Ep. 044 - Ep. 044 - Nvidia and the World Without Jensen | Jordan Schneider (ChinaTalk)

0:00

Hey, Dylan. Hey, Jordan. Thanks for joining.

0:00

Did you guys hear the CEO of the world's largest and most important company just resigned. No. Are you are you serious? Are you serious?

0:09

What could what could possibly had led to that?

0:23

I mean, there's some allegations of some pretty serious conduct violations here.

0:28

He was apparently you can't sleep with your subordinates and trying to keep it. Yeah. Wait, Jordan.

0:38

I mean, we've got Elon, we've got Sergey, everyone in Silicon.

0:43

Once you get to a trillion dollars, I think you're like legislated to start sleeping with your subordinates.

0:48

There's there must be something else.

0:50

I think the trillion dollar class might be fine, but once you get to five trillion, then you have problems with this sort of stuff. Then woke 2.

0:56

0 really starts coming after you. Oh, 3. 0.

1:01

Yeah, we're we're rotating back.

1:05

So yeah, this is truly a, you know, momentous day, momentous moment for the future of America, China, and artificial intelligence where Dylan somehow looks bored after all this, or just drained. I don't know.

1:21

How are you processing it?

1:25

Dude, I mean, I don't know.

1:25

Like you're like look up to some guy, this guy, and then like, you know, he does these things. It's not fun, you know. Yeah.

1:31

It's look, there's a lot of things going on and a lot of things to contemplate.

1:40

I think given we're ChinaTalk, we shouldn't even talk about the future of NVIDIA.

1:44

We should just talk about the future of open source and future of AI, and China US relations.

1:52

I think like, you know, I'm I'm I'm obviously I've obviously been a hawk or a fucking bald eagle, as we said in a prior podcast. For quite some time.

1:59

And But at the same time, there's like, I'm I'm worried this is going to be used as cover to pass really stupid laws.

2:11

You know, obviously, you know, NVIDIA without Jensen is a much weaker company.

2:17

Now, whether you think NVIDIA's interests are that of the national interest of America is something that I think like most people would say yes, but I think some people in the audience would probably say no.

2:30

I mean it's pretty clear's chips were ending up in China, knowingly or not, it's probably just knowingly.

2:39

Yeah, I mean, I think I think like look, Jensen and NVIDIA were the biggest defenders against not shipping chips to China, and now they are gone and out of play.

2:48

So what regulations pass?

2:48

You know, I hope this means the MATCH Act passes, right?

2:53

That is far more important for US national interest than the RASA Act.

2:58

RASA, RASA's pretty dumb bill.

3:03

I think like, you know, Biden style diffusion controls should totally be done.

3:08

But I think like again, like if if the US government goes and passes RASA, that's gonna be a big issue.

3:12

It's gonna be really bad for American interest, probably, in the long term and short term, medium term.

3:22

So yeah, I think I think this is the key key aspects of it, right?

3:26

And the other undertoning is like you can't just ban everything because China can and will utilize the fact that their components are in every single NVIDIA server and every single Google server and every single Amazon server, et cetera, et cetera, et cetera.

3:38

You cannot get materials, you know, you cannot build these things without them.

3:43

So there is a level of tit for tat here. Yeah, I don't know.

3:47

I think I think that's that's my main take. But then You can add it. You can add it.

3:53

Well, this I think it's a crossover one, right?

3:57

So we we we should talk about the future of AI a little bit. Too.

4:00

But we open with MATCH and RASA.

4:00

Let's start with MATCH, MATCH and RASA.

4:04

So MATCH Act, for those who have not been tuning in to my MATCH Act shilling over the past few months, is this idea that basically DU advanced DUV lithography alongside EUV, which already hasn't been able to get into China, shouldn't be able to go into Ch allegedly shouldn't be able to go into China and RASA, slightly different.

4:26

It is aiming to to close the quote cloud cloud access loophole by which you know Alibaba Tencent ByteDance today currently can legally access GPUs which are banned from being imported into China by setting up these, you know, cutouts in Malaysia, which you guys have done such a great job of sort of untangling.

4:50

Jordan, do you share Dylan's skepticism about RASA and his potential fallout?

4:59

Or maybe Dylan, why don't you why don't you expand on on why you think it's such a bad thing?

5:06

I think, you know, the what is the US national interest?

5:09

Is it it is to A maintain a lead on AI as large as possible, and B control China's access to AI as much as possible.

5:19

If your goal is to leave maintain a lead as large as possible for as long as possible and maintain control over China, you do not prevent China from having any compute, because that actually means that the US will have no compute too.

5:32

If you full cell ban anything in my world, I believe that China will just also start enacting economic war.

5:35

I know if you remember in 2022, I stayed up all night and I wrote the October 7th chip restrictions regulation, and I said China and USA are officially an economic war.

5:47

Or I said if it's a new cold war, it's an economic war.

5:54

And in there I found like 20 loopholes that I posted about that very night.

5:58

And I'm just one dude, not an army of lawyers.

6:01

And was hoping they'd get patched.

6:01

Many did not get patched and still are not patched.

6:07

Like one of the key aspects of this original restriction, which I think is, you know, mostly true, which is there's the civil military fusion within China.

6:13

The Chinese military and state will and can apprehend any resource to for military and intelligence purposes, even if it's a completely commercial thing, right?

6:26

We've seen it with things like Douyin and TikTok already, and that's the most consumer you can get.

6:32

So clearly they do utilize it.

6:32

At the same time, they will also enact and respond in certain ways.

6:37

We've seen this with the trade wars, we've seen this with some of the stuff on the auto tariffs, we've seen our auto you know, shutting almost shutting down American auto plants, things like this, threatening rare earth minerals and things like that.

6:50

There's been a lot of actions that China has threatened or done some and then pulled back to show their fangs.

6:58

And so if you ban chips completely from China, including rental and deployment, there is no reason for China to not just blow the whole thing up, right?

7:07

You have to have some tit for tat.

