Does the math check out, though? For reference, Nvidia added $17B of non-marketable securities in 2025, and its total revenue was $215B. So, assuming all these equity investments went to GPU purchases, they only amount to 8% of its annual revenue.
Since these firms (neoclouds and AI startups) will get consolidated if Nvidia puts too much money into them, I assume the majority of their funds are coming from elsewhere.
> Corporate development-led AI equity investments alone exceeded $40 billion in just the first four months of 2026
The accusation in the headline is that this activity contributed to the general shortage of GPUs which would, if true, increase their pricing power and profit margins even on chips sold to other customers. So I don't think it's a direct comparison to their equity deals.
Is this the only reason Nvidia made money? No, obviously. But is it adding fuel to the fire? If there is a bubble will this make the fallout worse? Seems likely.
More importantly this suggests a second plausible reason the board might have wanted to wrest control of the company away from him. (The other hypothetical one being that he was doing illegal stuff.)
According to Gemini, Nvidia added $43 billion in direct equity stakes,$36 billion in AI compute partnerships, and $100+ billion in loan lease guarantees, which works out to $63 billion this year, $20 billion for the next 3 years, and $10 billion for the last 6 years based on it's assumptions.
So about a third, a sixth, and a twelth of their annual revenue respectively.
I think the larger concern is that both the revenue and assets/agreements depend on Nvidia maintaining sales, market share, and their share of capacity at TSMC/etc... If they run into trouble on any of those fronts, their revenue and assets could take hits from multiple directions.
I actually got a very different impression. He could publish this interesting result, exactly because he had been working on this project i.e., had an interesting dataset to feed into Claude and had the expertise to interpret its output.
Compare him to those who are burning tokens to generate cringeworthy AI slop on social media. The value of a tool depends on the skill of the user.
Apparently, when Japanese farmland is abandoned, the landscape generally transitions into a dense and dark forest. So species that prefer open fields or ponds tend to disappear, because humans stop cutting trees and irrigating rice fields.
Overall, I got a faint impression from the article that it’s like saying “abandoning a zoo reduces biodiversity”. Strictly yes, but these species were there only because humans constantly intervened in nature.
Why do we die of old age in the first place? My money's on it being a strongly positive evolutionary force for a species. Almost all organisms age. It would make sense to me that the ones that tried not aging are gone now, likely because they were much much worse at adapting to changing conditions than ones that could reroll the dice every generation. I would further guess that the healthy oldsters filled up their ecological niche and blocked the young individuals from gaining a foothold.
But that argument doesn't prevent limited mechanisms, and conditional mechanisms in particular would still be rewarded. Conditional mechanisms provide much more short term adaptability to changing conditions, without damaging the long term species-wide adaptation.
There are intriguing parallels with centrally planned vs market economies.
Natural selection doesn't work like that. Having two offspring is good. Having three offspring is even better. Having ten offspring is really great. More than than ten is greater still...
There's a common theory that for mice, they're unlikely to survive for many years because of predation, so there's not much benefit to preventing aging in later years. Small benefit means any counterbalancing disadvantages (lower vigour?) could outweigh the anti-aging benefit, or that the force of selection is insufficient to overcome random mutations that degrade the anti-aging mechanism, but "that's good enough, no need for more" is not how evolution works.
Do you think AI is as good at writing bug-free code as finding bugs? If that's the case, this should be a one-time event during human coders -> AI transition. It's like catching a lot of bugs during a handover review.
What categories of bugs would you say we eliminated going from macro assemblers to compilers?
(Operating systems, by the way, are the exception to the rule for assembly: the Linux kernel today has an x86/x64 assembly source component around the same size as MSDOS 5.0: ~250kloc of it. Windows is smaller, based on the NT4 kernel leak at around 70kloc. As a percent of total they're much less, of course, but they also do much more. And have correspondingly more bugs, too.)
> What categories of bugs would you say we eliminated going from macro assemblers to compilers?
Entire classes of 'safety' bugs that just don't exist in modern compilers.
