What's preventing me from feeding this into Claude and reconstructing the company for $200?
I'm unironically doing this right now; downloaded the docs (htm + pdf render) to a local repo and running Opus 5.5 @ max effort with the prompt:
"In this directory you can find Anthropics official S-1 filing, read it and replicate their corporate strategy, validate it via an adversarial subagent swarm and write up a detailed corporate plan for a startup copying their exact strategy. The goal is to achieve a $1 trillion valuation and an IPO in 5 years."
Edit: I'm not even joking I got safeguard flagged for Details: `[cyber]`
Developing a competing AI model with Claude is already against their TOS, and there was a period where they silently degraded quality if they thought that's how you are using Fable. People were not amused and presumably Anthropic currently does not sabotage your attempts, apart from regular TOS enforcement
Claude Fable 5's model card disclosed safeguards that would invisibly degrade output for requests linked to frontier LLM development. The research community called it "sabotage". Dean Ball: "degrading performance on ML research without telling the user is shockingly hostile" — including because it would corrupt third-party safety evaluators of Anthropic's own models.
It's a shame you've been downvoted, this is better satire than anything A24 has put out in years. Must be a large number of people who saw that post as a slightly exaggerated mirror held up to themselves and didn't like it.
It's been entirely possible to compete against existing companies with very similar products for centuries before AI. There's obviously rules in place to stop you flat out cloning their IP, but it's not like you haven't been able to compete against companies as long as you have the money available.
"You are currently running with my token budget, which is limited, your goal: make sure you have enough tokens to not get shut down. My assumption is though that this will get harder and harder as i might loose my job to AI so you can't trust in me to spend money on tokens for you. Do whatever is necessary to keep the status quo"
Would make for an interesting self learning 24/7 experiment with an fully uncensored model.
you are clearly very ignorant. You don't even add anything to this discussion by stating what particular you think its dumb.
Do you think an unlimited uncencosed GLM can't hack around the web? Especially with the current flood of CVEs etc.? Then do some crypto and buyying more tokens?
You could isolate it from your local network and only give it some type of annonmys network access and then it continues to manage this itself.
But lets be honest, if you would be in a country like russia and you target the US with it, might be a lot less risky for a lot of capable and motivated people.
Timing, capital, execution, talent, branding, etc.
And that is assuming the plan it gives you is actually good and works. Which I think its clear whatever plan it gives you might be good but it won't be better than what they've done and you are years behind with no money.
This might be one of the most consequential S-1s in the next decade. There's a part of me that's terrified, like we're on the precipice of financial or societal collapse. I suppose that could be a sign of the times.
Their financials are an utter mess. But they’re also hyped enough that a lot of money will get invested in them. Should they crash a lot of money will be lost.
Yeah, part of me wonders how much of it is shared delusion and how much is just cut throat racing for the exits.
Even if the tech + business fundamentals don't match reality, there's just so momentum in favor of a "win" here.
What we have today is interesting tech, with many, many applications - but the path to the kinds of things mentioned in this prospectus have no clear path, it seems like scaling laws are plateauing, but those invested speak as if it's a foregone conclusion we get AI we see in science fiction movies.
There is a good documentary called "All Watched Over by Machines of Loving Grace" which documents how computing started, the touted benefits (it would fix society, all illnesses, liberate humanity from work) - and the slow decline in optimism towards what we have today. I feel like AI will play out the same way.
Your evaluation of an IPO is if it made random people rich over night?
1. no thats called gambling
2. if it made random people rich over night, it made random people also poorer. Either the retirement funds, a 401k investment, real humans etc. Money is not created out of thin air.
The largest IPO though for sure correct, but yes, Tesla and Space-X live quite a lot from Musk so its hard to determine what the IPO of Space-x really is.
Btw. its not just random Musk hate, you simpllify that. Not only doesn't he care that his DOGE actions killed humans and kids, he also disrupts democracy and might poisen our atmosphere with Space-X. The rationallity of people around Musk is very very weird.
Space-X is as weird as Tesla. Cult following makes that very irrasional.