7:07

Now, the optimum thing for American policy is what is the minimum amount of compute that China can have with the most amount of American oversight to where they will not blow up the supply chain.

7:23

Because we do not have the stomach to disentangle the supply chain, very clearly.

7:28

We just don't over and over and over again.

7:31

And so given that, you cannot have China have no chips, period, both internally and foreign.

7:36

That will blow up this blow this all up.

7:41

But to minimize the amount of compute and maximize the control, you can have Biden style diffusion laws, where it's certain companies, and yes, it's a basically a bill written for hyperscalers, but hyperscalers can rent to Chinese companies and companies with American sort of.

7:57

Where America has like power over to be like, hey, turn it off at any moment.

8:00

You want all of their compute to be like that.

8:06

And if you allow Alibaba or Oracle or ByteDance or Alibaba or ByteDance or whoever it is, Moonshot, to rent compute from American companies outside of China, and that is their main access of compute, this is supremely awesome because China will not want to send their data externally in many cases.

8:25

And if they do, we're probably stuff that the US intelligence apparatus can can handle.

8:28

And two, if shit goes tits up, you can turn it all off.

8:32

Whereas if you turn off remote security access, then there's even more incentive to a smuggle, B, you know, you know, ramp your own domestic industry.

8:43

They're gonna do that anyways, by the way, which is why you need the MATCH Act.

8:47

But C, you you sort of really prevent any like reason for them to continue the status quo, right?

8:55

Which is a lot of Chinese industry is dedicated to making American computer, at least in the electronic space for American companies.

9:04

The largest PCB supplier to NVIDIA is a Chinese company, for example.

9:07

And it's not just NVIDIA, of course.

9:10

So I think I think you have to come out come in it with like a like a game theory perspective.

9:14

And the best thing for America to do is allow remote access of GPUs because the alternatives are much worse. Yeah.

9:19

I mean, a few pushbacks on that.

9:23

I I think there is a piece of the Chinese ecosystem, you know, Xi Jinping likely included, that is deeply embarrassed by the fact that these companies really, really want to import the GPUs, right?

9:36

Part of what we learned over the course of the Trump administration so far is even when Trump is offering them, the Chinese government is like, I'm not entirely sure we even want these companies to buy it.

9:47

So yes, we have some equilibrium right now, but I I definitely see a world in which.

9:55

You know, chunks of the Chinese government see something like RASA and say, all right, great.

9:58

You know, these companies, they thought they could, you know, get cute and not really invest in the domestic ego.

10:04

System and now they're gonna really have to play balls.

10:06

I think there are also other kind of things that are going on with Xi that would make him less likely to want to declare economic warfare on the US by doing something as dramatic as you know, shutting off all global manufacturing of chip production, even if it, you know, ends up we end up figuring out how to do that.

10:26

Yeah, but we already shut them off, why wouldn't they re why wouldn't they retaliate?

10:29

Because because then you get into a tit for tat.

10:32

And by the way, this economy isn't growing and unemployment's at 20%.

10:36

I mean it's every single time we've we've gotten into that tit for tat, they're led to a compromise and there was no tit for tat, right? Tariffs got reduced.

10:45

There was there were some restrictions that went up, but mostly China didn't just take it.

10:47

They did retaliate, and the the US economy and and administration would not be willing to stomach the pain.

10:56

Yeah, I mean they didn't in 2022.

10:56

They didn't in 2022 the the loophole the rules had so many loopholes.

11:04

Yeah, they didn't okay, they didn't in 2023, they didn't in 2024.

11:08

And then I think like look, Trump like kept doing the war in Iran even though he w lost him 10 Senate seats.

11:12

Like I think there, I think there is a aspect of pain, especially in a sort of second term lame duck when you know, all all the cabinet is leaking against Lutnick because they're pissed and they've already, you know, there's an Axios report that said that like China's already breaking the trade deal and they're already doing economic warfare against the US.

11:32

I don't think that you know escalatory dominance is necessarily something that should be completely assumed on on the part of the Chinese government.

11:41

The other argument I would make is that in RASA, if China has access to compute outside of China, they are now paying humongous margin dollars to the US ecosystem, right?

11:51

If you take a megawatt of compute outside of China, there's multiple layers of margin stacking.

11:59

First of all, NVIDIA is making, you know, You know, $7 million of profit per megawatt that gets deployed.

12:09

In addition, the cloud provider is making another $7 to 10 million dollars of profit if it's selling at today's pricing.

12:14

There is a tremendous amount of profit that we get to get, which companies like Anthropic, OpenAI, Google, Amazon, Meta, et cetera, do either less, they don't have the cloud or they're deploying their own chips.

12:30

Margins, the cost structure is so different. Why does China win?

12:33

They have economies of scale and their cost structure is better.

12:36

Why does American AI win?

12:36

We have our cost structure is better and our economies of scale is higher.

12:40

And so, you know, allowing them to ship all this money out, right?

12:45

If ByteDance wants to compete with Meta in the global market on video recommendation algorithms, Meta has a fourth of the cost of compute potentially.

12:53

If you don't let ByteDance directly buy the chips, but you do let them rent them.

12:59

Yeah, well, but I mean here's the thing.

12:59

This is like this kind of, you know, we want to do ByteDance and meta.

13:01

Like Zuck was totally fine letting TikTok spend billions of dollars on Instagram ads, right?

13:10

But if you just don't let them play the game, then you kind of win by default.

13:13

And also, Dylan, I mean, isn't there an enormous sort of even even in our, you know, shaky AI times over the past week, an enormous demand?

13:23

Like, like, do you really think it wouldn't be sort of better for the US to be able to get ahead by having you know, however many gigawatts are currently flowing you know, through Malaysia into the Chinese ecosystem by just like helping OpenAI Anthropic and DeepMind make better models.

13:41

I think And make more money.

13:41

I think that there's also a reflexive nature of this, right?