E.g.: forgetting to "POP" something you "PUSH"-ed, especially when it was a conditional allocation for a temporary.
Similarly, type errors were much more common, especially signed/unsigned, or accidentally overwriting things in scenarios that just don't occur even with ordinary C code. Function args were easy to push "correctly enough" that it "worked", but not really.
The code it writes in the first attempt is at least as buggy as a human's. So the question is more whether it is as good at spotting bugs in its own code as it is at spotting bugs in human code
I'm tempted to say "no, it isn't". But I don't have numbers to back that up
This is correct. Out of the $80.7B loss, $72.4B was due to changes in the fair value of their convertible notes (and similar financial arrangements).
Going a bit meta, I'm curious why we don't observe a faster convergence of opinions here. I assume that HN users are well-versed in AI tech, and the current generation of LLMs will happily break down the relevant information in Form S-1. Nevertheless, we see a chaotic storm of opinions, with wildly varying S/N ratios.
Is it a limit of current LLMs, or is the bottleneck perhaps human nature?
Yes but those convertible notes only exist to grant them access to compute they would otherwise need to pay fair market value for.
If I drive a taxi making $200/day and need $150/day of gas.. but convince Shell to take a convertible note instead of $150.. Do I make a) $200/day profit, b) $50/day profit, or c) $200/day - (change_in_value[convertible_note]) in profit?
Some are arguing that C is somehow fake, but its obviously more real than A!
Another big component of the $80B loss was $3B in stock-based compensation. This line item is very expense-like (“instead of paying your drivers in cash, give them shares of your taxi company”), and Anthropic also adjusts it away when reporting their adjusted net income. Without this adjustment, their non-GAAP operating income was negative in 2026 1H.
Overall, I feel like a man in the parable of The Elephant in the Dark while writing this. Not only is my sight limited, but this elephant (Anthropic) is dancing wildly!
Human nature. Obviously we know some things are just true. It's common sense! Why would I invite a non-deterministic hallucination machine that can't think to challenge my world view with something inconvenient like facts?
People are emotionally invested in a deflation and slowdown in AI because they feel threatened by AI. They'll cling to reeds of hope and wilfully misread financial stats to make themselves feel better.
You can both appreciate Signal overall and be critical of some aspects.
What the user criticizes is due to sheer choices, which affect some users a lot; for example, numberless registration and full cloud backups can only be purchased with a Google or an Apple account, which clashes a bit with the privacy objectives.
And it's not a problem of not having the infrastructure to support other payment methods, since the donations support all the common ones, and there's now even a cryptocurrency integrated in the platform.
Disallowing and ostracizing other clients is a choice, they spent effort to enforce it (for example by adding mechanisms to recognize the official clients).
Hear hear. It’s been my observation that HN in particular has become much more combative. It’s full on like the meme where someone says “I like strawberries” and someone else complains “oh, so you hate pineapples?! They’re a fruit too, you know?! My great-grandfather ate pineapples. You’re a monster”¹. Yeah, chill, no one said anything about pineapples, stop assuming things or thinking everyone is against you. Don’t succumb to the tribalism.
¹ Unfortunately, I feel I must point out that I’m (I hope obviously) exaggerating for comedic effect.
What really disturbs me as I’m watching myself engage it more frequently. I’ve left other sites because I didn’t like the side of me that came out. Feeling a little bit like that these days here.
For what it’s worth, this article cites a longitudinal study done by German researchers.
They basically installed a bunch of bird feeders outside nursing homes to see if watching birds improves the health of the elderly people. The researchers concluded that it did:
Nursing home staff were the first to see the difference the birds made. "For weeks, they haven't seen a smile on [a resident's] face, and then they see the smile," Kals says.
Some of the memory and concentration improvements are evident after just three to four weeks, Kals says.
It seems that they used the following bird feeder in their study:
That sounds a bit like "enrichment" for pets. We all go a little nutso without it.
