The More You Buy, The More You Lose
We recorded 27 claims from this post. 8 were checked against evidence. 19 can’t be checked. 23 more are held. Read the post at its source.
Claims from this post (27)
In the order they appear in the post. Each quotation is the speaker’s words in this post. A label such as “Supported” or “Can’t be checked” is the claim’s current assessment; its page has the evidence and the reasons.
Assessment: UnresolvedSaid
Forecast
“Hyperscalers will have sunk over $1.3 trillion dollars into generative AI by the end of 2026”
Our reading: By the end of 2026, the hyperscalers (Meta, Google, Amazon, and Microsoft) will have sunk over $1.3 trillion into generative AI since March 2022.
Not yet due Evidence as of
Limitation: The measure of money 'sunk into generative AI' is not stated, so the record may not be scorable after the deadline either.
Assessment: UnresolvedSaid
Forecast
“This is a problem, as all four are certain to become repeat visitors to the bond markets.”
Our reading: Meta, Google, Amazon and Microsoft are certain to become repeat visitors to the bond markets.
No deadline stated Evidence as of
Limitation: Several primary filings were inaccessible to read_page.
Assessment: UnresolvedSaid
Causal claim
“hyperscalers have borrowed a bunch of money to fund AI, to the point that it’s pushing them into cash-flow negative territory, and every time they raise more money, bondholders become more worried about them paying it back, especially given that both Amazon and Google have now gone cash-flow negative as of their latest quarters.”
Our reading: Hyperscalers' AI borrowing is pushing them cash-flow negative (Amazon and Google already are), and each raise increases bondholder worry.
No horizon Evidence as of
Limitation: The cash-flow result depends on interpreting the undefined term as quarterly free cash flow.
Assessment: UnresolvedSaid
Causal claim
“This naturally creates a vicious cycle. The more AI servers that hyperscalers buy, the more demand they create for RAM and high-bandwidth memory, which increases the price of RAM and HBM, which makes the AI servers more expensive, which means hyperscalers need more money, and because AI has yet to provide meaningful improvements in revenue or cashflow, they’re forced to raise more debt.”
Our reading: The more AI servers hyperscalers buy, the more demand they create for RAM and HBM, which raises memory prices and server costs, so hyperscalers need more money, and because AI has yet to provide meaningful improvements in revenue or cashflow they are forced to raise more debt.
No horizon Evidence as of
Limitation: Some cost evidence consists of analyst forecasts rather than realized transaction prices.
Assessment: UnresolvedSaid
Observation
“Every AI startup is unprofitable”
Our reading: Every AI startup is unprofitable.
No horizon Evidence as of
Limitation: The original interview could not be read.
Assessment: UnresolvedSaid
Observation
“Data center operators have yet to show substantive proof of a sustainable business model renting out GPUs, let alone profits that would justify taking on billions in debt, and the payoff date seems to exist somewhere between “fuck knows” and “never.””
Our reading: Data center operators have not shown a sustainable business model renting GPUs or profits justifying their debt.
No horizon Evidence as of
Limitation: Consolidated results are not the same as mature GPU-rental project economics.
Assessment: Can’t be checkedSaid
Observation
“Nobody seems to want to discuss that every data center we’re describing is 2 to 3 years in the future, and the market is becoming increasingly impatient and showing signs of non-compliance with the greater AI narrative.”
Our reading: Markets are becoming increasingly impatient with the AI narrative while data centers remain 2–3 years out.
No horizon
Assessment: Can’t be checkedSaid
Causal claim
“Because it isn’t working right now, the only reason that any of you think it’s working is because CoreWeave (which lost $740 million last quarter) and Nebius ($399 million in revenue, $8.45 billion in debt) haven’t had trouble raising debt.”
Our reading: The AI data center and GPU business is not working right now, and the only reason people think it is working is that CoreWeave and Nebius have not had trouble raising debt.
No horizon
Qualifier: right now
Assessment: Can’t be checkedSaid
Forecast
“It’s unlikely that SB Energy raises all this money at once, which means that it’s going to fall into the same problem as hyperscalers are facing — that the more debt that AI data centers raise, the more expensive it becomes to raise debt for AI data centers.”
Our reading: SB Energy is unlikely to raise the funding at once and will face rising debt costs as AI data center debt grows.
No horizon
Assessment: Can’t be checkedSaid
Forecast
“Even with the cast-iron guarantee of mag7 findom NVIDIA, it’s hard to see how SB Energy pulls together what will likely be a succession of different $10 billion debt deals of the course of several years, especially given the above-discussed curdling of the AI data center debt markets.”
Our reading: Even with NVIDIA's guarantee, it is hard to see how SB Energy pulls together what will likely be a succession of $10 billion debt deals over several years, especially given the curdling AI data center debt markets.
