Analysis written by Van Ha Trinh, Financial Markets Analyst at Exness. 

The most interesting aspect of this IPO may not be what SpaceX has built. It is the fact that two rational, informed traders study the same prospectus and arrive at valuations nearly a trillion dollars apart. That tension is the story. Because nobody can agree whether it is visionary or delusional, and the gap between those two positions is measured in hundreds of billions of dollars.

What the prospectus actually reveals

SpaceX declares it has identified the largest actionable Total Addressable Market (TAM) in human history and then quantifies it at US$28.5 trillion. 

Source: SpaceX Form S-1, Exness

The chart tells you everything about how SpaceX wants to be seen:

• Space (launch, satellites, exploration): US$370 billion ~ 1.30%

• Connectivity (Starlink broadband + mobile): US$1.6 trillion ~ 5.62%

• AI (infrastructure, consumer, advertising, enterprise): US$26.5 trillion ~ 93.08%

The internal composition of that TAM is where it gets truly audacious. Of the US$28.5 trillion total, $26.5 trillion, nearly 90% of the entire figure, is attributed to xAi alone. Not rockets. Not Starlink. A category SpaceX did not compete in until it absorbed xAI six months ago. 

Now here is where the GDP comparison becomes the most useful analytical lens available. $28.5 trillion is almost exactly the annual GDP of the United States, the largest economy on Earth, representing roughly 25% of all global economic output.  Put differently, SpaceX is claiming it has identified a market opportunity equal in size to every good and service produced by 335 million Americans in an entire year. Global GDP sits at approximately $110 trillion. SpaceX's claimed TAM represents roughly 26% of the entire world's annual economic output, and that is excluding China and Russia. 

The TAM is not the problem. TAMs are always aspirational. The problem is the implied capture rate baked into the IPO price and whether a company that generated $18.7 billion in total revenue last year deserves to be priced as though it has already won a war it has not yet entered.

This comparison is not just an interesting bar chart. It is a diagnostic tool for intellectual honesty.

Understanding SpaceX’s business structure

Source: SpaceX Form S-1, Exness

Following the merger with xAI, the company operates across three core segments: Connectivity 

(Starlink), Launch and AI.

Starlink / connectivity: the core high-margin cash engine

• Financial Performance: Generated US$11.4 billion in 2025 revenue, accounting for 61.9% of the company's total top-line performance. It has successfully captured a massive 10.3 million subscriber base spanning 164 countries.

• Profitability: Delivered a stellar segment-level operating income of US$4.4 billion at an approximate 63% EBITDA margin. It stands as SpaceX's sole profitable division after clearing its capital-intensive deployment phase.

• Competitive Moat: Operates as a software-style global infrastructure monopoly with an unreplicable fleet of 9,600+ low-Earth orbit (LEO) satellites. This establishes an order-of-magnitude lead over emerging competitors like Amazon's Kuiper (~500 satellites) and OneWeb (~650 satellites).

Launch services: Starship development driving asymmetric upside

• Financial Performance: Contributes approximately 22% of total top-line revenue, posting over US$4 billion in 2025.

• Capital Constraints: The division is heavily exposed to a capital-intensive investment phase, posting a US$662 million operating loss in 1Q2026 with cumulative Starship development spend exceeding US$15 billion.

• Market Dominance: The Falcon 9 continues its run as the world's most reliable and frequently launched vehicle, executing approximately 161 launches in 2025 compared to competitors such as Rocket Lab and Blue Origin, with only 18 and 11 launches, respectively

• The Disruption Ultimate Goal: If Starship successfully achieves its target cost structure to reduce launch costs to ~US$100/kg (down from the current ~US$1,500/kg), the technology will render every existing launch vehicle commercially obsolete.

Artificial Intelligence (xAI): the high-beta infrastructure pivot

• Financial Performance: Recorded US$818 million in Q1 2026 revenue and US$3.2 trillion in 2025, but remains locked in a high cash-burn phase.