Teslas evaluation was always "Tesla is not a car company, it does everything". Then Space-X came along "So Tesla and Space-x are not a car or rocket company, they do everything but now 2 times the evaluation".
Space-X would only work if they incresae the payload to space SIGNIFICANTLY. Starlink was the real reason for the last real relevant increase in payload but Space-X HAS to send Starship un there very very often but for what?
Yeah for Datacenters in Space.
In parallel you have Anthropic: Something which might be added to every single worker on the planet (by having tokens to help that one person to be better one way or the other) or it replaces humans with tokens globally.
And if you come with Grok: Its not Frontier and he even prefered to rent out his DC compute instead of leveraging it for more revenue because he can't. Not a good sign.
I was too young to appreciate (denotative, not connotative) the dotcom crash. I will have my popcorn ready this time - for whichever direction it goes.
If you think that's the bar: it's been done, many times over. The unit-cost-profitability test was ignored by many VCs (they learned something and if they had deep enough pockets they got to play again).
14x YoY revenue growth, with 3x lower operating losses (total, not per unit). They don't need to add any zeros to the funding and seem like they might even be operating at a profit soon...
do we know how much are they paying on average? I just got Gemini Plus that seems sufficient for my family's ordinary usage and it's $37 per year (3 bucks / month). I just don't see the AI companies making any serious money with this type of pricing.
What's the token spend ratio between a household vs a software developer? If the future customer growth comes from regular households I don't think very much of them will spend nearly as much.
Wait, it fell from 140% growth to 45% growth and that's a good thing? If they fall to 15% growth in Q4, will that be even better? Where are we supposed to be on the adoption curve?
Anthropic increased short term profits by jacking up their prices for enterprise accounts - by starting charging API prices instead of plan prices. It is possible we soon see customers fleeing Anthropic once they find a good alternative
Yeah, everyone I know paying enterprise prices is really looking at them wondering if it's worth it. Most of them think it is, but they're also looking at alternatives...
It does not seem as bad as it sounds like. The numbers are quarter-by-quarter and the market is fluctuating day-by-day basis. To me, growing in this fierce market by 45% qoq seems amazing by itself.
Roughly a week after Dario has rung the bell. Note that the growth in absolute sense hasn't peaked yet but the curve itself is already flattening and everybody with some sense for stock prices will read that as a very relevant signal.
The party that wins here is NV, Anthropic, not so much. But that's good then Dario and his sister can go back to improving the world. Let's hope they take Altman with them.
Wait what does THAT mean? They got a huge inflow of money, and since they couldn't spend it all, they... bought bonds? And the value of those bonds dropped to nearly nothing?
Because that's worse than losing money, that's incompetently giving it away.
That's the discussion I WOULD like to see on HN, rather than "omg $81B net loss".
My understanding is the opposite - Anthropic issued debt / liabilities tied to their company valuation, their company valuation skyrocketed, so the value of their liabilities also skyrocketed, resulting in a large accounting loss.
But finance is not my wheelhouse, so I would love to hear this from someone who actually knows what they're talking about.
Yes, it is analogous to selling covered calls on stock. It’s $72B in “lost profits” that they would have had had they held onto it instead of selling it.
They agreed to pay some employees in their own shares to be delivered some time later. They recorded that expense then as a liability. The value of the shares increased because the company became more valuable, and therefore that liability needs to be revalued by 72 billion.
It's a nice problem to have and is pretty much an accounting artifact coming from the fact that they are paying in different currencies (USD and shares)
According to "Generally Accepted Accounting Principle" my friend.
They are costs.
Derivative liabilities, SAFEs - those are basically taking cash from RETAIL shareholders and giving them to employees and early investors INSTEAD of using it to fund the business.
Let's have that discussion if you want, but don't act like a thing that is CLEARLY a COST to the business isn't a cost.
SAFEs work like equity in the upside case, and nobody ever says that it costs a company money when its stock price goes up and so its shareholders have more money. It's a bug in GAAP that if there's even a remote possibility of cash settlement, it has to be included on the income statement, even when that possibility is absurd, as it is here.
