Rendered at 12:14:18 GMT+0000 (Coordinated Universal Time) with Cloudflare Workers.
ElProlactin 22 hours ago [-]
It's hilarious that Ed can't even take the gift Jensen Huang has given AI skeptics and do something reasonably intelligent with it.
Just a few points to highlight how Ed's hatred doesn't even allow him to do a thoughtful, accurate analysis.
> The true scale may be far, far greater — and Culper Research believes, as mentioned, the total volume of kit going to Chinese customers amounts to 20 percent of the company’s revenues for FY 2026.
This is not what Culper Research suggested. Culper Research suggested that 20% of Nvidia's compute revenue "remained driven by" China.
"We believe that in reality, over 20% of Nvidia’s FY 2026 compute revenues remained driven by China"
Nvidia's total FY 2026 revenue was $215.9 billion. 20% of that is exactly $43.2 billion. This is the number Zitron claims went to China. In actuality, compute revenue for FY 2026 was about $162 billion, so 20% of that would be ~$32 billion. An $11 billion difference.
You would expect a journalist or expert observer to be able to pay attention to the details and get them right. It seems Ed just skimmed Culper's paper and, without even understanding that "compute revenue" is a subset of Nvidia's total revenue, eagerly used his calculator app to figure out what 20% of Nvidia's total revenue was.
Zitron made an even worse mistake with this claim:
> ...NVIDIA has inflated its earnings by more than $4.8 billion using sales to what appears to be a passthrough entity for customers in Mainland China.
The $4.8 billion is the reported value of the purchases Megaspeed made. Revenue is not earnings, so this did not inflate Nvidia's earnings.
How anyone can take seriously a commentator who doesn't know the difference between revenue and earnings is beyond me.
pllbnk 22 hours ago [-]
Ed Zitron is now in the "I told you so" mode, and rightfully so. First, he always makes sure to say he is not a financial analyst and is learning as he goes, but he always does his homework to expose all the crooked deals and it doesn't make the corruption any better when you say "they illegally smuggled only $32 billion of goods, not $43 billion". I don't understand what most of it means, or whether all of it is illegal, but I trust what he says because he has been talking about it (not only he, but Gamers Nexus and others) for a long time while other journalists kept silent.
m_w_ 21 hours ago [-]
> I don't understand what most of it means, or whether all of it is illegal, but I trust what he says because he has been talking about it ... for a long time
> Ed Zitron is now in the "I told you so" mode, and rightfully so.
Ed has been calling bubble since 2023. Stocks go up. Stocks go down. If you've traded Ed's position for the past three years, you've missed out on some of the most consequential, life-changing investments in history. Literally. Even after the recent sell-off.
Anthropic, which Zitron argues might not even make it past 2028, is probably going public next week in one of the biggest IPOs in history.
This is not to say that concerns about AI financing and valuations are unfounded, but take, for example, CoreWeave. While I personally haven't touched it in the past 6 months and wouldn't touch it right now (I had a small position previously that delivered a near-100% return) because I think the questions about its commitments call into question its long-term prospects, it was up 25% the other day on the back of better-than-expected earnings and 168% year-over-year revenue growth. It has a revenue backlog of over $130 billion, 10% of the Fortune 500 now uses its services, and over half of its new customers in 2026 come from outside of the AI industry.
Meanwhile, Zitron is screaming "buying NVIDIA GPUs is a shitty deal for anybody other than Jensen Huang" and "if it wasn't, hyperscalers would be screaming about the incredible, actual revenues from their $1.4 trillion or so in capital expenditures" while completely ignoring that demand hasn't waned and the revenues are growing.
I mean, there are reasonable bear arguments around CoreWeave but Zitron is all over the place and couldn't string together a coherent argument to save his life. He seems childishly incapable of acknowledging that AI is a "real" technology with value and that there is demand for it.
The reality is that AI could be tremendously valuable and in-demand for years to come and that some of the companies that are involved in the boom are going to struggle or fail because their financial structures don't work. These are not mutually exclusive and, in fact, when you look at the introduction of any new technology, this type of outcome is quite normal and to be expected.
> First, he always makes sure to say he is not a financial analyst and is learning as he goes, but he always does his homework...
You don't need to be a financial analyst to understand that revenue is not earnings. He's been a full-time AI naysayer for 3 years and hawks a $70/year newsletter for people who want to pay to read his rants. That's more than enough time and incentive to learn the most basic accounting terms and principles.
It's just absolutely shoddy, sloppy, lazy "journalism" to publish a claim that Nvidia "inflated" its earnings by $4.8 billion when what really happened is that it sold $4.8 billion worth of chips to a company that might have been smuggling them to China (revenue).
> ...it doesn't make the corruption any better when you say "they illegally smuggled only $32 billion of goods, not $43 billion".
You can read my comment history. I think it's very likely that Nvidia has been playing dirty here and I'd speculate that this is one of the reasons Jensen Huang resigned so quickly.
But Culper's claims about the amount of revenue "driven by" China are unsubstantiated at this point, and because Nvidia is allowed to sell certain chips to China and to cloud providers in Asia that rent compute to Chinese companies, the notion that all of the revenue "driven by" China was the result of smuggling is just asinine.
The smuggling discussion also touches on something that ostensibly Zitron would find difficult to get his head around: China is smuggling these chips because it sees them as being tremendously valuable.