• Current Operational Losses: Posted a staggering US$2.47 billion operating loss in Q1 2026, acting as the primary driver behind SpaceX’s consolidated red ink.

• Monetisation & "Picks and Shovels" Model: Rather than competing directly with entrenched consumer AI rivals, the segment prioritises industry collaboration through an infrastructure leasing model. This is anchored by a disclosed US$1.25 billion/month compute contract with Anthropic (US$15 billion annually), establishing a clear path toward near-term profitability.

• Future Catalyst: The segment aims to develop space-based orbital AI data centres, providing a definitive solution to the intense power consumption and heat dissipation bottlenecks currently facing ground-based tech infrastructure.

The Bull Case: three compounding speculations, each explosive on its own

• Starlink’s software-like hyper-monetisation: Boasting an incredible 63% EBITDA margin, Starlink functions more like a high-margin SaaS giant than a telecom utility. As it aggressively scales from residential users to high-ARPU enterprise, maritime, aviation, and direct-to-cell markets, it will unlock an unstoppable, recurring cash fountain to fund the rest of the ecosystem.

• Starship’s dominance of space logistics: Achieving a cost structure of US$100/kg will give SpaceX absolute pricing power over the global space economy. This is the ultimate asymmetric upside in the prospectus, allowing SpaceX to launch its own massive constellations and heavy orbital infrastructure at near-zero internal cost while forcing traditional aerospace entities into obsolescence.

• Space-based AI centres dominate the next tech wave: The US$15 billion annual Anthropic contract proves that xAI’s true value lies in infrastructure provision rather than consumer apps. By moving supercomputers into low-Earth orbit, SpaceX offers a definitive solution to Earth's power grid shortages and heat dissipation bottlenecks, unlocking a Total Addressable Market (TAM) valued at US$26.5 trillion for orbital AI computing.

The Bear Case: brilliant business, mission impossible

Source: SpaceX Form S-1, Exness

• Stretched multiples and inflated TAM projections: The aggregate US$1.75 trillion valuation implies a standalone valuation of US$600 billion to US$900 billion for the AI segment. Underwriters allocating US$26.5 trillion of the US$28.5 trillion total TAM to AI is classic IPO marketing fluff. In reality, xAI's consumer vertical heavily lags behind incumbents, and the mảng is bleeding cash with a US$4.3 billion net loss in Q1 2026 alone.

• Endless Capex black holes and cash burn vulnerability: To maintain its lead, SpaceX must remain a hyper-aggressive cash burner. Starlink’s hard-earned profits are currently being entirely consumed by AI losses and Starship’s multi-billion-dollar development cycles. If global tech capital expenditure or the AI arms race cools down, SpaceX's heavily leveraged financial structure will face severe post-IPO strains.

• Thermal physics bottlenecks and governance red flags: Technically, operating high-density data centres in a vacuum environment faces unforgiving radiative heat dissipation hurdles that lack large-scale commercial precedent. Governance-wise, Elon Musk holding over 85% of voting rights with only ~46% equity is a major corporate governance warning sign for institutional funds demanding standard checks and balances.

The bull and bear debate is not a sign of market confusion. It is a sign that SpaceX is genuinely, structurally unlike anything that has ever been brought to public markets before. For traders who understand asymmetry, that same ambiguity is the setup.

The most interesting aspect of SpaceX may not be the rockets, the satellites, or even the AI ambition. It may be the company has managed to build something so complex, so multi-layered, and so dependent on one man's continued execution.

In the history of public markets, that has never happened before the listing day. It is happening now.

Sponsored

Not when the rockets blow up , and when Massive Net Losses: The company suffered a staggering net loss of roughly $4.9 billion in 2025 and an additional $4.3 billion in just the first quarter of 2026 ,if the money it receives from the

Government dries up they will have bigger loses.

regards worgeordie

No. But when have markets ever priced what anything is actually worth?

Anything is worth exactly how much people are willing to pay for it.

This should be an interesting stock to watch. Its about 100 years (?) ahead of its time in terms of being a space transportation company stock?