No horizon
Qualifier: even with the guarantee of NVIDIA
Assessment: Can’t be checkedSaid
Forecast
“I also think it’s fairly likely somewhere between nothing and very little happens as a result here, even if NVIDIA offers its backstop.”
Our reading: It is fairly likely that somewhere between nothing and very little happens as a result of the SB Energy deal, even if NVIDIA offers its backstop.
No horizon
Qualifier: even if NVIDIA offers its backstop
Assessment: Can’t be checkedSaid
Causal claim
“This is all a very, very bad sign for the AI industry at large. If there were real, diverse, meaningful and consistent long-term demand for generative AI or NVIDIA GPUs, NVIDIA wouldn’t have to create the world’s first circular financing within a circular financing”
Our reading: NVIDIA's circular-financing-within-circular-financing is a very bad sign for the AI industry, because if there were real, diverse, meaningful and consistent long-term demand for generative AI or NVIDIA GPUs, NVIDIA would not have to create it.
No horizon
Assessment: Can’t be checkedSaid
Rhetoric
“Allow me to be subjective for a second: this deal is bullshit.”
Our reading: The NVIDIA/SB Energy/OpenAI deal is not credible.
No horizon
Assessment: Can’t be checkedSaid
Causal claim
“As a result, the only way that NVIDIA can continue beating and raising each earnings season is to manufacture these massive deals, all to avoid discussing the blatantly obvious truth that diverse demand does not exist.”
Our reading: The only way NVIDIA can keep beating and raising each earnings season is to manufacture massive deals, which it does to avoid discussing the obvious truth that diverse demand does not exist.
No horizon
Assessment: Can’t be checkedSaid
Causal claim
“NVIDIA’s revenues are, for the most part, propped up by FOMO rather than any real relationship to revenues, productivity, or reality, much like the rest of the semiconductor companies profiting off of AI.”
Our reading: NVIDIA's revenues are mostly driven by FOMO rather than real productivity or revenue gains.
No horizon
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“It’s all deeply unsustainable, vile and wasteful, and only made possible in a lax regulatory environment, a captured tech and business ecosystem, and an economy dominated by growth-at-all-costs thinking.”
Our reading: The AI industry's layered arrangement is deeply unsustainable, vile and wasteful, and is only made possible by a lax regulatory environment, a captured tech and business ecosystem, and an economy dominated by growth-at-all-costs thinking.
No horizon
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“Their actual AI products — outside of providing compute to Anthropic and OpenAI — are mediocre also-rans that range from embarrassing to actively harmful, deeply unremarkable simulacrums of whatever OpenAI or Anthropic’s product du jour might be”
Our reading: Hyperscalers' own AI products, apart from providing compute to Anthropic and OpenAI, are mediocre also-rans ranging from embarrassing to actively harmful, deeply unremarkable imitations of whatever OpenAI or Anthropic's current product is.
No horizon
Scope: outside of providing compute to Anthropic and OpenAI
Assessment: Can’t be checkedSaid
Forecast
“At any given time, the market could simply refuse to wait any longer.”
Our reading: The market could at any time stop tolerating hyperscalers' delayed AI payoff.
No horizon
Assessment: Can’t be checkedSaid
Observation
“None of the hyperscalers have created a single new product line or service that actually matters, nor do they have any unique IP or technology that could turn into one.”
Our reading: None of the hyperscalers has created a single new product line or service that actually matters, nor do they have any unique IP or technology that could turn into one.
No horizon
Assessment: Can’t be checkedSaid
Forecast
“The fact that we are even debating whether AI is helping these companies and that nobody can just show me the actual numbers to prove me wrong are the signs that something is very, very wrong in a way that’s unlikely to change.”
Our reading: AI's lack of demonstrable benefit to hyperscalers is unlikely to change.
No horizon
Assessment: Can’t be checkedSaid
Conditional forecast
“It’s also unlikely to change with another trillion dollars’ worth of capital expenditures.”
Our reading: Another trillion dollars of capex is unlikely to make AI demonstrably pay off for hyperscalers.
No horizon
Only if: with another trillion dollars of capex
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“People are going to be very mad at me for saying this, but when you strip away the media hype and the investment rounds, the actual things that generative AI can do are, at best, kind of cool and for the most part awkward and mediocre.”
Our reading: Stripped of media hype and investment rounds, the things generative AI can actually do are at best kind of cool and for the most part awkward and mediocre.
No horizon
Qualifier: when you strip away the media hype and the investment rounds
Assessment: ContradictedSaid
Observation
“Outside of coding, AI really has no fundamental use cases beyond being slightly better at interpreting search queries, brainstorming, and searching documents.”
