Hello and welcome back to another episode of Sharp Tech, presented to you by Stripe. Joining me in the studio, Ben Thompson. Ben, how you doing?
I thought, Andrew, the hello is a little bit too enthusiastic.
We are here on a day of mourning.
You know, it was supposed to be a day of celebration. This is our 300th episode of Sharp Tech.
But the celebration has been cancelled. It has been cancelled.
I feel like I cancelled harder than Jensen Huang.
You're sitting here in all black. I missed a memo.
I thought about going to buy a leather jacket, but I did not have time.
I had to wear black. I was looking around for a black jacket. I had a Bucks jacket.
Which, number one, it's great because it's black, which is the effect I was going for.
Number two, it's a franchise that technically is not dead, but might as well be dead.
Not in great shape. They did lose the MVP.
Yeah. So, and they're what, one in 11 or something, so it actually matches.
Yeah. Well, I'm going to say I'm wearing Nvidia green here today, supporting the new era with Colette. In any event, that's where we will begin and probably where we will live for much of this episode. Wednesday evening on the East Coast, the news came through. The New York Times cited five sources at Nvidia who said that Jensen Huang had an affair with an executive at the company. Company resources were used to cover it up. Shortly thereafter, the Wall Street Journal reported that Nvidia's board was preparing to oust Huang.
By the way, what were you thinking in terms of what are we going to do for Sharp Tech if he hadn't been fired yet?
I was under no illusions that he wasn't going to be fired in time. Eventually he was going to go, is what I knew.
I was going to say because we record on Thursday, publish on Friday morning, that we would have to acknowledge at the beginning and then not talk about it at all, because the chances of going stale. But it does, maybe video is good for us. Good luck. I've complained about timing as far as tech events go.
In this case, tech Thursday morning. Here we are. So Thursday morning it became official. Let me ask a stupid question on behalf of the audience at the outset here. So Nvidia stock down about 15% in the past five days. It did rebound a little bit on Thursday. But if the stock is down double digits in the wake of the news that Jensen could be on the way out as CEO, why is it in shareholders' best interest to make a change? Why did this happen?
It's actually a very interesting question.
A little governance question at the top.
Well, it is one of the eternal challenges, one of the biggest challenges in terms of business and management theory is alignment. There's a specific term for it, where usually it's in terms of the company's management being the principal agent problem, because their incentives are not always the same as the shareholders' incentives. Like for example, most executives want to keep their jobs. Maybe not Jensen Huang, but he definitely wanted to keep his job. But so they will act in ways that maybe are not best for the long-term company. Like this comes up in terms of...
Quarterly earnings versus longer horizon.
That's right. Sure. It comes up in incentive questions. If an executive is compensated via stock, then in theory that's in shareholders' interest. But if they know I'm only going to be in the job for a year, or they're worried about losing their job...
Juice that share price. Absolutely.
That's right. So you, but that principle, so theoretically the board should represent the shareholders and sort of keep a rein on the executives to keep them aligned, but the board can also be not aligned with the shareholders. So in this case, you know, setting aside why he unquestionably, you know, was going to get fired, to your point, yeah, this is an alignment issue. The shareholders definitely want him to stay, the board does not, and they have to be concerned about liability for themselves. And the governance of Nvidia was not admirable. Which by the way...
I'm going to defend slightly when we get to that point. But yeah, it is an interesting observation that this is an outcome that the vast majority of Nvidia shareholders did not want.
And probably employees as well. But board members, they can face personal liability if they're aware of misconduct.
And people will make the case, and there's a case to be made, that this is in the long-term interest of Nvidia. And you know, and so there was a fiduciary duty. And at the end of the day, like there was at a minimum obvious violations of policies that exist in every company, like don't sleep with your subordinate.
Yeah. Like the sleeping with a subordinate is the... you always get fired.
It's a big red line. We're going to be talking a lot about China today. China has red lines. That is a corporate red line in America in the current environment. And there may be other stuff that the board was aware of that has not yet become public. But in any event, we are where we are at this point. Where might Nvidia go from here as we look ahead? It's a loaded question.
