>However, Anthropic’s disclosure to KBRA excluded several important financial metrics, including free cash flow and earnings before interest, taxes, depreciation and amortisation. Credit rating analysts had to come up with their own estimates based on assumptions, the documents show.
>Given the limited information, KBRA produced a “risk score” to assess Anthropic’s creditworthiness rather than the traditional full financial models.
This and the fact they went to a small ratings agency does not inspire confidence in the rating.
No, but a trillion dollar company going to a ratings agency to get an assessment of its credit rating and not giving that ratings agency enough information to do that doesn’t inspire confidence in that trillion dollar company, either.
Why didn’t they give them the info? They either withheld it or don’t have it. Both make them look bad.
that's a bigger red flag. there may be reasons (bad) to not give them all the info, but going to a smaller agency is the largest of red flags when you are that big of a company.
No, the title is poorly phrased, I think you misunderstood.
All ratings agencies have their own scales, but broadly all ratings fall into one of two overarching categories, investment grade vs. junk.
The article explains that the rating ANT received was the highest within the junk range. So “highest junk rating” means within the junk category, it’s in the best sub-area. In other words, it’s just scratching below the investment grade rating.
This is important for institutional investors who can only buy investment grade, meaning if they can improve even one single grade up, they get access to a lot more debt.
You can hold both things in your mind and they aren't contradictory. It can be a bad system but it can still be suspicious if someone behaves outside the norm inside it.
First SpaceX, now Anthropic. I wonder what happened the last time the big rating agencies gave massive amounts of sketchy debt investment-grade ratings to avoid alienating their customers?
"The rating action, which has not previously been reported, shows Anthropic is close to achieving the investment-grade status that would allow the start-up to lower its borrowing costs and tap a wider pool of institutional investors including life insurance companies and pension funds.
Bankers for Anthropic and rival OpenAI have lobbied the largest rating agencies to grant the AI labs investment-grade ratings once they go public."
The point of the headline isn’t to communicate that their debt is currently rated junk, it’s to communicate that it’s impossible for their rating to improve any further without becoming investment grade.
"The rating action, which has not previously been reported, shows Anthropic is close to achieving the investment-grade status that would allow the start-up to lower its borrowing costs and tap a wider pool of institutional investors including life insurance companies and pension funds.
Bankers for Anthropic and rival OpenAI have lobbied the largest rating agencies to grant the AI labs investment-grade ratings once they go public."
Bonds and a high risk high reward business model feel like a bad fit.
Like if there is a significant risk of bankruptcy then there must on the other side of the scale be a possibility of making bank like you have with stock or derivatives.
Subprime loans blew up because it was a bunch of opaque securities (mortgages sliced and diced together) that lenders thought were safe, but really weren't. Anthropic debt doesn't have this issue because everyone knows about anthopric and their business model.
GPU backed securities also don't have that issue because they're far more transparent. The problem with mortgage backed securities is that nobody has the time to inspect the credit worthiness of each mortgage in a MBS, each creditor has hidden correlated risk (eg. if the whole economy tanks), and on top of all that trenching obfuscates the risk. Lenders therefore had to take the credit ratings on face value. It's hard to make the mistake with datecenter/gpu/anthropic debt.
> The problem with mortgage backed securities is that nobody has the time to inspect the credit worthiness of each mortgage in a MBS
Credit worthiness of each mortgage wasn't the issue with mortgage-backed securities. It was that low value mortgage-backed securities were repackaged into a basket with a mix of securities and given a higher credit rating than what Subprime would have gotten on their own.
This repackaging is absolutely going to happen again with GPUs. If there are GPUs of different generations within the same security, it's already happened.
> Bankers for Anthropic and rival OpenAI have lobbied the largest rating agencies to grant the AI labs investment-grade ratings once they go public. Investors expect the initial public offering this year to hit a valuation of more than $2tn.
Let’s be clear about what this means: it gives them access to the corporate bond market. What they’re lobbying for is that pension funds, life insurance, etc., can take exposure to these companies. Suddenly the everyman, your mom and dad, have significant exposure to these companies and their circular financing deals without even knowing it. This is disgusting.
What a lot of people fail to realize is that the rules around IPOs changed since the dot com bubble. Previously, employees counted as shareholders and companies above a certain threshold of shareholders were required to IPO. This is why companies like Microsoft, Amazon, etc., had the majority of their growth while being publicly traded. This allowed the everyman with a brokerage account to participate in these gains.