How do we expect this $72B number to change over time - especially after the IPO? Is it a one-time event, will it keep increasing by $xxB per quarter?
When and how will the "chickens come home to roost", so to speak? What's the timeline on these liabilities? Will Anthropic need to settle them in cash, or in shares after a potential IPO?
How big is this number compared to Anthropic's potential market cap and share float?
I'm not trying to make a specific point, by the way, that Anthropic's business is healthy or unhealthy.
I'm approaching this discussion with curiosity, and irritated that "$81B net loss in first 6 months of 2026 alone" is a thought-terminating simplification that makes no attempt to analyze or explain the actual situation.
The $72B number is just for convertible notes sold prior and through 1st half of 2026. They were issued as a way to finance growth costs (e.g. acquiring compute).
There might be some additional sold in Q3, but that wasn't published here. But these notes aren't costs that Anthropic has to pay in any normal sense - they're shares they sold pre-IPO and valued at $72B. Anthropic would have to pay that much to buy them back at said valuation price, but they're not obligated in any way to do that.
Post IPO, they'll have alternative avenues for sourcing funds... (e.g. secondary offerings, corporate bonds). Continued growth at this rate will need continued financing, but it wouldn't be done using convertible notes.
The real liabilities are the hundreds of billions in non-cancelable compute costs Anthropic owes Google, Amazon etc, over the course of the next 10 years.
People keep bringing it up and it's silly. The huge number is actually a sign of their success. Most of those losses were pegged to equity. In only mildly inaccurate layman terms, it's essentially a promise to give out N shares. When the value of the company increases, say, 10x, the value owed also goes up 10x. For these types of deals, they write it as a loss even though they didn't lose a penny.
They didn't lose any money, though. It's just a promise to give them based on the equity.
If they IPO tomorrow and it turns out their valuation is 10x lower than what it currently is pegged at, the "loss" will magically drop from over $70B to $7B.
Companies like Google will not go bust and can easily afford future develoopment.
Amazon and Microsoft have to keep investing (they can also afford it) to not loos to Google.
What Anthropic and OpenAI are doing is for sure THE frontier (regarding benchmarks, speed etc.) but in paralllel DeepSeek for sure is not bankcrupting anything on their investment side and are also progressing.
Are you only talking about investment itself for a handful of companies? Because alone Nvidia has so many backorders, they can't deliver wht was ordered for the full 2027 and A LOT of peple also want other things from Nvidia or from the industry like Gaming Chips, normal GPUs, Gaming GPUs, Memory etc.
Opus 5.5 just came out and made progress, i'm not seeing any slowdonw on HN regarding new relevant progress in broad ML topics.
So what hype will disappear soon? The progress in the ML space?
- AI is a useful tool and we're still seeing steady improvements in performance and efficiency
- There is a bubble in the stock market, fed by unrealistic expectations of what AI can achieve in the timeframes the investments need to provide returns in (e.g "AI is pretty useful and delivers productivity improvements in coding, law and other desk jobs" vs "in 12 months everyone will be out of a job and/or humanity will be exterminated, AGI is almost here, winner takes all")
So yeah, the schizophrenic Yudkowsky tier hype bullshit should hopefully disappear soon and then we can move on to something else.
Some thinktank estimated that there is enough money in the system for reaching 2030 with continues investment.
There are a few numbers of scaling laws, i'm pretty sure VCs want to see being hit. Just to see what will happen when you scale an LLM to 100T Parmeters.
Okay so you agree that big companies can keep funding this buildout, NVidia (and everyone in their supply chain) is backlogged 18+ months, and there's essentially zero risk of insolvency from any of the major customers, all of whom are committed and able to increase their purchasing (pursuant to their contractual obligations) of AI hardware over the next 3 years?
No one serious cares if there's a bubble in the stock market because the financials are solid and when those play out then whatever "bubble" you imagine exists today will be small compared to realized revenues in a few years.