It's pretty dire. And it's dire on multiple levels. So number one, I mentioned the corporate governance sort of aspect. There is this issue where I think Jensen Huang has made this analogy. I first got it from Tae Kim, who wrote the book The Nvidia Way. I don't know if he got the analogy from Jensen or Jensen has now adopted it from Tae Kim and is now using it.
I wish I had read that book, and now if I read that book, I bet it's actually going to be pretty sad to read through the 300 pages from Tae Kim.
Absolutely it will, because this company... Jensen is a singular figure in tech history.
Not just that he's built this company for 30 years to at one point being the most valuable company in the world, but he did it multiple times in the face of, and took multiple very large bets in the face of basically no one believing in it.
And drove it through his own personal belief and conviction from starting the company. No one saw the market for chips like that. There were other accelerator companies like from Voodoo Labs and the Voodoo card or 3dfx.
This is the 1990s. Their first graphics card was a total failure because they sort of designed it wrong. They actually got bailed out by Sega because they were making it for Sega and they had to go and say it didn't work, and Sega actually forgave them in exchange for equity in the company, which unfortunately Sega then went out of business. But so coming out with the graphics card, just the overall concept of the GPU, and then the brilliant piece was the concept of the shader, where the GPU, the first versions of GPU... So I had a TNT2, I think, or it was the first Nvidia card that I had. This was like back when I was in college. And those were... there was... oh, what's the... So there was two... there was 3D... man, I should have read for this. There was two versions of how to draw triangles on screen, basically OpenGL. And then Microsoft had one, there's something 3D, Direct 3D, I think, or Direct... man, struggling. But the...
They were sort of dumb. So they would process inputs and draw onto a screen, but it was like all totally pre-programmed and it was designed around you knew what the OpenGL commands were. You could handle all those commands and like generate triangles because everything is a triangle and put it on the screen to create graphics on the screen. What Nvidia did was come up with the concept of actually those little things should not be just dumb. They should be programmable. So you can actually like adjust them and you can make changes and in the long run they can actually run programs themselves. That was the concept of the shader. That is the fundamental idea of the GPU. And that's where CUDA comes from. CUDA is the way you do the programming on those computers. And all this came out of being a gaming company because the thing about being a gaming company is you're drawing on the screen. It is an embarrassing... it's called embarrassingly parallel problem.
If you cut the screen into smaller and smaller squares, you can go faster and faster because there's less work for any one of these processors to do. And so you could just scale up and do all these calculations in parallel at the same time.
That's what AI is like. That is where the whole GPU comes from. It's doing a bunch of calculations all at the same time in parallel that are relevant.
When did it become clear that Nvidia's hardware was going to be essential for AI development? Is that like 2018, 2019?
I think people point to... so there was, I think it was called AlexNet, this sort of image recognition that was built on Nvidia chips using CUDA.
And that was the huge wakeup call.
So CUDA was, you know, it started out as being a bet to do better graphics, like everything about Nvidia comes from graphics, and it was considered to be a fairly insane bet. It nearly sunk the company because the first version that supported it, the additional complexity meant the chip ran hot. It actually wasn't that performant. It wasn't AMD then. It was a company called... what is it? What did AMD buy? ATI or something like that, that actually made all their chips were just way better.
And in the graphics sort of era, it was sort of like frontier models today, where everyone would just switch to the one that was better because they were all relatively dumb, right? They all used OpenGL, they all used Direct 3D, DirectX, whatever it is, because I'm going to come up with it by the end of the show. Great. Maybe Jim will type it on the screen and have mercy on me. So you have the...
And so Nvidia went through this generation when they introduced this new capability and their chip sucked because of it, and they almost went out of business. Like it was really bad. But they pushed through, and then this capability to program for them was used for AlexNet to do image recognition. And that was the big wakeup call of this can be used for...
This is going to be transformational and unlock all kinds of upside that just wouldn't be there without this hardware.
And so DirectX, thank you Jim. So it used to be called, I think it used to be called Direct 3D.
DirectX was not the original name, but...
So yeah. So meanwhile, Jensen, I don't know when he came up with the idea that this can be used for more than games, if it was pre-AlexNet or post-AlexNet, but in general, that's when CUDA development really takes off. It's not just for gaming. It can be used for anything. And the, you know, all the old GTCs, Jensen showing off like 50 gazillion things, like this is nuts. Because all he was doing was just different variations of CUDA libraries, right? We could use it for this, use it for this, use it for that. And they're making all these demos basically throwing CUDA demos against the wall to see what sticks.