Our reading: Outside of coding, AI really has no fundamental use cases beyond being slightly better at interpreting search queries, brainstorming, and searching documents.
No horizon Evidence as of
Limitation: The assessment uses the ordinary practical meaning of 'fundamental'.
Assessment: UnresolvedSaid
Forecast
“LLMs are economically and practically unsuited to the tasks they’ve promised to solve, and no amount of extra capital expenditures or venture capital dollars will magically fix that.”
Our reading: LLMs are economically and practically unsuited to their promised tasks, and more capex or VC money will not fix that.
No deadline stated Evidence as of
Limitation: The undefined tasks and economic standard materially affect what would count as a counterexample.
Assessment: Can’t be checkedSaid
Causal claim
“AI, as it stands, is an exercise in kayfabe — a thing people are taking seriously because the rich and powerful are saying they have to and a media ecosystem built to celebrate them says that you need to.”
Our reading: AI, as it stands, is kayfabe: people take it seriously because the rich and powerful say they have to and a media ecosystem built to celebrate them says they need to.
No horizon
Qualifier: as it stands
Assessment: Can’t be checkedSaid
Rhetoric
“Every layer of GPU (and CPU) intensive agentic bullshit is a tacit admission that LLMs are not a tool with any of the features of “artificial intelligence” anyone has dreamed of.”
Our reading: Every layer of GPU- and CPU-intensive agentic software is a tacit admission that LLMs are not a tool with any of the features of 'artificial intelligence' anyone has dreamed of.
No horizon
Assessment: Can’t be checkedSaid
Forecast
“And I fear everything will end in tears, all because those with the responsibility to tell the truth or stand in the way of grifters both opened the doors and sung their praises, attacking and othering skeptics who wanted to avoid what I fear is now an inevitably painful future.”
Our reading: He fears everything will end in tears and that an inevitably painful future now lies ahead, because those responsible for telling the truth or stopping grifters instead opened the doors, sang their praises and attacked skeptics.
No horizon
Qualifier: I fear
Held from this post (23)
These records did not pass every check, so nothing about them is published. Held is not a finding.
- The two reviews reached different assessments. Our reading, not assessed: SB Energy will have a great deal of trouble raising the money for the campus, and even assuming it raises the money, he struggles to see how a company that has never built a data center can build the largest data center campus in history.
- The two reviews read the claim differently
- The criteria are ambiguous. Our reading, not assessed: The NVIDIA-backed SB Energy 10GW data center deal for OpenAI, including SB Energy's IPO, will not happen as described.
- The criteria are ambiguous. Our reading, not assessed: By end of 2026 hyperscalers will have invested over $1.5 trillion in LLMs, which have not produced revenue worth disclosing.
- The two reviews read the claim differently
- The two reviews read the claim differently
- The two reviews read the claim differently
- The two reviews read the claim differently
- The two reviews read the claim differently
- The two reviews read the claim differently
- The two reviews reached different assessments. Our reading, not assessed: Nobody other than hyperscalers, Anthropic and OpenAI seems to be demanding gigawatts of compute, because outside the AI labs and those supporting them there is less than a fifteenth of the demand needed to support the 190GW of planned data center capacity.
- The two reviews reached different assessments. Our reading, not assessed: Because CoreWeave has not yet built its needed capacity, it has not bought all the equipment, and the longer it takes to buy that equipment, the more expensive it will get.
- The two reviews read the claim differently
- The two reviews read the claim differently
- The criteria are ambiguous. Our reading, not assessed: AI code generation is not a business model; Anthropic and OpenAI let users burn $8,000–$14,000 a month for a $200 subscription.
- The two reviews reached different assessments. Our reading, not assessed: Even if it has broken ground, the SB Energy data center will find it extremely difficult to meet its timeline of finishing phase one (around 800MW) sometime in 2028, and at a rate of 800MW every two years it will be more than a decade before it opens.
- The two reviews read the claim differently
- The two reviews read the claim differently
- The two reviews reached different assessments. Our reading, not assessed: Even if margins remain stable and both OpenAI and Anthropic stay alive for the next five years, hyperscalers' revenue growth would still be overshadowed by capital expenditures that represent 100%, 118% and 181% of cloud revenues as of their last quarters.
- The two reviews read the claim differently
- The record does not meet the current rules. Our reading, not assessed: A deal may be signed, but SoftBank entities will struggle to raise $30B+ for chips and $14B for construction.
- The record does not meet the current rules. Our reading, not assessed: Blackwell being '10x better' than Hopper does not appear to have made any AI startups profitable, or even more profitable, nor to have lowered anyone's costs in a way measurable in dollars and cents.
- The two reviews read the claim differently