13:46

I'm just gonna I'm just gonna devil's advocate this shit for a while because I I think we're generally pretty close, but it's fun to do the devil's advocate, just so people hear.

13:56

You know, and this is this is sort of what Jensen's argument was to some extent was hey, you know what?

13:59

AI AI diffusion across the US is largely aided by Chinese models, right?

14:05

If we look at total AI ARR of the AI ecosystem, if you add up all the application layer companies that are running on open models, it's actually well north of $10 billion, right?

14:18

It's not like it's some like tiny little amount of open model revenue that is going to American companies.

14:22

And DeepSeek makes DeepSeek V4 Flash, and their revenue is less than fucking Baseten and Fireworks and and and all these guys serving the Chinese model.

14:30

Actually, America is making so much money off of the Chinese models, and it's like amazing.

14:37

And it lets us accrue and invest more and more and more.

14:42

Yes, ByteDance bought a bunch of ads on TikTok and Zuck didn't really see the threat immediately.

14:47

Yes, of course, everyone's short sighted.

14:49

Everyone makes mistakes, right?

14:49

You know, Jensen got a little horny, right?

14:52

That's his short sighted mistake at the end of the day.

14:55

But but it's he did more, he did more than that, right?

14:58

I mean, it's clear that he had one ships over a long period of time, yes.

15:02

It it's not again, it's not just the women.

15:05

It's doing a little more than turning a blind eye to the fact that his suppliers were not only doing the Malaysia thing, which like again is like kind of legal even today, but bringing the chips in the world.

15:23

I don't think it's not I don't think it's kind of legal. Yeah, no, no.

15:25

I mean I mean it's like this is like like ethical hoops.

15:28

This is like it's like unethical hoops.

15:28

I think it's about the letter of the law versus the spirit of the law.

15:31

Like Jensen is in these closed door meetings with a lot of the leaders of the U. S.

15:34

Who clearly have the spirit of wanting to slow down China's advance and decisions that he can make.

15:41

And they did not necessarily put it in the world.

15:43

But I don't think I don't think that's why he's losing his job.

15:45

I I think he lost his job because there is some evidence that he had, you know, had that that they knew the chips were going into China.

15:54

Which has been the affair.

15:54

Yeah, and and look, man, I mean, circumstantially around Jensen, his Taiwan sales manager is indicted for smuggling chips out of Taiwan.

16:07

His closest OEM partner, Supermicro, had their CEO, and co-founder, or you know, head of sales, whatever, Wally, arrested.

16:14

One of their sales managers is currently a fugitive in Taiwan.

16:19

So it's not necessarily like you know, it's not a good out of the blue that they're cracking down on this, and that people around him are being caught with clear evidence, right?

16:34

Like the picture of the Supermicro Taiwan guy with the hair dryer in his hand moving the stickers from the fake servers onto the real ones, or from the real ones onto the fake servers that pass in the audit in September of last year.

16:48

It's pretty like damning evidence that they're aware of this.

16:51

Yeah, I mean, and there was no honeypot required to get Jensen to do that with Supermicro.

16:57

Well, I think the the evidence that's required that needs to pass the test would be Jensen actually knowing about this and Jensen being complicit in it and doing anything, but Either way.

17:05

If the guess I bet the board has.

17:05

If I had to guess, the board has evidence on this.

17:09

I mean, like, how do you explain no longer reporting Singapore revenue. Yeah. True. Yeah.

17:22

Like the circumstantial evidence is there.

17:26

And when I I think this is the the big thing is like people, even in the case of his these affairs, they would point and say, Well, it's not technically illegal, except the spirit of his doctrines is against the interests of other people who have a different opinion about the relationship that the US should have with China and the technology that they're working there.

17:46

And a convenient time, right, right around the midterms is when he gets whacked.

17:51

So I mean the cover is gone, right?

17:54

Like you know, it was basically like the the the cabinet members like you know took the blue wave as okay, it's fine.

17:59

We we're we can now blow the cover and it's it's over for for Lutnick and Jensen.

18:06

Yeah, basically because it's I mean it's pretty clear Lutnick always you know and and and Trump wanted to make deals and they used export restrictions as a deal making thing, but like, you know, in another sense, right, like it's the fucking law, right?

18:16

You can't negotiate with the law.

18:19

You make a law, you make a law, right? And you enforce it.

18:23

So there's clear like issues there.

18:23

No, I think enforcement has been a big question so far.

18:27

Like if you look at the Biden era diffusion framework, it was passed and then just not enforced. It's still on the books.

18:34

Trump just directed people not to enforce that framework. Really?

18:37

The diffusion laws aren't like they didn't remove them? It was super weird.

18:41

Like he had like a comment and there was like a thing and a footnote, but they did not, they never repealed it.

18:47

In 2025, they, you know, Wait, sorry, these are laws in effect?

18:49

I didn't realize the diffusion I thought we would have to repass them.

18:53

It's not really a law in the traditional sense.

18:57

It's like a Executive action. No, no.

18:57

It's it's like guidance to the BIS, basically. Yes.

19:03

And then he never published like a real Kessler never this Kessler is just I don't know what the fuck he's been doing.

19:10

Apparently he shows up in the office.

19:10

But like, you know, he never passed a real piece of guidance saying we are no longer doing this.

19:17

He just like made a footnote somewhere, which is like, by the way, guys, don't worry about this.

19:20

Like, we're gonna bring a new thing for you soon.

19:23

And that was in May of 2025.

19:23

So no, it's it the the level of like indecision and and incompetence that we've seen and just like, you know, cognitive dissonance on the part of this policy process is really breathtaking.

19:37

But anyways, I want to come to Jordan.

19:39

Jordan, what's your what's your RASA take?

19:39

Where where do you want to put the temperature on just how many chips we should let China have access to remotely?

19:45

Well, yeah, I think what Dylan said about having the US be able to monitor what Chinese is doing, China's doing with the chips, as well as being able to have the option to cut off their access and then start to use the chip for what they want.