And the big shift was ChatGPT, and that stuck to say the least. Well, and from a normie perspective, my knowledge of Jensen is obviously his technology became economically essential basically overnight, and then he was tenacious about taking advantage of that and capitalizing on it across all sorts of different dimensions.
Right. So what's really interesting, so number one, if you look at the Nvidia stock price, you like do full-on like Google because the AI thing's so big, it's basically a totally flat line on the bottom of the chart.
Yeah. Everything's relative.
There's a little bump that I think is downstream of their sales really surging because of Bitcoin mining, or Ethereum actually in particular was done on Nvidia chips. That's also an embarrassingly parallel problem. And then it kind of goes down again when the Bitcoin era crashed, and then it just shoots to the moon. It's one of the wildest stock charts you'll ever see.
I think there's an aspect of AI was bad for Jensen and CUDA because he was out there trying to... like this is clearly a very useful technology to do this in parallel, this scalable capability. He's talking about everything's going to be rewritten for these sort of systems, which I always push back on, like no, it's hard to program with and no one's going to go back and rewrite everything. Well, maybe now they all will. But he was looking for something, and this was the most fun time to watch Nvidia because CUDA was a thing and you could see them trying to figure out what they were going to do with it, and it was...
Well, I was just going to say you can go back to 2021. You did an interview with him and you could see he was sort of fumbling in the darkness for how exactly this would be unleashed and wildly useful, and then a year later obviously, but there was a period where it was like...
Throwing spaghetti at the wall.
That's what happened with when the interview started. When we first did the interview, he really enjoyed it and he emailed me right away, says like, let's do it after every GTC.
Yeah. And the reason was he was out there campaigning for people to use this accelerated computing, and he saw it, he's like, oh, you understood what we're doing, pretty technical. And so, and then yeah, once he goes on, there's a reason he didn't need me anymore. Yep. So it was very funny actually, the first interview I did with him was I think I was flying to Taiwan that night, but Jensen Huang, I'm going to figure out how to do an interview. So I actually rented a... or I got a hotel room nearby O'Hare, and it was so noisy in this hotel room that like, and I think I got the mattress like up against the door because it was super noisy, and I'm trying to construct some sort of like relatively soundproof and I don't look like a total on the video sort of thing. It was quite the experience. I'm like scrambling to make this room somewhat soundproof.
We'll put links in the show notes if anybody out there wants to get emotional listening to old Jensen Huang interviews this weekend.
Right. It's super interesting because there's not... there's a little bit about LLMs, but you know, not very much.
Once this was super clear, in there was a GTC in San Jose. I think it was 2023, 2024. I think it was 2024 where it was a huge stadium. It's like this is the first like Jensen Huang the rockstar sort of bit, and it was super boring.
It's like they talk about the roadmap. It's like what happened to all the cool demos? It's like well no, this is the business. They brought that back a little bit the last few, but that one in particular stands out because it was like...
And suddenly least interesting Jensen.
Less imagination. Okay. And at the same time, you suddenly had very real reasons for people to fight against CUDA, and you had very real threats to their position. So in all this growth over time, those margins, were they going to be sustainable? And if people are building on top of LLMs and they're not building on top of CUDA, is CUDA really a moat? And that became the only question that Nvidia was organizing.
So Nvidia, Huang was a... I can't overstate what a good CEO he was. So one thing they did do around 2019, 2020 is they bought a company called Mellanox. Mellanox, Mellanox. What Jensen saw was we're going to scale not just on a chip, lots of processors, but lots of chips together functioning as if they're one chip, but they're actually multiple chips. It's like recursive, like chips upon chips upon chips. But to do that you needed networking, and Mellanox had this like very high-speed chip interconnect capability that was later called NVLink and a couple other things. And so he buys that so they can do this systems approach. So they are still... so their moat today isn't just CUDA. Another moat they have is they sell systems, not chips.