These rules were essentially removed by Obama in 2012 through the JOBS act. What this means is that companies can stay private for a majority of their growth phase, which is why we’re seeing IPOs being floated with $2T dollar amounts on them. They’ve privatized the gains, and they only IPO once they want to cash out.
So let’s understand clearly what’s about to happen: early VC investors in oai/ant are going to make billions selling dogshit to your parents pension funds, your parents pension funds are run by idiots who will gleefully take the AI exposure and are legally allowed to take that AI exposure because the AI companies lobbied the credit rating agencies to allow it, and when the bills come next year and these companies are completely unable to meet their liabilities, your parents have to postpone their retirements or come back out of retirement to make ends meet.
It’s truly evil. I hope the employees at these companies with dollar signs in their eyes awaiting their multimillion dollar payoff understand the system that they’re participating in and how evil it is.
Point blank: 2008 was about manipulation of the credit rating agencies. This bubble will also be about manipulation of the credit rating agencies. Absolutely deplorable.
And no, just to nip this in the bud, AI isn’t a fad. It’s the most transformative technology of our lifetimes. It’s going to change everything. But that doesn’t have anything to do with chaos in financial markets or malinvestment during the gold rush. It has to do with whether these companies have any kind of moat and whether the billions being poured into current generation chips are going to depreciate over the timeframes that investors think they will.
Because for a pension fund the exposure will be much smaller if they're an early investor in a 100m Business than being a late investor in 2T business.
Basically when you invest at the peak of the bubble you're going to be hosed.
I don’t like that they’re fleecing the corporate bond market after privatizing the majority of their upside. Yes, I wanted them to issue their junk bonds even earlier.
What I really want above all else is for credit rating agencies to not face pressure to rerate them favorably because they’re deemed too big to fail.
How are they fleecing anybody if their rating is junk?
They can lobby all they want but until we see ratings agencies bending to their lobbying in the face of extreme scrutiny from anyone who would buy these corporate bonds I wouldn't go about worrying too much.
And if default risk rises, yes the value of those bonds will get crushed. But Id argue the value of equity will be more so - since they are not A) positive EBIT B) EBITDA
Those two numbers are variables that compose measures of default risk.
> Let’s be clear about what this means: it gives them access to the corporate bond market. What they’re lobbying for is that pension funds, life insurance, etc., can take exposure to these companies. Suddenly the everyman, your mom and dad, have significant exposure to these companies and their circular financing deals without even knowing it. This is disgusting.
What's the problem? Fund managers—who are in a position to understand the risks more than your everyday investor—can decide if they want buy these bonds or not, just like any other public corporate bond.
Most pensions and life insurers and so on are mandate constrained to investment grade only.
2008 wasn't fund managers deciding all together in unison to buy bad mortgages. It was AAA ratings on diversified MBS that should have been rated much lower giving them cover to load up.
I just wish we could invest in the Chinese companies without the USG coming in later and claiming it's illegal to invest in them and seize the investment. It's clear when the US models financially crumble, everyone will jump on the Chinese models. (https://www.foxrothschild.com/publications/investments-in-ch...)
What's not clear is whether the west can build a separate but equal ecosystem here when VC only funds privatized models, but it was hilarious when some influencer went on and on that Chinese open weight models are unsafe because we don't have access to their training code and data. Serious WTF territory IMO.
It's because their datacenter businesses are growing so fast YoY. Why do you care about the difference in a few percentage points of interest if you're getting 37, 41, or even 82% returns YoY?
Correct hyperscalars are doing fine - as it stands.
The problem is OAI/Anthropic - but if OAI and Anthropic fail the whole thing comes apart. Goog, Meta etc will be fine as they had tremendous cash flow generating operations prior to datacenter offerings w.r.t LLMs.
If OpenAI and Anthropic somehow become insolvent to the point that they are unable to continue running inference on their existing trained models, their tens of billions of dollars of monthly revenue will need to go somewhere.
Every now and then I read articles from the Financial Times just to confirm that they are still a terrible source of information.
BB+ is technically junk. But the it is one notch from investment grade. Also Junk ratings are normal for young. So the framing is sensationalized.
"Growth has slowed" is technically true but misleading. Q2 to Q3 still went from $11.5bn to $16.8bn. And at those levels, it’s not wildly unexpected that it has to slow down.
And the bias is obvious when they write "Despite limited disclosures" which makes it sounds suspicious, when it’s normal thatrivate companies don't publish financials.