While the idea of investing in this is appealing to me on one hand, on the other hand they are still in development with their rockets, and their test philosophy is to push their test rockets to the limit (and deliberately risk blowing the rocket up) so to better understand the test rocket limitations. I think many investors/speculators may not understand such, and hence we could see a fair amount of volatility each time they blow up a test rocket.

I also do not like airline stocks, as fuel prices and flying incidents (not to mention competition) make such airline transportation stocks too risky for me. With Space-X being a transportation company in many ways (albeit space transportation), I decided to stay out, and just watch.

My best wishes and good luck to all who invested in Space-X.

50 minutes ago, oldcpu said:

This should be an interesting stock to watch. Its about 100 years (?) ahead of its time in terms of being a space transportation company stock?

While the idea of investing in this is appealing to me on one hand, on the other hand they are still in development with their rockets, and their test philosophy is to push their test rockets to the limit (and deliberately risk blowing the rocket up) so to better understand the test rocket limitations. I think many investors/speculators may not understand such, and hence we could see a fair amount of volatility each time they blow up a test rocket.

I also do not like airline stocks, as fuel prices and flying incidents (not to mention competition) make such airline transportation stocks too risky for me. With Space-X being a transportation company in many ways (albeit space transportation), I decided to stay out, and just watch.

My best wishes and good luck to all who invested in Space-X.

How much does the value depend on Musk, and how much depends on the company itself?

Edited June 18Jun 18 by Hummin

23 minutes ago, Hummin said:

How much does the value depend on Musk, and how much depends on the company itself?

That IMHO is hard to say. I think unquestionably Musk is the driving force behind Space-X.

Having typed that, Space-X has some very very amazing and incredibly technically capable people working at the company, enabling the company to achieve the almost sci-fi instructions that Musk instructs them to design/implement/operate.

I personally believe, that if Musk were to no longer being involved, that without his demanding push, that the company would lose its 'edge'. But clearly that belief is no more than a judgement equivalent to speculation by myself.

The only company involved in spacex that makes any money is starlink. The rest are cash drains. Like Tesla, the value of the stock is all Musk hype.

Same as gambling on bitcoin.

What I don’t understand is this: XAi is not the only AI in the business.

It has many rivals in the US without even considering China, and any others that might be being developed. Unless these each find a specific niche, they will all be competing for the same business, and only the best ones will win.

Either that or a price race to the bottom. So that potential market might not be as rich as they all think.

10 hours ago, gargamon said:

The only company involved in spacex that makes any money is starlink. The rest are cash drains. Like Tesla, the value of the stock is all Musk hype.

Same as gambling on bitcoin.

I won't dispute the gambling assessment. IMHO a position in any company stock, to a lessor and greater degree, is gambling.

However what i read is the partly re-usable Falcon-9 rocket is also making money. Contrary to your view that Starlink is the only money maker.

10 hours ago, phetphet said:

What I don’t understand is this: XAi is not the only AI in the business.

i think the 'niche' (?) Space-X plan to do, is put XAi servers in a TBD earth orbit, using solar power that is not diluted from the earth's atmosphere, to power the XAi servers/computers. Assuming Space-X can minimize the satellites (with XAi servers embarked) exposure to the occasional lunar eclipse of the sun and minimize exposure to the occasional earth eclipse of the sun, the satellites will obtain massive sunlight. That sunlight will produce lots of power via the solar panels. i believe Space-X is the only organisation with an Ai product planning such.

Why the interest in power? Purportedly AI requires massive electrical power. This (electrical power) tends to be a MAJOR expense for companies with AI products. Space-X presumably plan on massive cost savings here. I have not studied this enough to risk providing more details, that i could get wrong.

So this may (or may not) give Space-X's XAi a commercial advantage.

Edited June 18Jun 18 by oldcpu

I was around when the dot.com hype came to an end.

We are in a phase where "missing out of something" is actually the driving force. Index tracker funds will have to buy into it, fuelling the hype some more.

Let it run, enjoy the ride. But by the time "everybody" including your uncle is involved, this would be a good time to sell.