19:59

In a future where the intelligence wars are playing out and there's an escalation, I think is important optionality for them.

20:07

I think the other thing is that it's so clear how much American companies benefit from a mutually benefit from a relationship with China.

20:16

With Chinese companies, with Chinese researchers, with Chinese software, and just with Chinese people.

20:22

So there's lots of, you know, Chinese citizens that work for the lab today, right?

20:28

Who are born in in mainland China.

20:28

And it's really complicated for me to understand.

20:33

Like we work with engineers who are based in China sometimes.

20:38

That work for NVIDIA, that work for AMD, that have access to early products, that bring up the new GPUs, that work on the software, low level hardware, stuff like this.

20:44

And as you escalate tensions, all of this relationship breaks down, and everything just gets harder for everybody.

20:52

So insofar as you are as as anybody is pro AI, this is a bad day for AI progress.

21:02

So as as test as I mean that's what it seems like a bad day for AI progress.

21:07

That would be my high level take.

21:07

The the one thing about RASA is that that that makes it clear, you know, to me is like, like you said, China is trying to get off GPUs.

21:15

So like in our tracking in February, GLM was released with support for seven domestic accelerators, not just Huawei, but they included Moore Threads, Hygon, Cambricon, Kunlunxin, MetaX, Enflame.

21:26

DeepSeek published a megakernel with Huawei Ascend for their V4 model.

21:32

Meituan claims they trained LongCat entirely on Ascend.

21:37

Like people don't even train on AMD.

21:43

You know, it's it's like only NVIDIA or TPU that you can train on in the US.

21:47

So that's happening, like with or without the remote access laws being passed.

21:53

It's just more about motivation.

21:53

I think it's a lot about inference as well.

21:56

Like, are you gonna let ByteDance compete?

21:56

Globally with inference capacity from these providers on NVIDIA GPUs, or are you just gonna like not let them into the US and not let them have a business or you know in Southeast Asia or anything like that, or make them serve it from within China.

22:14

That's that's an open question.

22:14

But you know, maybe the the final thing is just like there are there are so many ways in which me personally and many people I know who work at the American Labs just like learn from people in China.

22:31

And if you cut off those like like direct working relationships, I think it's a one way street.

22:38

Like it's it's I think it's really hard to understand how you just roll this stuff back.

22:41

But then to Dylan's point, I'd love to get your take on this.

22:44

Like I think you can do all of the meaningful restrictions just focused on ASML and the MATCH Act.

22:51

Like just expand it beyond EUV, stop them getting access to UV tools too.

22:56

They need a bunch if they're gonna produce any of these chips that go into these datacenters.

23:00

And make them build lithography from scratch, right?

23:04

That's kind of the easiest way to stop them making any progress here.

23:08

The main thing I think that people just don't understand is that they think like if you remove one item at all of at the top of the supply chain, they won't do everything.

23:16

And I've used this analogy a hundred times, but it's the jigsaw puzzle, right?

23:20

You remove one piece, they'll fucking engineer the one piece.

23:22

If you blow up the board, it's harder.

23:25

At the end of the day, this is an economic war.

23:29

If you think you can get away with something without much blow, you know, without huge ramifications across supply chains, you're not gonna be able to solve the issue.

23:35

You have to just flip the table if you want it to happen.

23:38

The best way to flip the table is to prevent them from buying relatively low value or low total market cap value goods and make them buy the expensive stuff.

23:49

Don't buy don't buy ASML equipment and make chips and make models.

23:54

Make them not allowed to buy the ASML equipment.

23:56

Don't let them buy chips, make them rent the chips.

23:58

Don't let them rent the chips, make them buy the AI, right?

24:00

You know, make them transfer up the value chain where the value chain is much larger and larger and larger.

24:04

And if you do that, then China has an economic disadvantage, massively so, and this is a scale game.

24:10

And therefore the West will win.

24:13

And so I think that that's that's the key reason why, you know, look, they're gonna be pissed about whatever you ban, ban the thing at the bottom of the totem pole and chop the legs out underneath them.

24:23

Don't just attack only the top sphere, the AI aspect, the chips.

24:27

Focus on how you make the chips, focus on the equipment to make the chips.

24:30

That's what's more important.

24:30

Yeah, I think it's also a time thing.

24:34

Like we're kind of watching China on the rise.

24:37

And I think that in response to a lot of these actions, they don't want to escalate because the longer that they wait in these intelligence wars, the more time they have to catch up.

24:50

And there's a lot of discourse about is the gap wider or smaller.

24:56

I would say the gap in the model capabilities is wider at this point, but the gap in chip capabilities is like as close as it's ever been.

25:05

You know, Ascend chips are are pretty comparable to hopper chips.

25:07

And you can look just look at the phone chips from Huawei right now.

25:13

Like if you look at Kirin and compare it to what Apple is shipping for the phone SOCs, it's it's as good or better.

25:25

That's the that's the thing to focus on.

25:25

Not the models that are, I think, much more clearly behind the American frontier at this point still.

25:33

Well I mean there's a lot here. I think to start, right?

25:37

It's not just like yes, it may be bad for progress, but that the the the lodstar from a you know national interest ChinaTalk perspective is is relative capability, right?

25:46

So even if this is something that you know, if if we're if our ecosystem is 2% cut off, but like the delta between US China is getting tighter, then that's you know, maybe a worse thing than if the ecosystem is 40 or 50% cut off at the different kind of levels, however you want to slice it, but the delta is is is increasing over time.

26:04

Now, you know, you mentioned the the the the Kirin chips versus iPhone and, you know, where where Huawei is versus NVIDIA chips, but the question the question's not just like capability, it's also scale, right?

26:17

And that's where it comes back to the equipment. Like, yes, it sucks.

26:23

Or it's not ideal that you know China's been able to break through the sort of 14 nanometer line, which we like kind of sort of tried to draw on the sand in in in 2022.

26:33

But you know, this again, it's a scale game, right?