It’s also technical errors. Broadcom agreed to support the residual value of the chips, they did not guarantee the loan.
When journalists and news organizations are this terrible at basic reporting, Within their own specialized domain, it's hard to find sympathy for their complaints about AI-generated content taking over.
Edit: after another look I found that the agency that gave the rating also is a new small one with not very good record eighter (Kroll Bond Rating Agency).
Alternative title: Small Agency, which is previously has been fined for Ratings Without Method, Guesses that Anthropic's Numbers are close to normal for startups; FT Reports the Guess as News.
‘Growth has slowed" is technically true but misleading. Q2 to Q3 still went from $11.5bn to $16.8bn. And at those levels, it’s not wildly unexpected that it has to slow down.’
The growth slowdown is decidedly unexpected. The market priced in a $2T valuation that assumed it wouldn’t slow down.
From the article:
>However, Anthropic’s disclosure to KBRA excluded several important financial metrics, including free cash flow and earnings before interest, taxes, depreciation and amortisation. Credit rating analysts had to come up with their own estimates based on assumptions, the documents show.
>Given the limited information, KBRA produced a “risk score” to assess Anthropic’s creditworthiness rather than the traditional full financial models.
This and the fact they went to a small ratings agency does not inspire confidence in the rating.
No, but a trillion dollar company going to a ratings agency to get an assessment of its credit rating and not giving that ratings agency enough information to do that doesn’t inspire confidence in that trillion dollar company, either.
Why didn’t they give them the info? They either withheld it or don’t have it. Both make them look bad.
>they went to a small ratings agency
that's a bigger red flag. there may be reasons (bad) to not give them all the info, but going to a smaller agency is the largest of red flags when you are that big of a company.
[delayed]
so Antropic hid all the bad info, went to more loyal agency...
...and score still showed "high probability of being junk"?
No, the title is poorly phrased, I think you misunderstood.
All ratings agencies have their own scales, but broadly all ratings fall into one of two overarching categories, investment grade vs. junk.
The article explains that the rating ANT received was the highest within the junk range. So “highest junk rating” means within the junk category, it’s in the best sub-area. In other words, it’s just scratching below the investment grade rating.
This is important for institutional investors who can only buy investment grade, meaning if they can improve even one single grade up, they get access to a lot more debt.
honestly not changing my overall feeling that credit ratings are junk science
You can hold both things in your mind and they aren't contradictory. It can be a bad system but it can still be suspicious if someone behaves outside the norm inside it.
but, but... HN told me that Anthropic is profitable. Their margins are supposedly insane!
First SpaceX, now Anthropic. I wonder what happened the last time the big rating agencies gave massive amounts of sketchy debt investment-grade ratings to avoid alienating their customers?
From 2007-2008? Taxpayers bailed out the perps so that they could keep their bonuses.
Junk ratings are not investment grade. They’re rating this debt as at least a little bit sketchy.
From the article:
"The rating action, which has not previously been reported, shows Anthropic is close to achieving the investment-grade status that would allow the start-up to lower its borrowing costs and tap a wider pool of institutional investors including life insurance companies and pension funds.
Bankers for Anthropic and rival OpenAI have lobbied the largest rating agencies to grant the AI labs investment-grade ratings once they go public."
It's not a big ratings agency, and it's not investment grade.
The point of the headline isn’t to communicate that their debt is currently rated junk, it’s to communicate that it’s impossible for their rating to improve any further without becoming investment grade.
From the article:
"The rating action, which has not previously been reported, shows Anthropic is close to achieving the investment-grade status that would allow the start-up to lower its borrowing costs and tap a wider pool of institutional investors including life insurance companies and pension funds.
Bankers for Anthropic and rival OpenAI have lobbied the largest rating agencies to grant the AI labs investment-grade ratings once they go public."
This is not particularly newsworthy. Kroll is not one of the main ratings agencies, and certainly not the one bond purchasers use.
Isn't the "bad" signal here use of such an outlier agency to begin with?
As in "main ratings agencies" see no reason to bother
It’s a fair point. It’s certainly not good but that doesnt = bad. It could be an artifact of FT deciding to report on a nothingburger.
I’d reserve judgement until i see them report more detailed financials and the big ratings agencies step in with assessment.
Bonds and a high risk high reward business model feel like a bad fit.
Like if there is a significant risk of bankruptcy then there must on the other side of the scale be a possibility of making bank like you have with stock or derivatives.