Rinse and repeat. As before.

On 6/18/2026 at 10:33 PM, swissie said:

I was around when the dot.com hype came to an end.

We are in a phase where "missing out of something" is actually the driving force. Index tracker funds will have to buy into it, fuelling the hype some more.

Let it run, enjoy the ride. But by the time "everybody" including your uncle is involved, this would be a good time to sell.

Rinse and repeat. As before.

If the stock market has a significant correction sometime soon, which may well happen, depending, SPCX could take a significant hit. At some point, it might be good for a dead cat bounce trade. That's what I'd wait for. Make some money on the bounce, pull out my cost, and let the winnings ride.

On 6/18/2026 at 9:40 PM, oldcpu said:

i think the 'niche' (?) Space-X plan to do, is put XAi servers in a TBD earth orbit, using solar power that is not diluted from the earth's atmosphere, to power the XAi servers/computers. Assuming Space-X can minimize the satellites (with XAi servers embarked) exposure to the occasional lunar eclipse of the sun and minimize exposure to the occasional earth eclipse of the sun, the satellites will obtain massive sunlight. That sunlight will produce lots of power via the solar panels. i believe Space-X is the only organisation with an Ai product planning such.

Why the interest in power? Purportedly AI requires massive electrical power. This (electrical power) tends to be a MAJOR expense for companies with AI products. Space-X presumably plan on massive cost savings here. I have not studied this enough to risk providing more details, that i could get wrong.

So this may (or may not) give Space-X's XAi a commercial advantage.

One or two problems from the top of my head.

First, how do you get rid of the heat generated by the servers? It is possible, but much harder in space without an atmosphere to carry the heat away.

Second, space junk and debris. One small object at high speed can do serious damage.

3 hours ago, Hummin said:

One or two problems from the top of my head.

First, how do you get rid of the heat generated by the servers? It is possible, but much harder in space without an atmosphere to carry the heat away.

Its a good question. And the answer is straight forward. It easy. Satellites with devices that need cooling, typically have radiators that point to deep space, to provide cooling. Satellites that i worked on, had radiators that could lower the temperature to less than 100 kelvin, which obviously is far far cooler than is needed for servers.

It is more likely the satellite with the servers will need additional heaters !!!! so not to get too cold. Yes, the hot servers will help a LOT in warming up, but quite possibly, they won't be enough.

Clearly the engineering design, and carefully calculated size of the satellite's radiators will be incredibly important - illustrating the importance of your point - albeit perhaps not in the manner you envisaged.

3 hours ago, Hummin said:

Second, space junk and debris. One small object at high speed can do serious damage.

Spacejunk is an ongoing problem. In addition to NORAD, and US (and Canada) Space Command tracking of space junk, so do European government organization and commercial organisations (who sell collision risk info) track and provide warnings of potential collisions with space junk. It is an ongoing issue ... but one the press typical warns of the danger, but says NOTHING about the ongoing mitigation aspects. ... I guess its not surprising. Mitigation aspects rarely get headlines or people interested to read.

Don't care now....got my IPO allocation and sold at $210.....party on.

3 minutes ago, oldcpu said:

Its a good question. And the answer is straight forward. It easy. Satellites with devices that need cooling, typically have radiators that point to deep space, to provide cooling. Satellites that i worked on, had radiators that could lower the temperature to less than 100 kelvin, which obviously is far far cooler than is needed for servers.

It is more likely the satellite with the servers will need additional heaters !!!! so not to get too cold. Yes, the hot servers will help a LOT in warming up, but quite possibly, they won't be enough.

Clearly the engineering design, and carefully calculated size of the satellite's radiators will be incredibly important - illustrating the importance of your point - albeit perhaps not in the manner you envisaged.

Spacejunk is an ongoing problem. In addition to NORAD, and US (and Canada) Space Command tracking of space junk, so do European government organization and commercial organisations (who sell collision risk info) track and provide warnings of potential collisions with space junk. It is an ongoing issue ... but one the press typical warns of the danger, but says NOTHING about the ongoing mitigation aspects. ... I guess its not surprising. Mitigation aspects rarely get headlines or people interested to read.