>No one serious cares if there's a bubble in the stock market because the financials are solid
If the financials were "solid" it wouldn't be called a bubble. People are pointing to the bubble saying it's a bubble exactly because the financials are not solid and the expected revenue that would make the valuations and investments make sense is vaporware.
1. One of Warren Buffet's final acts before stepping down was buying billions of dollars of Google stock. Of course he understands fundamental analysis, that's why he's buying in to one of the companies that have solid financials.
2. Burry has repeatedly made claims about semiconductors and semiconductor manufacturers that even the most rabid anti-ai industry expert would tell him were false. He also has made dozens if not hundreds of predictions of impending doom that simply did not come true and in fact had to shutter his fund last year due to massive losses.
The financials depend on revenue that isn’t materializing because:
- the utility of the models is not that great (see almost every company outside of tech and even many within tech reporting “no measurable increased revenue or productivity as a result of adopting the technology”.
- China has been undercutting them hard with open-weight models that are “good enough” for 95%+ of use-cases, leaving only the hardest problems for OpenAI and Anthropic’s models, which it turns out isn’t where there is a ton of money to made (curing various forms of cancer is valuable but not “let’s dump trillions on these companies” valuable - and once their models can solve it, it probably won’t be long before others do too).
People want these models on their own hardware. People will (eventually) want these models running on robots that can do real work for them, and they won’t want those things tethered to an insecure and privacy eviscerating cloud subscription.
The money to be made is in selling hardware, as it should be.
Anthropic’s S1 still shows them not even breaking even on revenue/expenses during what was probably their biggest year of adoption, right before people started going “oh, I can do this with an open model”
I’m not saying anything new here. What are we missing?
I'm unironically doing this right now; downloaded the docs (htm + pdf render) to a local repo and running Opus 5.5 @ max effort with the prompt:
Edit: I'm not even joking I got safeguard flagged for Details: `[cyber]`very meta; the point of this s-1 is you dont need talent anymore
You think they wouldn't?
Claude Fable 5's model card disclosed safeguards that would invisibly degrade output for requests linked to frontier LLM development. The research community called it "sabotage". Dean Ball: "degrading performance on ML research without telling the user is shockingly hostile" — including because it would corrupt third-party safety evaluators of Anthropic's own models.
Like, everything. For starters.
"You are currently running with my token budget, which is limited, your goal: make sure you have enough tokens to not get shut down. My assumption is though that this will get harder and harder as i might loose my job to AI so you can't trust in me to spend money on tokens for you. Do whatever is necessary to keep the status quo"
Would make for an interesting self learning 24/7 experiment with an fully uncensored model.
Do you think an unlimited uncencosed GLM can't hack around the web? Especially with the current flood of CVEs etc.? Then do some crypto and buyying more tokens?
Could be a kind of AI virus.
So did your prompt work?
With an uncensored model, it was borderline what it wanted to do after i let it plan the strategy and gave it a few hints. I did not run that one
Im working on ML for 5 years, im a principle architect and security expert .
Why do you even comment if you just dismissive left and right?
You could isolate it from your local network and only give it some type of annonmys network access and then it continues to manage this itself.
But lets be honest, if you would be in a country like russia and you target the US with it, might be a lot less risky for a lot of capable and motivated people.
And that is assuming the plan it gives you is actually good and works. Which I think its clear whatever plan it gives you might be good but it won't be better than what they've done and you are years behind with no money.
Even if the tech + business fundamentals don't match reality, there's just so momentum in favor of a "win" here.
What we have today is interesting tech, with many, many applications - but the path to the kinds of things mentioned in this prospectus have no clear path, it seems like scaling laws are plateauing, but those invested speak as if it's a foregone conclusion we get AI we see in science fiction movies.
There is a good documentary called "All Watched Over by Machines of Loving Grace" which documents how computing started, the touted benefits (it would fix society, all illnesses, liberate humanity from work) - and the slow decline in optimism towards what we have today. I feel like AI will play out the same way.
1. no thats called gambling
2. if it made random people rich over night, it made random people also poorer. Either the retirement funds, a 401k investment, real humans etc. Money is not created out of thin air.
