LordVile95
Posts: 1,033 +721
Welcome to AI where the cash flow is made up and the “profits” don’t matter
Anthropic is heading toward the largest IPO ever, at a possible $2 trillion valuation
- Thread starter Skye Jacobs
- Start date
Axeia
Posts: 1,225 +1,602
Luck of the draw? For PHP, Python and C# I've had better results from Claude than from ChatGPT and Gemini.
Although I try to to use to latter two more as they aren't as stingey with their limits.
Only syntax errors I encounter are from all of them having a hard time sticking to a certain version of the language or frameworks used. They'll try to use features that have been removed or vice versa not introduced yet.
Gemini in particular has a habit of trying to present me with WPF code as well whilst I'm using Avalonia with C#. Which works similar but often slightly different.
Frustratingly if you have them rework the same piece of code a few times they might sneak back in a syntax error from version #2 back into version #8 as well, even within the same session.
In my own experience and what I see from tests and comments on tech sites Claude is quite consistently ranked the highest for coding. Rankings after that get muddy, possibly due to the rest having some languages down better than others or due to when they're tested (version of the LLM).
havok585
Posts: 530 +411
wrong on so many levels, I know you are a hater because if you actually did your homework, the biggest profits made were done under spaceX with the big fat government contracts that actually deliver heavy loads to ISS.
LordVile95
Posts: 1,033 +721
However Elon rolled xAI and Twitter into the company before the IPO which are both unprofitable money pits driven by his own Ego and caused the entire company to run at a loss. The only part that makes a profit is starlink and that’s debatable with the amount of satellites they’re going to have to replace soon
Megalomaniac
Posts: 1,373 +1,458
Megalomaniac
Posts: 1,373 +1,458
I don't even bother explaining to plebians how this works. These people invest in meme coins for God's sakes.
“Plebians”?
We somehow went from discussing an IPO to addressing the landed aristocracy.
The amusing part is that the actual numbers don't support either extreme being thrown around here.
Government contracts are important to SpaceX, but they represented roughly 20% of 2025 revenue. Starlink is its largest profit engine, but Reuters reported that SpaceX was generating substantial profit before xAI was folded into the company, with its reusable launch business also contributing to earnings.
xAI then brought enormous losses and capital requirements into the consolidated company, which is why simply saying “SpaceX loses money” or “only Starlink makes money” misses what actually happened.
You don't need meme coins to understand that.
You just need financial statements.
Apparently those are harder to obtain than a title of nobility.
DanteSmith
Posts: 199 +343
“What matters is net profit, not revenue” is true only if you're trying to reduce investing to one line on an income statement.
Investors aren't buying yesterday's profit. They're buying a claim on future cash flows.
That is especially important with a growth company. A business can be losing money because the underlying economics are terrible, or because it is deliberately spending heavily on R&D, infrastructure, hiring and expansion to build something much larger. Those are two very different kinds of losses.
So yes, profit matters. But so do revenue growth, gross margins, cash burn, capital requirements, free cash flow and whether those losses are actually creating future earning power.
The real question with Anthropic isn't simply...“Is it profitable today?”
It's...“Are today's losses building a business that can eventually generate enough cash to justify a $2 trillion valuation?”
And that's where I become skeptical.
Because even if Anthropic becomes an enormously successful company, it can still be a poor investment if investors pay too much for that future success today.
That's investing though, taking risk on what a business may become, not waiting until every uncertainty is gone and then acting surprised that everyone else already priced it in.
Net profit matters.
It just isn't the entire spreadsheet.
LordVile95
Posts: 1,033 +721
That’s what a pleb would say“Plebians”?
We somehow went from discussing an IPO to addressing the landed aristocracy.
However the profitable arm has hit a few stumbling blocks and China are currently well on their way to compete with it.
Exactly.
The nobility apparently use vibes, titles and inherited confidence.
We plebs are still stuck with revenue, margins, cash flow and financial statements.
China is certainly becoming a serious competitor, and anyone assuming SpaceX will have the field to itself forever would be foolish.
But “well on their way to compete with it” needs a little perspective.