I know radiators exist. The space station has them, and satellites use them too.

My point is scale.

Servers are a different animal. A real server facility produces a lot of heat per square meter. Almost every watt going in becomes heat, and that heat has to be removed all the time.

I just became more aware of this with my own solar system lately, how much ventilation and cooling air matters. That is why the challenge in a no atmosphere environment strikes me. So yes, radiators work. The real question is how big they must be, how heavy, how expensive, how exposed, and how many failure points you create when you scale it up to server farm level.

That is also why I am pointing at the Moon as a future possibility. If this is not realistic as a free flying satellite or server farm, could the Moon make more sense long term? It has space, material, and maybe the possibility to build proper facilities with an internal cooling atmosphere to move heat around.

The heat still has to go somewhere. That is the core issue.

Edited June 20Jun 20 by Hummin

15 minutes ago, MIke B Bad said:

Don't care now....got my IPO allocation and sold at $210.....party on.

Oh, you're such a financial genius...

2 minutes ago, Hummin said:

I know radiators exist. The space station has them, and satellites use them too.

My point is scale.

Servers are a different animal. A real server facility produces a lot of heat per square meter. Almost every watt going in becomes heat, and that heat has to be removed all the time.

I just became more aware of this with my own solar system lately, how much ventilation and cooling air matters. That is why the challenge in a no atmosphere environment strikes me.

So yes, radiators work. The real question is how big they must be, how heavy, how expensive, how exposed, and how many failure points you create when you scale it up to server farm level.

That is also why I'm pointing at the moon Moon as a future possiblity. If this is not realistic as a free flying satellite or server flying farm, could the Moon make more sense long term? It has space, material, and maybe the possibility to build proper facilities with an internal cooling atmosphere to move heat around.

The heat still has to go somewhere. That is the core issue.

Deep space is incredibly cold. There is nothing on earth like it. Nothing.

Its a mere matter of radiator design to propagate/dissipate the heat from the servers to space. But in doing so, it allows cooling going back the other way. Obviously the radiators need to be pointed to deep space (and not pointed at the earth). And MOST important, in addition, additional heaters will absolutely be needed to prevent server damage from being too cold - likely cycling on/off to to prevent too cold a temperature.

If this is a topic of interest to you, there are many books on satellites.

Edited June 20Jun 20 by oldcpu

9 minutes ago, Hummin said:

I know radiators exist. The space station has them, and satellites use them too.

Indeed. Yes I worked on the ISS program, plus both North American and European satellite programs for decades, as an engineer responsible for the satellite's telescope, detection sensors and also computers (and also as an operator). But one not need first hand knowledge (like i have), .... as noted, there are good text books on this.

Did you also work on such? It would be good/fun to converse with a fellow engineer/satellite operator at times (possibly better via private exchange).

Soon, Elon will begin mining the ASTEROIDS....

When this happens, SpaceX will easily be worth USD250 Trillion, with a T.

(in today's dollars)

Edited June 20Jun 20 by GammaGlobulin

3 minutes ago, oldcpu said:

Indeed. Yes I worked on the ISS program, plus both North American and European satellite programs for decades, as an engineer responsible for the satellite's telescope, detection sensors and also computers (and also as an operator). But one not need first hand knowledge (like i have), .... as noted, there are good text books on this.

Did you also work on such? It would be good/fun to converse with a fellow engineer/satellite operator at times (possibly better via private exchange).

I have worked with deep well fluid solutions and tools in oil well service, operating in confined spaces, so I know a little about heat transfer, pressure, flow, and what happens when theory meets equipment in the field.

I am more curious than trying to prove a point, so I appreciate your reply

It's time the government take it over and let Bernie Sanders run it.

44 minutes ago, Hummin said:

That is also why I am pointing at the Moon as a future possibility. If this is not realistic as a free flying satellite or server farm, could the Moon make more sense long term? It has space, material, and maybe the possibility to build proper facilities with an internal cooling atmosphere to move heat around.