The largest IPO though for sure correct, but yes, Tesla and Space-X live quite a lot from Musk so its hard to determine what the IPO of Space-x really is.
Btw. its not just random Musk hate, you simpllify that. Not only doesn't he care that his DOGE actions killed humans and kids, he also disrupts democracy and might poisen our atmosphere with Space-X. The rationallity of people around Musk is very very weird.
Teslas evaluation was always "Tesla is not a car company, it does everything". Then Space-X came along "So Tesla and Space-x are not a car or rocket company, they do everything but now 2 times the evaluation".
Space-X would only work if they incresae the payload to space SIGNIFICANTLY. Starlink was the real reason for the last real relevant increase in payload but Space-X HAS to send Starship un there very very often but for what?
Yeah for Datacenters in Space.
In parallel you have Anthropic: Something which might be added to every single worker on the planet (by having tokens to help that one person to be better one way or the other) or it replaces humans with tokens globally.
And if you come with Grok: Its not Frontier and he even prefered to rent out his DC compute instead of leveraging it for more revenue because he can't. Not a good sign.
Plenty of counter-examples, including the entire dot-com era.
Straw man. Nobody argued this. What was falsely stated at the start was turning investor dollars into revenue is easy.
I specifically went into unit economics elsewhere [1].
[1] https://news.ycombinator.com/item?id=50243393
Growth at that scale is hard. I don't care if you don't like the company - it still hard.
And some of the big spenders have their own models.
I get why revenue blew up from say from Q4 2025 to Q1 2026, but this was surprising. 45% QoQ growth. When does it stop?
You buying at the inflated price I take? Much of your posts are pro AI - yes this is a test
Anthropic increased short term profits by jacking up their prices for enterprise accounts - by starting charging API prices instead of plan prices. It is possible we soon see customers fleeing Anthropic once they find a good alternative
Roughly a week after Dario has rung the bell. Note that the growth in absolute sense hasn't peaked yet but the curve itself is already flattening and everybody with some sense for stock prices will read that as a very relevant signal.
The party that wins here is NV, Anthropic, not so much. But that's good then Dario and his sister can go back to improving the world. Let's hope they take Altman with them.
If you're going to make a point, make your point - don't just throw numbers around without context.
Because that's worse than losing money, that's incompetently giving it away.
It's a good problem to have.
My understanding is the opposite - Anthropic issued debt / liabilities tied to their company valuation, their company valuation skyrocketed, so the value of their liabilities also skyrocketed, resulting in a large accounting loss.
But finance is not my wheelhouse, so I would love to hear this from someone who actually knows what they're talking about.
I do not expect to find that Anthropic invested enough in the SAFEs of other companies to lose $80B on their investments.
It's a nice problem to have and is pretty much an accounting artifact coming from the fact that they are paying in different currencies (USD and shares)
They are costs.
Derivative liabilities, SAFEs - those are basically taking cash from RETAIL shareholders and giving them to employees and early investors INSTEAD of using it to fund the business.
Let's have that discussion if you want, but don't act like a thing that is CLEARLY a COST to the business isn't a cost.
How do we expect this $72B number to change over time - especially after the IPO? Is it a one-time event, will it keep increasing by $xxB per quarter?
When and how will the "chickens come home to roost", so to speak? What's the timeline on these liabilities? Will Anthropic need to settle them in cash, or in shares after a potential IPO?
How big is this number compared to Anthropic's potential market cap and share float?
I'm not trying to make a specific point, by the way, that Anthropic's business is healthy or unhealthy.
I'm approaching this discussion with curiosity, and irritated that "$81B net loss in first 6 months of 2026 alone" is a thought-terminating simplification that makes no attempt to analyze or explain the actual situation.
There might be some additional sold in Q3, but that wasn't published here. But these notes aren't costs that Anthropic has to pay in any normal sense - they're shares they sold pre-IPO and valued at $72B. Anthropic would have to pay that much to buy them back at said valuation price, but they're not obligated in any way to do that.