As of June, China's Guowang and Qianfan/SpaceSail constellations had only a few hundred satellites combined. Starlink had roughly 10,400. Reuters even quoted a Chinese industry executive estimating that an optimistic scenario might see China reach Starlink's current scale around 2033.
Meanwhile Starlink's latest quarterly revenue grew 66%, its subscriber base doubled to roughly 12 million, and SpaceX intends to launch another 1,000 next-generation satellites within a year.
Yes, Starlink has challenges...falling revenue per subscriber, enormous capital requirements and a constellation that must continually be refreshed. Those are real things investors should examine.
But competition existing isn't the same thing as the incumbent business suddenly becoming unprofitable or obsolete.
China successfully recovering a reusable booster is impressive too. SpaceX has already done that hundreds of times.
So I'd describe China as rapidly developing a competitor, not standing shoulder to shoulder with SpaceX today.
Otherwise we might as well declare every company doomed the moment somebody else opens a laboratory.
LordVile95
Posts: 1,033 +721
It’s more the profitability of the profitable arm of the company is likely to go down whereas the money pits will continue to be money pits or worst case turn into white elephants which you could argue twitter already is, considering public information shows it only generates about 350m in revenue vs about a 3bn cost and is bailed out a lot by a very generous 2bn a year licence from xAIExactly.
The nobility apparently use vibes, titles and inherited confidence.
We plebs are still stuck with revenue, margins, cash flow and financial statements.
China is certainly becoming a serious competitor, and anyone assuming SpaceX will have the field to itself forever would be foolish.
But “well on their way to compete with it” needs a little perspective.
As of June, China's Guowang and Qianfan/SpaceSail constellations had only a few hundred satellites combined. Starlink had roughly 10,400. Reuters even quoted a Chinese industry executive estimating that an optimistic scenario might see China reach Starlink's current scale around 2033.
Meanwhile Starlink's latest quarterly revenue grew 66%, its subscriber base doubled to roughly 12 million, and SpaceX intends to launch another 1,000 next-generation satellites within a year.
Yes, Starlink has challenges...falling revenue per subscriber, enormous capital requirements and a constellation that must continually be refreshed. Those are real things investors should examine.
But competition existing isn't the same thing as the incumbent business suddenly becoming unprofitable or obsolete.
China successfully recovering a reusable booster is impressive too. SpaceX has already done that hundreds of times.
So I'd describe China as rapidly developing a competitor, not standing shoulder to shoulder with SpaceX today.
Otherwise we might as well declare every company doomed the moment somebody else opens a laboratory.
Last edited:
Where I think you're going too far is treating those losses as though we've already established they're permanent money pits.
There's also a problem with the X numbers. The IPO disclosures showed X advertising revenue alone at roughly $1.84 billion in 2025, not $350 million, before subscriptions and other revenue. The broader AI/X segment generated around $3.2 billion.
That segment still lost a fortune, so I'm certainly not pretending it's healthy today.
But there's an enormous difference between...“This operation currently loses billions while management invests heavily in it” and “This operation will always lose billions and become a white elephant.”
The first is supported by the financial statements.
The second is a forecast.
And that's really the entire investing argument we've been having. The question isn't whether xAI is profitable today. It clearly isn't. The question is whether the billions being spent today eventually produce a business worth substantially more than those expenditures.
Maybe they do. Maybe they don't.
That's the risk investors are purchasing.
As for Starlink competition reducing margins, 100% absolutely possible. That's something I would model into the valuation.
But “competition is coming, therefore profitability will decline” isn't automatic either. Markets can grow faster than competitors capture share, costs can decline, pricing can change and new services can expand revenue.
I'd definitely put a discount on SpaceX because of xAI's losses and Musk's habit of assembling his companies like corporate Lego.
I just wouldn't declare the Lego castle a white elephant before we've finished reading the instruction booklet.