I am very interested in the current Artemis program.

Reference your pondering on this, while a lunar-based AI facility is an intriguing concept, IMHO it is likely at least a decade away from practical viability.

I have only worked on GEO and LEO so i can only speculate re: Lunar. Compared to Low Earth Orbit (LEO), the Moon imposes severe logistical constraints:

• the extreme Delta-v requirements and launch costs for lunar (compared to LEO) drastically limit lunar payload capacity,

• the technical complexity of a successful soft lunar landing and autonomous lunar deployment and assembly introduces risks far greater than those for basic LEO satellite platforms.

• Furthermore, the Moon's 14-day night cycle creates a massive power availability challenge, necessitating heavy energy storage or nuclear infrastructure to survive the long periods of darkness.

Given these factors — along with the difficulty of maintaining even more complex than LEO systems, without human intervention — the risk-to-reward ratio for launching, deploying, and operating a server farm on the Moon, at present time, remains significantly less favorable than continuing to advance the capabilities of LEO-based computing.

Still - maybe someday, in the future.

Edited June 20Jun 20 by oldcpu

3 hours ago, gargamon said:

Oh, you're such a financial genius...

Too kind.

All minor problems.

Outside of the protective magnetic shield of the earth, all space dwellers will die of constant solar/universal bombardment of "particles" that will destroy our GNA. End of story. Tell it to Elon.

Otherwise, has anyone else a cure for radiactive bombardement that makes life impossible? Speak freely.

On 6/20/2026 at 9:46 PM, swissie said:

All minor problems.

Outside of the protective magnetic shield of the earth, all space dwellers will die of constant solar/universal bombardment of "particles" that will destroy our GNA. End of story. Tell it to Elon.

Otherwise, has anyone else a cure for radiactive bombardement that makes life impossible? Speak freely.

There is work going on examining this very point.

A number of the speculative proposals for Lunar (and Martian) longer term accommodations are proposing putting lunar regolith (or Maritan regolith) over the structure. That regolith not only is to mitigate the enormous temperature swings, but also acts as being considered to act as a decent radiation shield.

Other aspects being considered are electronic 'shields' or 'magnetic shields, where active research is ongoing into magnetic and electronic (electrostatic) shielding methods as alternatives or enhancements to traditional "passive" material shielding (like heavy aluminum or hydrogen-rich plastics). This is especially relevant to longer term space travel.

At this point in time? Just research.

Edited June 22Jun 22 by oldcpu

On 6/20/2026 at 4:46 PM, swissie said:

All minor problems.

Outside of the protective magnetic shield of the earth, all space dwellers will die of constant solar/universal bombardment of "particles" that will destroy our GNA. End of story. Tell it to Elon.

Otherwise, has anyone else a cure for radiactive bombardement that makes life impossible? Speak freely.

What about natural protection?

The Moon has lava tubes, and Mars has volcanic formations and possible cave systems too. I read about these ideas years ago already as possible solutions. Would that not be one of the possible answers against radiation, micrometeorites, and extreme surface exposure, instead of only thinking about exposed structures?

And what about subsurface ice or frozen water deposits, especially on Mars? Not only as a water resource, but maybe also as part of the thermal question, if used carefully.

I am asking more than claiming. It seems to me the long term answer may be to use the planet or Moon itself as protection, not build everything fully exposed on the surface.

On 6/20/2026 at 7:33 AM, GammaGlobulin said:

Soon, Elon will begin mining the ASTEROIDS....

When this happens, SpaceX will easily be worth USD250 Trillion, with a T.

(in today's dollars)

Yes, he will send a spaceship to collect a bunch of Asteriods. The metalls worth 500$. The mission has cost 20 million. Boy, I want in on this deal.

PS: Not making the headlines, but in Frebruary this year, Elon declared that he will focus on a "moon mission" instead of a "mars mission". Great. That means that the financial wizard of the world has decided that it would be a good idea to familiarise himself with some basic phisycs.

Like anything else, it is worth whatever people will pay for it.