26:37

It's not just that you can make a thousand chips, you need to make a hundred thousand.

26:39

You need to have megawatts upon megawatts of like advanced hardware in order to be able to both train train and deliver this AI it's such that it like makes a difference to your economy.

26:52

And you know, as long as you're getting the equipment the the equipment side right, it's just like just makes it so much harder for the Chinese ecosystem to compete with the American ones if they can't sort of rely on the Japanese and the Dutch.

27:08

And by the way, it's like they already can't buy the stuff from the Americans.

27:12

So it's not it's not like screw over US companies on the SME side.

27:17

It's making sure that the Japanese and Dutch and Koreans can't like you know, make up for the fact that we're not able to to sell into them.

27:25

I hear you, Jordan, on the on the interpersonal stuff.

27:28

Like there is an aspect of this which is deeply tragic.

27:30

Like they're like AI has the potential to deliver like incredible things for humanity.

27:34

And you know, there's an there is a side of this, which is like a kind of beautiful, you know, humanity scale experiment. So make the future now.

27:42

But you know, problem is like US and China have a lot of issues with each other, and China happens to be run by authoritarian leader who runs a line in a system. And it's like, you know.

27:53

Not necessarily Liang Wen Feng's fault that that is the case.

27:58

I mean, it it it absolutely isn't.

27:58

Like there's old blog posts of him talking about how he's not interested.

28:02

And you can kind of read in in his writing like a vision of the world, which is much more kind of unified and holistic than what you see out of Xi Jinping in his speeches and and you know his frame of the world is this you know, struggle for commanding heights.

28:20

But like, you know, Liang Wenfeng doesn't run the country, right? She does.

28:22

And he's been doing it 15 years, and he's gonna be around for a long time to come.

28:27

And that's the reality that people need to operate under. And Jensen clear didn't.

28:34

Like he did not buy in to this framework.

28:34

He didn't think that sort of China advancing at the same at the same clip as the US was something which was a problem.

28:44

And there is enough of a consensus among America's elected officials that that's wrong that we've come to the place that we're not. We're at now.

28:57

Dylan, any any closing policy thoughts?

28:57

I do want your guys' take on like just how bad this is that Jensen's now off the table for the future of American innovation.

29:06

And and we should talk about open models.

29:10

I mean, he was the big champion, right?

29:10

Like, Dylan, you're talking earlier, like I I still don't understand quite why just someone didn't spend a few billion dollars and make something better than Kimi or DeepSeek can open source it, but it's money. Not that easy.

29:22

It's not money, it's talent, man.

29:26

Like, you know, like all of the most talented people primarily work at major Anthropic OpenAI, DeepMind, and meta TBD, right?

29:31

And then you know, those who don't are split across a hundred different startups. There's no start.

29:40

I mean, like, yes, Thinking Machines has a cracked team, but is Thinking Machines as good as Moonshot?

29:43

Like, no, like Moonshot skill wise is probably just as good as the US labs.

29:50

They just have a big deficit in compute and deployment.

29:54

You know, so I think this is like if we if we if if we don't have a com a scale advantage, then you don't win.

29:59

These you know the yes, these smaller labs have spent a lot of money on models.

30:06

But DeepMind, Grok, Meta, have regularly fallen behind the frontier of Chinese labs, and they're spending more.

30:16

So it's not just, oh, spend a few billion dollars and you can make a model that's better.

30:19

Like that's, you know, the innovation and the quality of the engineering and researchers at Ch in the Chinese labs is just very fucking Wait a second, by the way, NVIDIA did spend this money and try to do it and still not produce anything with Nemotron that was better than the open source models from the Chinese lab.

30:34

So even though he's a big supporter of of open source, it's not like you're guaranteed to be effective just because you pour money and time into it. Kind of hard. Like definitely hard.

30:45

You need you need compute, you need people, and you need data We just got to get Alex Wang on the case.

30:53

Is that that's the only zero to one we've had which has been successful in the past three years.

31:01

I mean, they hired a ton of great researchers and have a ton of compute for sure and spend a bunch of bunch of money on high quality data as part of it. I don't know.

31:08

Look, I think there's so there's two things that I see the discourse shifting towards, which I'm a little like not completely against, but just try to be more nuanced on.

31:19

One is like ramping up export controls in the form of RASA and MATCH and the diffusion framework is like an obvious good thing in the US should be immediately doing it.

31:30

Like I'm not sure people are thinking through all the implications, which is what we tried to discuss here.

31:32

And then the second thing is people are like immediately, oh, Jensen leaves NVIDIA and it's like guaranteed the company's going to zero and and we could never run without them.

31:42

And it's such a unique work structure.

31:42

And this is all true, but I think that in that case as well, the more nuanced take is like in the short term, NVIDIA is fine.

31:50

Like they need to deliver on a commitments to customers and build a bunch of GPUs with the supply chain that they've got.

31:57

And like nobody set them up for success as well as Jensen could, but now they just need to execute on the stuff he set up.

32:03

Them for for like Rubin and Rubin Ultra, and I think a lot of people know what to go do.

32:08

I think they know their marching orders, they know their responsibility, they know the mission.

32:11

I think the really big test for NVIDIA is like Jensen has been incredibly good at retaining talent, especially amongst his executive team.

32:21

There are very few high part high profile departures from the NVIDIA executive team.

32:25

A lot of people retire, and basically only Rene Haas has gone to ARM and like gone on to be a CEO anywhere else.

32:29

Like there's no EVPs, SVPs, VPs that like really leave that often from NVIDIA.

32:38

This is a time where, you know, Jensen and the mission isn't at the helm.

32:43

Somebody needs to continue to give them compensation and direction and keep them and retain them.

32:47

And this is probably when the sharks are in the water.

32:50

The VCs are going to come in and offer people billions of dollars to start their own chip companies, their own neolabs, hey, you should come join us over here.

32:57

I mean, we're gonna see in the medium to long term weather.

33:01

NVIDIA can operate with Jensen.