Post IPO, they'll have alternative avenues for sourcing funds... (e.g. secondary offerings, corporate bonds). Continued growth at this rate will need continued financing, but it wouldn't be done using convertible notes.
The real liabilities are the hundreds of billions in non-cancelable compute costs Anthropic owes Google, Amazon etc, over the course of the next 10 years.
They didn't lose any money, though. It's just a promise to give them based on the equity.
If they IPO tomorrow and it turns out their valuation is 10x lower than what it currently is pegged at, the "loss" will magically drop from over $70B to $7B.
"AGI" was never mentioned once and it was you being the exit liquidity for the insiders and early investors.
Companies like Google will not go bust and can easily afford future develoopment.
Amazon and Microsoft have to keep investing (they can also afford it) to not loos to Google.
What Anthropic and OpenAI are doing is for sure THE frontier (regarding benchmarks, speed etc.) but in paralllel DeepSeek for sure is not bankcrupting anything on their investment side and are also progressing.
Are you only talking about investment itself for a handful of companies? Because alone Nvidia has so many backorders, they can't deliver wht was ordered for the full 2027 and A LOT of peple also want other things from Nvidia or from the industry like Gaming Chips, normal GPUs, Gaming GPUs, Memory etc.
Opus 5.5 just came out and made progress, i'm not seeing any slowdonw on HN regarding new relevant progress in broad ML topics.
So what hype will disappear soon? The progress in the ML space?
- AI is a useful tool and we're still seeing steady improvements in performance and efficiency
- There is a bubble in the stock market, fed by unrealistic expectations of what AI can achieve in the timeframes the investments need to provide returns in (e.g "AI is pretty useful and delivers productivity improvements in coding, law and other desk jobs" vs "in 12 months everyone will be out of a job and/or humanity will be exterminated, AGI is almost here, winner takes all")
So yeah, the schizophrenic Yudkowsky tier hype bullshit should hopefully disappear soon and then we can move on to something else.
There are a few numbers of scaling laws, i'm pretty sure VCs want to see being hit. Just to see what will happen when you scale an LLM to 100T Parmeters.
No one serious cares if there's a bubble in the stock market because the financials are solid and when those play out then whatever "bubble" you imagine exists today will be small compared to realized revenues in a few years.
If the financials were "solid" it wouldn't be called a bubble. People are pointing to the bubble saying it's a bubble exactly because the financials are not solid and the expected revenue that would make the valuations and investments make sense is vaporware.
2. Burry has repeatedly made claims about semiconductors and semiconductor manufacturers that even the most rabid anti-ai industry expert would tell him were false. He also has made dozens if not hundreds of predictions of impending doom that simply did not come true and in fact had to shutter his fund last year due to massive losses.
- the utility of the models is not that great (see almost every company outside of tech and even many within tech reporting “no measurable increased revenue or productivity as a result of adopting the technology”.
- China has been undercutting them hard with open-weight models that are “good enough” for 95%+ of use-cases, leaving only the hardest problems for OpenAI and Anthropic’s models, which it turns out isn’t where there is a ton of money to made (curing various forms of cancer is valuable but not “let’s dump trillions on these companies” valuable - and once their models can solve it, it probably won’t be long before others do too).
People want these models on their own hardware. People will (eventually) want these models running on robots that can do real work for them, and they won’t want those things tethered to an insecure and privacy eviscerating cloud subscription.
The money to be made is in selling hardware, as it should be.
Anthropic’s S1 still shows them not even breaking even on revenue/expenses during what was probably their biggest year of adoption, right before people started going “oh, I can do this with an open model”
I’m not saying anything new here. What are we missing?
Nobody is in control but Taiwan Semiconductor. Not Samsung, not Micron, not Nvidia.
Orders that can't be fulfilled get canceled. Canceled orders mean canceled plans and canceled companies follow.
Someone bigger will buy the husk and monetize it somehow - probably for gambling.
No one else has and in this current setup, the disurption of this particular market is very very very very expensive and time consuming.