Starting to sound like subprime loans to me...
Subprime loans blew up because it was a bunch of opaque securities (mortgages sliced and diced together) that lenders thought were safe, but really weren't. Anthropic debt doesn't have this issue because everyone knows about anthopric and their business model.
As long as we aren't creating GPU-backed securities, we should be fine...oh wait https://nvidianews.nvidia.com/news/nvidia-partners-with-apol...
GPU backed securities also don't have that issue because they're far more transparent. The problem with mortgage backed securities is that nobody has the time to inspect the credit worthiness of each mortgage in a MBS, each creditor has hidden correlated risk (eg. if the whole economy tanks), and on top of all that trenching obfuscates the risk. Lenders therefore had to take the credit ratings on face value. It's hard to make the mistake with datecenter/gpu/anthropic debt.
> The problem with mortgage backed securities is that nobody has the time to inspect the credit worthiness of each mortgage in a MBS
Credit worthiness of each mortgage wasn't the issue with mortgage-backed securities. It was that low value mortgage-backed securities were repackaged into a basket with a mix of securities and given a higher credit rating than what Subprime would have gotten on their own.
This repackaging is absolutely going to happen again with GPUs. If there are GPUs of different generations within the same security, it's already happened.
Non-paywall link: <https://finance.yahoo.com/technology/ai/articles/anthropic-s...>
Thanks, very helpful. I was shocked to read they are securing $35B for chips and then to read it's buying chips from alphabet/Broadcom.
Obligatory Arrested Development reference: Upgraded from Triple-Sell to Don't Buy (https://www.youtube.com/watch?v=qK5nqmHz0UA)
> Bankers for Anthropic and rival OpenAI have lobbied the largest rating agencies to grant the AI labs investment-grade ratings once they go public. Investors expect the initial public offering this year to hit a valuation of more than $2tn.
Let’s be clear about what this means: it gives them access to the corporate bond market. What they’re lobbying for is that pension funds, life insurance, etc., can take exposure to these companies. Suddenly the everyman, your mom and dad, have significant exposure to these companies and their circular financing deals without even knowing it. This is disgusting.
What a lot of people fail to realize is that the rules around IPOs changed since the dot com bubble. Previously, employees counted as shareholders and companies above a certain threshold of shareholders were required to IPO. This is why companies like Microsoft, Amazon, etc., had the majority of their growth while being publicly traded. This allowed the everyman with a brokerage account to participate in these gains.
These rules were essentially removed by Obama in 2012 through the JOBS act. What this means is that companies can stay private for a majority of their growth phase, which is why we’re seeing IPOs being floated with $2T dollar amounts on them. They’ve privatized the gains, and they only IPO once they want to cash out.
So let’s understand clearly what’s about to happen: early VC investors in oai/ant are going to make billions selling dogshit to your parents pension funds, your parents pension funds are run by idiots who will gleefully take the AI exposure and are legally allowed to take that AI exposure because the AI companies lobbied the credit rating agencies to allow it, and when the bills come next year and these companies are completely unable to meet their liabilities, your parents have to postpone their retirements or come back out of retirement to make ends meet.
It’s truly evil. I hope the employees at these companies with dollar signs in their eyes awaiting their multimillion dollar payoff understand the system that they’re participating in and how evil it is.
Point blank: 2008 was about manipulation of the credit rating agencies. This bubble will also be about manipulation of the credit rating agencies. Absolutely deplorable.
And no, just to nip this in the bud, AI isn’t a fad. It’s the most transformative technology of our lifetimes. It’s going to change everything. But that doesn’t have anything to do with chaos in financial markets or malinvestment during the gold rush. It has to do with whether these companies have any kind of moat and whether the billions being poured into current generation chips are going to depreciate over the timeframes that investors think they will.
I don't understand your position.
First you say you don't like that they're accessing the corporate bond market, which is something all public companies can do.
Then you say they should've IPO'd earlier when they were even earlier stage and riskier .. which would allow them to issue bonds even earlier?
What is your position? Or are you complaining for the sake of complaining?
Because for a pension fund the exposure will be much smaller if they're an early investor in a 100m Business than being a late investor in 2T business.
Basically when you invest at the peak of the bubble you're going to be hosed.
I don’t like that they’re fleecing the corporate bond market after privatizing the majority of their upside. Yes, I wanted them to issue their junk bonds even earlier.