33:01

I think that question is up in the air because if a year from now we've seen like a bunch of departures from NVIDIA, I'm going to be really concerned about their future.

33:12

If a year from now everybody's still like growing in the same direction and Colette is driving things, or they've got somebody else, like especially a technical leader who has emerged, like Ian Buck steps up or something.

33:21

Maybe they transition Michael Kagan out of C CTO or something.

33:25

I don't know, whatever that looks like.

33:29

Then I would be coming back and saying, hey, I'm, you know, much more confident in NVIDIA.

33:32

But do I mean I I really think people are underestimating the dominance they have in the supply chain, in the sales and support that they have, all the relationships with the neoclouds, all the investments they have in the labs and all the startups, the ability they can they can they can use their balance sheet to backstop everybody and make them be partners with them.

33:49

They can invest in everybody, they can acquire all these companies as soon as they're successful.

33:56

Like it's still NVIDIA's to lose.

33:56

And I think there's a very, very high chance that they're still doing just fine two, three, four years from now.

34:06

And another another open question is like, okay, say like they, you know, some people get poached, they raise big rounds and they start doing new things.

34:13

Is that really like the worst thing?

34:13

Are we sure that's like the worst thing for the Western ecosystem to have, you know, different people trying new bets?

34:24

The other question is is like we're going from a founder to someone who is not the founder.

34:28

And I imagine there are just like aside from executing on the road map, there must be some like, you know, big wacky out of consensus things, which in V which Jensen probably would have done over the next five or 10 years, which maybe wouldn't have paid out for another five or 10 years, which I guess potentially wouldn't have happened.

34:49

And you know, to what extent does that you know, happen, but like half as much, to what extent is that just like, okay, instead of it happening within NVIDIA, it happens from, you know, some EVP's new startup.

34:57

That's kind of the question that I that I have, going from like, you know, the jobs to the to the Tim Cook era of of NVIDIA.

35:08

I'm certainly not sitting here saying like this is a really good day for NVIDIA.

35:14

I don't understand why NVIDIA's not down a lot more.

35:16

Yeah, I think it's I think it's really bad for sure.

35:20

But people like I'm talking to people and they're like NVIDIA's a zero.

35:23

And I'm like, you use the Apple versus Steve Jobs analogy.

35:27

If you bought when Steve Jobs went on leave and then passed away, you'd return a ton of money with Apple stock since that time.

35:34

Yeah, but didn't he come back and like No gone gone 2. 0.

35:36

I think we're I think you know Apple this is this is Jobs leaving in 2015, not leaving in like 90.

35:41

Sorry, okay, got it, got it, got it, got it, got it.

35:44

I thought you meant the original gone.

35:47

App okay, Apple in 2015, looking back is the Steve Jobs was not a maniacal guy who ran every single part of his company.

35:55

He was, yes, he was a maniacal guy who cared about every aspect of his firm.

36:00

But at the end of the day, I don't think Steve Jobs exerted as much control over Apple as Jensen does over NVIDIA or Elon does over his companies. Really? I don't think so.

36:09

I mean, I I I don't know Steve Jobs.

36:11

I don't know Apple as well.

36:11

I've only read some books, but like it doesn't seem like he was absolutely so tremendously important, but like I just don't think it's the same level of like work.

36:20

Steve Jobs is a pure visionary and like Jensen is like negotiating every supply chain contract, and Steve Jobs is not doing that.

36:32

So Jensen is a Dylan, like how do you think they if they if they start faltering, like on what dimensions will they start faltering?

36:40

I think Rubin, you know, if there's any manufacturing issues with Rubin, people are distracted and Jensen's not there.

36:43

Things get solved slower, things get solved by committees, things don't get escalated up.

36:51

I mean, I don't know how many times at SemiAnalysis-analysis is if we fucking email Jensen and for a thing to get solved. Like what the fuck?

36:57

Like why are we emailing Jensen to fix something on some open source software?

37:01

And yet that's the only way we get it solved.

37:01

And we're irrelevant, right?

37:03

Like compare that to like actual companies in the space.

37:06

How often is it Jensen going out there and solving a problem?

37:09

I can I from what I've heard from everyone around the ecosystem, it's a lot.

37:12

How many times is it like, you know, Elon tells Jensen something, Jensen goes and tells his team and the team fixes it because the team was trying to work with the team. Yeah. And it doesn't work.

37:23

All of these examples of the reasons why NVIDIA would technically have problems going forward.

37:26

I think we're actually already present with Jensen at the hell.

37:30

Like they have a a big problem with all the custom ships coming up that are Yeah, exactly.

37:34

But I'm saying, I'm saying Rubin could slip, the the yield curve could look worse.

37:37

Feynman could slip, Feynman could be not as aggressive.

37:42

You know, one of the popular stories I love saying is like Volta didn't have Tensor Cores until three months before tape out.

37:50

Jensen is the one who fucking decided this, right?

37:52

There's bickering little people, and Jensen's like, no, AI is the way, let's do fucking Tensor Cores.

37:57

Like this is this is like, and it's like fuck everything else.

38:00

I don't care that, you know, we're three months away from tape out, and you say it's gonna take six months to incorporate the change, you're gonna do it, and you're gonna do it now.

38:05

And and and stack on a hundred of these decisions.

38:08

Feynman's gonna be much worse than it is, and it's gonna come later.

38:11

Rubin's gonna be could be Yield Ramp could be slower, Rubin Ultra could be slower.

38:15

Oh, the D spec on Rubin Ultra would take longer to do slash might not even get done, and therefore the volumes that NVIDIA can ship are lower.

38:24

Like there's a bajillion things that you can decide. XYZ things on software.

38:27

Oh yeah, like DeepSeek open sourcing all the software and DPP and like all the stuff that Moonshot's doing and making KDA work, that shit's gonna take forever to get up and running at high efficiency at NVIDIA relative to before.