Sources:
We finally know how much Elon Musk's X is making in ad revenue — and how that's changed over time
Elon Musk and X's efforts to win over advertisers after fights over content moderation haven't resulted in a bonanza.
www.businessinsider.com
SpaceX's IPO paperwork offers first look at the finances of Elon Musk's rocket company
SpaceX's S-1 paperwork, which pulls back the curtain on the rocket company's finances, is a required step ahead of its hotly anticipated IPO.
www.businessinsider.com
LordVile95
Posts: 1,033 +721
350 quarterly
It’s not really investment that’s the losses it’s the interest on the loans taken and that’s not actually paying off the loans.
Based on the fact AI isn’t scaling well with compute and open weight models popping up and being more useful to the individual it’s more and more likely that any data centre will have to be continually upgraded for diminishing returns to compete on the cutting edge of a sector with no actual profitable product. If you can run an open weight model on your phone why would you need anything else?
There’s only so much space to put things in and launching new satellites at the rate required seems unsustainable. Musk pushing to force a subscription to new Tesla’s also isn’t a good sign.
The instruction booklet is the same as always. Musk hypes something, the sheep follow and then the rug is pulled.I'd definitely put a discount on SpaceX because of xAI's losses and Musk's habit of assembling his companies like corporate Lego.
I just wouldn't declare the Lego castle a white elephant before we've finished reading the instruction booklet.
Sources:
We finally know how much Elon Musk's X is making in ad revenue — and how that's changed over time
Elon Musk and X's efforts to win over advertisers after fights over content moderation haven't resulted in a bonanza.
SpaceX's IPO paperwork offers first look at the finances of Elon Musk's rocket company
SpaceX's S-1 paperwork, which pulls back the curtain on the rocket company's finances, is a required step ahead of its hotly anticipated IPO.
There are several legitimate risks in there, but I think you're moving too quickly from “this could hurt the economics” to “therefore the economics will fail.”
Open weight models are absolutely putting pressure on frontier model pricing. That's real, and I mentioned essentially the same problem regarding Anthropic earlier.
But “if I can run a model on my phone, why would I need anything else?” treats radically different workloads as though they're interchangeable.
A small local model can be excellent for summarization, classification, simple assistants and plenty of everyday tasks. That doesn't mean it replaces large frontier systems doing complex coding, multimodal processing, long context reasoning, scientific workloads or enterprise scale automation.
In fact, the likely outcome seems increasingly hybrid...use the cheapest model capable of doing each job. That's bad for anyone assuming unlimited frontier model pricing power, but it isn't the same thing as eliminating demand for large scale compute.
There is also an accounting issue with your description of xAI's losses.
Interest expense certainly matters, but xAI's losses are not simply interest on loans. SpaceX's filing shows the AI/X segment lost roughly $6.4 billion at the operating level in 2025.
Operating income is calculated before interest expense.
So yes, there's debt and financing expense on top, but the underlying operation itself is presently spending substantially more than it earns.
That's exactly why we consider consider it risky.
Where we differ is that you're treating today's operating economics as their permanent state. I'm saying we don't know that yet.
Same with Starlink. Orbital capacity, satellite replacement and future competition are real constraints. But Starlink generated about $4.4 billion in operating income in 2025 while already replacing satellites and launching new ones. So saying replacement “seems unsustainable” needs more than the observation that satellites eventually wear out.
That's a cost the business model was designed around.
Maybe competition compresses those margins badly. Maybe replacement costs rise. Maybe Chinese constellations eventually take substantial market share. Those are risks I'd price in...They aren't accomplished facts.
And once we get to “Musk hypes something, sheep follow and then the rug is pulled,” we're no longer really discussing the financial statements.
That's a thesis about Elon Musk...Goal is trying to value businesses.
The nice thing about numbers is they don't care whether any of us like the CEO.
Steady on there with your nuanced takes!
Man, for some reason the comment section here sometimes makes it seem like we're on the luddite subforum of the democratic socialists website instead of a tech blog. It's just endless "capitalism/Elon Musk bad" and "AI bad/bubble and everyone but me is just sheeple". So clever.
Careful, that kind of nuance is liable to get you reported to the Department of Approved Internet Opinions.
Apparently we’ve reached the stage where saying ‘the valuation may be crazy but the underlying business is real’ counts as radical centrism.
Similar threads