What I really want above all else is for credit rating agencies to not face pressure to rerate them favorably because they’re deemed too big to fail.
How are they fleecing anybody if their rating is junk?
They can lobby all they want but until we see ratings agencies bending to their lobbying in the face of extreme scrutiny from anyone who would buy these corporate bonds I wouldn't go about worrying too much.
Its rated as junk.
And if default risk rises, yes the value of those bonds will get crushed. But Id argue the value of equity will be more so - since they are not A) positive EBIT B) EBITDA
Those two numbers are variables that compose measures of default risk.
Calm your panties down. Kroll != S&P.
> Let’s be clear about what this means: it gives them access to the corporate bond market. What they’re lobbying for is that pension funds, life insurance, etc., can take exposure to these companies. Suddenly the everyman, your mom and dad, have significant exposure to these companies and their circular financing deals without even knowing it. This is disgusting.
What's the problem? Fund managers—who are in a position to understand the risks more than your everyday investor—can decide if they want buy these bonds or not, just like any other public corporate bond.
Most pensions and life insurers and so on are mandate constrained to investment grade only.
2008 wasn't fund managers deciding all together in unison to buy bad mortgages. It was AAA ratings on diversified MBS that should have been rated much lower giving them cover to load up.
Someone working for a fund and whose job is to INVEST is going to loose their job for NOT buying into the ponzi scheme that is running 20% growth YoY
CoreWeave is making the same case for its bonds.
And CoreWeave's stocks are trending down due to concerns over its debts.
[1] https://www.wsj.com/livecoverage/stock-market-today-dow-sp-5...
[2] https://www.tradingview.com/news/seekingalpha:173ffc32a094b:...
[3] https://www.cnbc.com/2026/10/15/coreweave-credit-default-swa...
Correct - there are no freebies here!
You can't really rip off bond-holders when you are not positive EBIT and EBITDA...
I just wish we could invest in the Chinese companies without the USG coming in later and claiming it's illegal to invest in them and seize the investment. It's clear when the US models financially crumble, everyone will jump on the Chinese models. (https://www.foxrothschild.com/publications/investments-in-ch...)
What's not clear is whether the west can build a separate but equal ecosystem here when VC only funds privatized models, but it was hilarious when some influencer went on and on that Chinese open weight models are unsafe because we don't have access to their training code and data. Serious WTF territory IMO.
It's a nice thought but so far none of the hyperscalers seem to give a single fuck about how much the cash costs to borrow.
It's because their datacenter businesses are growing so fast YoY. Why do you care about the difference in a few percentage points of interest if you're getting 37, 41, or even 82% returns YoY?
Correct hyperscalars are doing fine - as it stands.
The problem is OAI/Anthropic - but if OAI and Anthropic fail the whole thing comes apart. Goog, Meta etc will be fine as they had tremendous cash flow generating operations prior to datacenter offerings w.r.t LLMs.
If OpenAI and Anthropic somehow become insolvent to the point that they are unable to continue running inference on their existing trained models, their tens of billions of dollars of monthly revenue will need to go somewhere.
Every now and then I read articles from the Financial Times just to confirm that they are still a terrible source of information.
BB+ is technically junk. But the it is one notch from investment grade. Also Junk ratings are normal for young. So the framing is sensationalized.
"Growth has slowed" is technically true but misleading. Q2 to Q3 still went from $11.5bn to $16.8bn. And at those levels, it’s not wildly unexpected that it has to slow down.
And the bias is obvious when they write "Despite limited disclosures" which makes it sounds suspicious, when it’s normal thatrivate companies don't publish financials.
It’s also technical errors. Broadcom agreed to support the residual value of the chips, they did not guarantee the loan.
When journalists and news organizations are this terrible at basic reporting, Within their own specialized domain, it's hard to find sympathy for their complaints about AI-generated content taking over.
Edit: after another look I found that the agency that gave the rating also is a new small one with not very good record eighter (Kroll Bond Rating Agency).
Alternative title: Small Agency, which is previously has been fined for Ratings Without Method, Guesses that Anthropic's Numbers are close to normal for startups; FT Reports the Guess as News.
‘Growth has slowed" is technically true but misleading. Q2 to Q3 still went from $11.5bn to $16.8bn. And at those levels, it’s not wildly unexpected that it has to slow down.’
The growth slowdown is decidedly unexpected. The market priced in a $2T valuation that assumed it wouldn’t slow down.
https://www.reuters.com/business/finance/investors-temper-an...