My apartment has a great view on nature. I see the sunrises and the sunsets, swallows racing across the evening sky, a chestnut's whole cycle, and all that lives in my balcony garden.
I cannot overstate how much it does for my happiness. I love the morning routine of caring for my garden. I love to watch things grow and seasons change.
I think our caveman brains need this sort of thing.
I'm curious about your perspective on the impact of LLMs on physics.
My impression is that (theoretical?) physics is increasingly becoming a branch of abstract math. It has somewhat reached a practical limit for experimental verification -- perhaps the last landmark result was the confirmation of the Higgs boson 15 years ago?
So I see no reason LLMs cannot crank out a lot of theoretical physics papers, maybe combining forgotten/obscure theoretical branches explored in the past. After all, the most advanced theoretical physics is about internal consistency just as pure math is.
It's a great question. AI guided results of theoretical calculations are already a given (e.g., see https://www.anthropic.com/research/vibe-physics). Right now, I use LLM's as a book-smart colleague, who can explain standard things that I might be weak or unclear about (I am an experimental condensed matter person, so my high-energy and general relativity knowledge isn't very deep.)
I do sort of worry that an experimentalist like me will become more or less a technician in the next decade or so, with AI suggesting experiments, and doing the calculations and data analysis.
But that said, in physics (and particularly biophysics) there are many things that are simply going to have to be done in the real world. Even the action of the simplest enzymes are pretty much beyond our ability to compute from first principles; and any engineered thing that goes into a living system has to be tested, as there are many pathways that are just unknown. I don't care how "smart" an LLM is, it doesn't know what it doesn't know, just like us monkeys.
Over the years I have seen many beautiful theories murdered by ugly experimental facts. AI isn't going to change this.
The real question is if the AI or humans run the experiments.
So much of science experiments these days are pulling signals of statistical significance out of experimentation noise. Repeating simplified versions of some of these experiments is easy to do. Correlating data from experiments is something computers have done for decades, having an overarching 'smart' computer to do this even faster just seems like a natural progression.
Yep, automation has made a lot of high SNR experiments possible. But someone still has to construct the experiment in the physical world and take the data. When I start seeing robots plugging BNC cables around tricky corners, and then wrapping tin-foil around things to suppress some damned ground loop, I will know the jig is up. But I think we are a long way from that being economical (a general purpose 100 Watt monkey that self-reproduces is going to still be competitive for a long time to come, at least for one-off type activities...)
I work in industry, developing scientific measurement equipment. So I'm right in the thick of things.
There are still plenty of problems in theoretical physics that are not quantum gravity. A better understanding of superconductors would be welcome right now. Likewise, nuclear fusion. Many interdisciplinary problems such as climate science and batteries.
But in terms of sheer numbers, the life sciences dwarf physics, and have been racing to automate lab work for decades. There's real money involved, not puny government grants. Larger commercial bio labs have already centralized management of technicians and data. The vendors of products in that space have all announced AI features.
In terms of my own career (at age 62), the most likely scenario is that people pay me to invent tools that make other people more productive. There's no economic reason to come after my job... yet.
Meanwhile, don't overlook the value of a "technician" who has a situational awareness of subtle conditions in the lab, and a personal stake in success of the experiment.
You are talking about high energy physics. In high energy maybe AI can help analyze accelerator data and physicists are probably already using it for that. There is theory all over the place, including condensed matter, a branch related to quantum physics of many problems and includes some hard problems like high tc super conductivity. Many (most?) theorists that manage to get interesting publications have early access to experimental results and can help direct experiments. AI maybe not have much to say about the occasional experiment that showed some unexpected result. Also in general in physics the objective is rarely to solve a mathematical problem exactly, it mostly finding the right model and right approximation and compared to math it is a bit like fuzzy logic. I am not sure if the process has been codified to the extent that AI can pick it up from its training set, but we never know.
Since these firms (neoclouds and AI startups) will get consolidated if Nvidia puts too much money into them, I assume the majority of their funds are coming from elsewhere.
reply