38:39

Maybe it's only a week, but each of these things and decisions and him driving and leadership and and the supply chain giving them 70% of their supply instead of, you know, much less because it's Jensen, you know, all of these things are exacerbating, and it could be like in 2029 the best chips aren't NVIDIA's and they don't have the majority of the supply chain. Yeah.

38:57

Well, I mean that that that's relevant to the stock, right?

39:00

But I I think like this is this is 2026, not twenty twenty-two.

39:05

There are lots of other competent teams who are running at this in a way in which there kind of was only one before.

39:08

So the fair watermark they NVIDIA is still the leader.

39:14

They're still driving the industry.

39:14

If NVIDIA slows down, it's not like someone else accelerated.

39:17

NVIDIA slowed down until someone else is the best. Yeah.

39:21

It's it's like a question mark, maybe depends on your situation for inference.

39:26

And then like No matter what, the industry is progressing slower.

39:31

Hardware, it progresses slower without Jensen at the helm of NVIDIA, period.

39:34

Yes, there are competitors, but without NVIDIA in the lead, people are running less fast, or even if they're running as fast as they were before, because NVIDIA that was in the lead and they're no longer, you know, as relevant, progress has slow.

39:45

This is a real slowdown in progress.

39:51

This is a pretty pretty morbid for all of us here. I mean it's sad.

40:02

This is very Shakespearean. Dude, I was so sad.

40:02

I'm I've been a Jensen fanboy since I was a teenager.

40:07

I don't think people like understand how much I fucking like look up to the man and love him.

40:10

Like, you know, it's crazy.

40:13

And it's sad the circumstance.

40:13

Like I think a lot of us were always impressed by NVIDIA, and just to see.

40:18

I think a lot of people like ex trying to excuse conduct that's like it's just just so bad to read about this.

40:29

I see all the memes about it. I mean, I don't know.

40:33

Dude, the memes are fire.

40:33

This is probably some of the best memes in like you know, this the memes the past few days is is up there with like Republican primaries in 2014-15.

40:45

Like the memes are that fucking good.

40:50

Yeah, I laugh and then I'm sad.

40:50

That's what the memes are doing to me right now. Eliot Spitzer.

41:00

You know, as a little high schooler into politics, he was my hero.

41:03

He was like crusading against corruption, governor of New York State.

41:06

And then it comes out, he had passed an anti-prostitution law.

41:09

And then it comes out that he was, you know, had $2,000 an hour escorts, and then has to leave the governorship.

41:19

And I that was my wake up, realized these like you know, men with egos aren't necessarily all their cut out to be moment.

41:27

So, you know, it's it's not surprising, but it is depressing, and it just kind of reminds you about how you know humans are frail and you know, don't don't ever meet your heroes. But I don't know, man.

41:41

Like we can still look up to Lisa Su, right?

41:44

She's she's still she's still there for us.

41:44

Like Dylan, who's your dream new NVIDIA CEO?

41:49

Who are we who are we putting in if we can snap our fingers?

41:56

If I could snap my fingers, Lisa Su, but I don't think she'll do it. Charlie Kawwas.

42:00

Johny Srouji, Charlie Kawwas.

42:00

Charlie Kawwas would be the best. How about you, Jordan?

42:06

Well, if Charlie's on the table, I'll just say Hock Tan, man. Bring the guy in. Tear it to pieces.

42:16

How good of a businessman?

42:16

I think I think Charlie is really good at is is is a bit more technical and less businessman than Charlie than Hock, even though he's a really good businessman.

42:24

It's like I think like that dynamic, you actually want Charlie.

42:27

Yeah, I think my personal opinion is lady name Colette Kress.

42:33

I really look forward to working with her.

42:36

And I Well we do love Colette. She's very cracked.

42:37

Yeah There is no I know I know you guys are buds.

42:42

There is no way the CFO of NVIDIA isn't implicated in whatever has been happening with these chips.

42:50

It's just it's not possible.

42:50

I could not be able to- We don't know how deep this web goes.

42:53

Yeah, we don't know how deep this web goes, but like I don't see how it's possible that like.

42:59

You know, Jensen and Colette don't both aren't both like, I don't know. Yeah, it's hard.

43:02

I mean, Collette's not the one that's in Asia, like making the agreement.

43:05

So maybe she could like pretend like, oh, I didn't know.

43:09

I signed a lot of pieces of paper. Like, I don't know.

43:11

But like she's gonna look at she's gonna she's gonna have to testify in front of Congress just how much she didn't know.

43:19

And there's gonna be, you know, congressional investigations to this.

43:21

So like look, it may, it it may be true, but like like the way it and that's that that's kind of why you want the technical CEO or why you want someone from outside, because I just don't understand how, you know, someone who is deeply involved with the operations and and kind of interacting with the city.

43:39

Yeah, and then a CEO is not gonna be called to testify and like for like days and days and days, and they're they get to focus on the company.

43:46

Yeah, look, I I think the fact that this originates from the intelligence community, kind of rules out the fact that the person who's immediately installed at C at CEO has like in depth deep knowledge of it.

43:59

Like The board has clearly been briefed on something initially that Rubio knew back in May before they did the visit to China and stuff.

44:06

I it's I think that would be really strange. No.

44:09

It's not the Chinese spy.

44:09

It's it's it's like look, the the all the facts with Supermicro, it's not the affair anymore.

44:16

It's Supermicro, it's Megaspeed, it's the it's it's chip smuggling into China.

44:21

And the sort of willful you know, covering of your eyes of Yeah Willful ignorance, yeah.

44:28

Yeah, I think anyone who's kind of complicit in that is gonna have a really hard time staying out of the line of fire of all this.

44:38

Yeah, I I'm watching the Supermicro indictment really closely to see how they actually play out prosecuting this stuff because they flip the guy who is doing the hair dryer stuff, taping, moving the stickers, who deleted his encrypted messaging app the day he got arrested.

44:56

And apparently he's working with the prosecutors now. So we'll see.

45:00

You know what happens there, right?

