter.com/x/status/2067706794990780917

 

[Route246]

Good point.

It is hard to separate SpaceX activities from Elon's vision (not to mention the key man risk to the stock), but I agree that I don't want to provoke people into the type of discussion we do not want on a friendly thread.

Thanks for pointing it out.

I believe the moderators want to avoid divisive discussion. The title of this thread is Spacex, not a person. If you go back and read my comments I am focusing on the corporation, the market reaction, stock price, lack of profits and pure facts unrelated to a personality.

 

[joesxm3]

I believe the moderators want to avoid divisive discussion. The title of this thread is Spacex, not a person. If you go back and read my comments I am focusing on the corporation, the market reaction, stock price, lack of profits and pure facts unrelated to a personality.

I think the point I was trying to make when I mentioned Elon is that while SpaceX's future milestones may seem unreasonable, it you examine their past activities you can see that each apparently unrelated thing is actually a step towards the long range goals and is done in a way that generates revenue that can be invested in developing the next step.

You can take this a step farther when looking at all of the Musk companies. Tesla built cars to generate revenue to invest in batteries, chip design and robotics. SpaceX is building the launch and communication capabilities. Boring company is building tunneling and possibly mining machines. xAI is building agentic AI that can help design chips and can control swarms of Optimus robots.

So you can picture a fleet of Starships, carrying a legion of Optimus robots, who drive around in Cyber Trucks and live in underground habitats dug by the Boring Company while communicating back to Earth using Starlink laser networks.

Right now it may seem that SpaceX is losing a lot of money due to xAI and X, but they have shown that they can set up data centers faster and cheaper than anyone and have established relationships to be able to buy or build the scarce chips. Renting out compute may be the next step in the pan to generate revenue to move forward. And that is even before hosting servers in orbit.

Once the initial hysteria wears off and if the stock price drops, even to the IPO level SPCX could be a good investment for those with a long enough time horizon.

 

[Chuckanut]

I’ll just ad that if you want to get an idea of what goes through Mr. Musk’s head listen to his interview with Joe Rogan, the one were he took a drag from a joint. (Yes, incredibly foolish.) The joint part takes about one minute in what is a 2-3 hour interview. Then judge for yourself.

With Blue Origin’s and ULA’s new rockets on hold for a while, and Boeing’s Starliner still grounded until they can make is safe for humans, SpaceX is the only real choice we have for large or even medium sized payloads, and manned flights. I suppose we could use Russian rockets. Ahh…. No, I don’t think so.

 

[Route246]

I think the point I was trying to make when I mentioned Elon is that while SpaceX's future milestones may seem unreasonable, it you examine their past activities you can see that each apparently unrelated thing is actually a step towards the long range goals and is done in a way that generates revenue that can be invested in developing the next step.

You can take this a step farther when looking at all of the Musk companies. Tesla built cars to generate revenue to invest in batteries, chip design and robotics. SpaceX is building the launch and communication capabilities. Boring company is building tunneling and possibly mining machines. xAI is building agentic AI that can help design chips and can control swarms of Optimus robots.

So you can picture a fleet of Starships, carrying a legion of Optimus robots, who drive around in Cyber Trucks and live in underground habitats dug by the Boring Company while communicating back to Earth using Starlink laser networks.

Right now it may seem that SpaceX is losing a lot of money due to xAI and X, but they have shown that they can set up data centers faster and cheaper than anyone and have established relationships to be able to buy or build the scarce chips. Renting out compute may be the next step in the pan to generate revenue to move forward. And that is even before hosting servers in orbit.

Once the initial hysteria wears off and if the stock price drops, even to the IPO level SPCX could be a good investment for those with a long enough time horizon.

Again, I'm not going to talk about a person, however your statement about Xai setting up data centers is questionable. I know a little about the DC business and there are data centers and then there are data centers. The ones being referenced as quick setup have a lot of compromises and shortcuts built into them. Your chatbot of choice can confirm this. It is one thing to build a data center and get servers provisioned and then it is another thing to do it properly, sustainably and by strict standards. Reuse of old warehouses is a common shortcut but the infrastructure of old warehouses is not the same as a purpose-built structure able to handle the weight, power and cooling requirements of a modern AI/ML workload. There may be something that "works" but it is not optimal and design shortcuts were taken and the usual issue is thermal overrun in specific regions of the footprint. The quality of the power and backup system is also something that can be costly and take a long time if done properly and can be cheap, quick and dirty if you are focused on speed to occupancy. I know someone who did some work at the Memphis site and the stories he told were "interesting" about the execution.

Understand that a data center is sometimes called a "carrier hotel" or similar nickname. The hotel metaphor is appropriate. The quick and dirty data centers (and there are plenty of them out there) are akin to a low-end motel on the outskirts of town where you have to sleep with one eye open. The high-end data centers from the big guys (Vantage, Equinix, Digital Realty, et. al.) are more akin to Hilton class on the lower end and Four Seasons (Equinix) on the high end. The large companies that specialize in DC buildout are mostly REITs and specialize in the facility and infrastructure only. Much of AWS is in-house but they also have a huge lease footprint with the big guys. I think MSFT, Google, and for sure Meta usually lease from the big guys. Doing it in-house is not a core competency and quality and efficiency suffers.