The Chinese government has designated AI strategically important technology. The government's own words: it presents a "rare, historic opportunity" and will be a "core driver of national security, technological self-reliance, and global influence".
There aren't billion-dollar smuggling rings and gray and black markets in China for Nvidia chips where buyers pay 50%+ premiums because the chips are worthless to everyone except Jensen Huang. They're being smuggled and buyers in China are paying a premium for them because there is a real demand for what they do.
Just a few points to highlight how Ed's hatred doesn't even allow him to do a thoughtful, accurate analysis.
> The true scale may be far, far greater — and Culper Research believes, as mentioned, the total volume of kit going to Chinese customers amounts to 20 percent of the company’s revenues for FY 2026.
This is not what Culper Research suggested. Culper Research suggested that 20% of Nvidia's compute revenue "remained driven by" China.
https://culperresearch.com/wp-content/uploads/2026/05/Culper...
Quote:
"We believe that in reality, over 20% of Nvidia’s FY 2026 compute revenues remained driven by China"
Nvidia's total FY 2026 revenue was $215.9 billion. 20% of that is exactly $43.2 billion. This is the number Zitron claims went to China. In actuality, compute revenue for FY 2026 was about $162 billion, so 20% of that would be ~$32 billion. An $11 billion difference.
You would expect a journalist or expert observer to be able to pay attention to the details and get them right. It seems Ed just skimmed Culper's paper and, without even understanding that "compute revenue" is a subset of Nvidia's total revenue, eagerly used his calculator app to figure out what 20% of Nvidia's total revenue was.
Zitron made an even worse mistake with this claim:
> ...NVIDIA has inflated its earnings by more than $4.8 billion using sales to what appears to be a passthrough entity for customers in Mainland China.
The $4.8 billion is the reported value of the purchases Megaspeed made. Revenue is not earnings, so this did not inflate Nvidia's earnings.
How anyone can take seriously a commentator who doesn't know the difference between revenue and earnings is beyond me.
https://danluu.com/zitron/
Ed has been calling bubble since 2023. Stocks go up. Stocks go down. If you've traded Ed's position for the past three years, you've missed out on some of the most consequential, life-changing investments in history. Literally. Even after the recent sell-off.
Anthropic, which Zitron argues might not even make it past 2028, is probably going public next week in one of the biggest IPOs in history.
This is not to say that concerns about AI financing and valuations are unfounded, but take, for example, CoreWeave. While I personally haven't touched it in the past 6 months and wouldn't touch it right now (I had a small position previously that delivered a near-100% return) because I think the questions about its commitments call into question its long-term prospects, it was up 25% the other day on the back of better-than-expected earnings and 168% year-over-year revenue growth. It has a revenue backlog of over $130 billion, 10% of the Fortune 500 now uses its services, and over half of its new customers in 2026 come from outside of the AI industry.
Meanwhile, Zitron is screaming "buying NVIDIA GPUs is a shitty deal for anybody other than Jensen Huang" and "if it wasn't, hyperscalers would be screaming about the incredible, actual revenues from their $1.4 trillion or so in capital expenditures" while completely ignoring that demand hasn't waned and the revenues are growing.
I mean, there are reasonable bear arguments around CoreWeave but Zitron is all over the place and couldn't string together a coherent argument to save his life. He seems childishly incapable of acknowledging that AI is a "real" technology with value and that there is demand for it.
The reality is that AI could be tremendously valuable and in-demand for years to come and that some of the companies that are involved in the boom are going to struggle or fail because their financial structures don't work. These are not mutually exclusive and, in fact, when you look at the introduction of any new technology, this type of outcome is quite normal and to be expected.
> First, he always makes sure to say he is not a financial analyst and is learning as he goes, but he always does his homework...
You don't need to be a financial analyst to understand that revenue is not earnings. He's been a full-time AI naysayer for 3 years and hawks a $70/year newsletter for people who want to pay to read his rants. That's more than enough time and incentive to learn the most basic accounting terms and principles.
It's just absolutely shoddy, sloppy, lazy "journalism" to publish a claim that Nvidia "inflated" its earnings by $4.8 billion when what really happened is that it sold $4.8 billion worth of chips to a company that might have been smuggling them to China (revenue).
> ...it doesn't make the corruption any better when you say "they illegally smuggled only $32 billion of goods, not $43 billion".
You can read my comment history. I think it's very likely that Nvidia has been playing dirty here and I'd speculate that this is one of the reasons Jensen Huang resigned so quickly.
But Culper's claims about the amount of revenue "driven by" China are unsubstantiated at this point, and because Nvidia is allowed to sell certain chips to China and to cloud providers in Asia that rent compute to Chinese companies, the notion that all of the revenue "driven by" China was the result of smuggling is just asinine.
The smuggling discussion also touches on something that ostensibly Zitron would find difficult to get his head around: China is smuggling these chips because it sees them as being tremendously valuable.
The Chinese government has designated AI strategically important technology. The government's own words: it presents a "rare, historic opportunity" and will be a "core driver of national security, technological self-reliance, and global influence".
There aren't billion-dollar smuggling rings and gray and black markets in China for Nvidia chips where buyers pay 50%+ premiums because the chips are worthless to everyone except Jensen Huang. They're being smuggled and buyers in China are paying a premium for them because there is a real demand for what they do.