45:03

Like there's not really obvious ways to hide this stuff and be willfully ignorant in the technology connected internet, you know, internet world that we live in right now, where you're trying to do things in Taiwan from the US. So we'll see.

45:21

But I I would find it again, I would find it really strange if they immediately installed somebody who was like completely complicit in all the same stuff that Jensen was that's motivating to change. Maybe I hope I hope not. I hope not.

45:35

I hope she does a great job.

45:38

I hope she, you know, delivers hope all our new chips online and, you know, brings me brings me AGI, but we'll see. Any closing thoughts? Yeah, sorry.

45:51

Yeah, I think there's there's like the closing thought is just that there's potential that NVIDIA just goes into civil war here over this, because she's interim, she's not full time.

45:57

I think there's a constant clash in any company between finance and technology. And sales. Kind of three big areas. Sales wants to do deals.

46:08

Technology wants to make big bets.

46:08

Finance needs to figure out how to pay for it.

46:12

Finance is in charge now.

46:12

Well, we know what's going to happen to the NVIDIA business going forward.

46:15

It's going to be probably very profitable for a little while.

46:19

And then medium to long term, it's probably gonna look a lot different.

46:23

But yeah, closing thought on like maybe the political side is just like I hope, I really hope the people that are MATCH Act.

46:33

Yeah, I really hope for the fucking MATCH Act.

46:37

Think this stuff through, guys. Yeah, please.

46:37

I I I think like you can be a hawk and have a debate on like RASA, but it's hard to hard to have a debate on Yeah so obvious.

46:52

Anyway, We'll be watching closely, guys. Yeah.

46:55

I have I tweeted on Monday, I'm like I I did my emergency pod.

47:04

Or we should let's just end it. Say it, say it. What what's up?

47:07

Oh no, I was I I I was I think I did my emergency pod on Friday.

47:14

I was like, oh man, how many days is it gonna take before I have to do another one? Six days. Okay, yeah, turn around. Yeah. A daily emergency pod.

47:21

That's what we were saying internally.

47:26

Every single one of our pods has to be named emergency right now. It's just like. Yeah. See you next week.

47:29

Dylan, what do you think is gonna happen next week?

47:35

What how can we go how can we ramp up from here?

47:38

We thought the Anthropic through the capital markets, but now the CEO of the biggest, most valuable, most important company in the world goes away. What are we doing here?

47:47

What can what can you do?

47:49

AI 2027, they said China wakes up to AI in November.

47:52

So we've got we've got a big bear awakening. Next November?

47:58

No, no, we got to nationalize it.

48:00

We gotta nationalize Anthropic and SpaceX.

48:04

And then Dario gets thrown in jail. Oh no, no, no.

48:09

I'm just like for season two, you know, we gotta we gotta up the stakes right for our next emergency pod.

48:15

Like, you know, if they want it to go crazier.

48:17

Just to be clear, what how do you think functionally it plays out to nationalize a company of that size? I have no idea.

48:25

No, it's not actually gonna happen. Sorry.

48:28

We gotta end this podcast like three minutes away. We lost somebody. We lost the thread.

48:33

Okay, we can cut all this, but I mean I know I know Jordan's I know Jordan is a socialist and he loves and admires China so much. Socialist Canadian. What?

48:41

No, Schneider, Schneider, Schneider. Oh Schneider.

48:45

You need to pass some more power law, AI, you know, AGI.

48:49

Like there needs to be some legislation around it.

48:53

I don't think the executive just nationalize that.

48:55

But what I'm just what I'm getting at is that I think you're just we are watching the nationalization of NVIDIA play out right now.

49:02

Like this is the equivalent of Jack Ma disappearing. I disagree too. I don't think so at all.

49:06

I disagree completely, Jordan.

49:10

All he had to do was not sell chips to people smuggling them to China, and he would have been fine. That's all.

49:15

So he done something against the national interest, and the national intelligence community comes in, takes him out, and installs somebody that they have more influence over.

49:23

Yeah, when you break the law, there are consequences.

49:26

Don't break the fucking law.

49:26

Yeah, just don't break the law. Okay, sorry, guys.

49:29

Sorry for the he can he can advocate for all he wants.

49:32

He can scream to the rap.

49:36

He could have done H20 all he wanted to, and then we would like he didn't break the law and he didn't break the law.

49:41

Wait, all you're saying right now is he deserved to be nationalized, which, like, fair enough. No.

49:44

There was like so much between don't break the law and nationalized.

49:49

This is a fucking massive chasm.

49:53

Are you saying that they are not? This is not national.

49:56

Nationalizing NVIDIA right now. That's crazy. No chance. Absolutely no chance. Absolutely no chance. What is the difference?

50:01

Again, what would they have to do for it to look like national?

50:08

I don't know who's going to be a good nationalization mean to you.

50:13

It means it does it mean a CEO.

50:17

Yeah, they have to do that. To be the CEO. No, you no. Like, I don't know.

50:19

Any of the people we've said for CEO are not nationalizing the company.

50:28

They are just logical CEOs. Right. So let's see who it is.

50:35

Jared Kushner for CEO and Donald, when Donald Jr.

50:37

Becomes CEO, then we'll know. No, no.

50:42

Really, really who was behind this from the beginning. Oh my God. I don't know. Okay.

50:47

Maybe this is too far too far gone for you guys, but I think that's exactly what's happening right now. Anyway. Cut this part out.

50:52

Well, it's it's it's aligning the country.

50:56

It's aligning a country with the laws of the books.

51:02

Like leave it in, Jordan. All right.

51:02

All right, we'll leave it in. All right.

51:07

You guys got on camera pushing back at my wild conspiracy theory that I'm going to do. And this. And I didn't repeat it. And this everything.

51:16

It's not financial advice.

51:16

Don't come after me, please.

51:17

Wait, who's gonna who's gonna end the press the button?

51:19

Dylan, can you press the button? Is it on your computer?

51:21

We've got people in the side who are who are recording this. Von