Like AMD, I'm not sure what the status is right now. Like a previous generation, their chip was faster than the fastest Nvidia chip, but it's not a chip game anymore. It's a systems game, and AMD is not building systems. So that's moat number two. Number three was the supply chain. Like the... pulling... I'm monologuing here.
It's Nvidia history, you know, before we get to the present.
Well, this really ties into right now and what is going to be a big issue going forward. Huang saw how big this was going to be, and you saw him out there. This is also 2024. The whole thing's out there drinking in Seoul, partying in Taiwan with the girl next to him, right? And what was he doing?
He was securing memory. He was securing...
I was going to say the Japanese drinks, that was all about memory, right?
That's right. Korean, but it's okay. You know, we all look the same to you, the Korean. I got to tap into my Asian. So...
So he's out there securing the supply, and this is super important on a few levels. It's not just that this gave Nvidia competitive advantage and that they could get memory and other people couldn't, and the other components. It's that one of the challenges, this is we've discussed this in the context of TSMC with the supply chain moving to Asia, is just a fundamental conservatism relative to like Silicon Valley hype and going...
Yeah. Because they're, I mean, they're usually like closely held businesses. They don't want to overinvest and be left holding the bag. It takes a while for the, no, this AI thing is real. Start building.
Jensen is the one that got them building. He's the one that got the whole supply chain geared up. Real, this is going to be a big thing. If it wasn't for Jensen Huang, the number of chips in the world today, the number of like all this stuff would be much, much less. Not just chips, but all the components, all the pieces. There's, you know, tens of thousands of pieces that go into an Nvidia rack. He was out there selling this vision.
And he talked about this, what, like he does so many keynotes a year, like he would do like five or six, and he actually made this very clear. I was at the... last time I interviewed him was... no, second to last time I interviewed him was in Taipei, and it was interesting because he's there and he's very clear he's speaking to suppliers, like to the ecosystem, and he's like, and he said this on stage, the reason I do these keynotes is I am communicating, it's the most efficient way for me to communicate to you in mass about this is where we're going, get ready, prepare these things, go forward. And he, for that reason, above and beyond just Nvidia chips enabling AI, he was the salesman in the supply chain that got Asia believing in AI so they could supply everything that's going on today.
Wow. Yeah. Well, and he got a lot of people on the American side funded, and I mean the creative financing that he has been overseeing over the last couple of years will also... threat to their position. So this is actually the key. This is the root of where this all went wrong for Jensen.
At the end of the day, the thing that he almost... that the company almost died for, and was the most brilliant thing that he did, was CUDA.
And he got so wrapped up in preserving CUDA as a moat that it led into all sorts of issues.
Well, one of them, but to mention the Neoclouds, which I... what do I call them? I call them something. I call them Neolabs in my update this week. I don't know what I was doing.
I read that, I was like, are there Neolabs in addition to Neoclouds? Okay.
Good correction live on the show here. The Neoclouds are because the problem with Jensen is supplying the hyperscalers and he's supplying the frontier labs, sometimes via the hyperscalers, but they all... it's a big fight in the long-term AI value chain who captures the margin. Right now there's not enough chips, so Nvidia's capturing all the margin. But in a world where there's plenty of chips, surely their margin's going to compress because I can buy chips from anyone, or I'll make my own chips if I'm an Amazon or a Google.
And so in the long run, Nvidia was going to face real pricing pressure if they only had a few buyers. If their only buyers were Amazon or Google or Microsoft or OpenAI or Anthropic, when you have fewer buyers, they have more power over you because they are more capable of coming up with alternatives, doing something else. Why has AMD made so much more traction against Intel in the hyperscalers, whereas enterprises and like the federal government all runs on Intel? AMD's not made any inroads there. Why?
Because they're relatively small, not sophisticated. They have some program that works on Intel processors, not on AMD processors. If you're an Amazon, you will put in the effort to make sure the AMD processors work well and can run everything because you want multiple suppliers. So that puts pressure on Intel, and because AMD can be better and can be cheaper, whatever it might be.
Nvidia did not want to become Intel, or they would in the long run... Intel, they needed other buyers who would be Nvidia-only.
That's what I was saying on Sharp China earlier this week, is Jensen obviously... the way I made it sound on Sharp China is it was all done altruistically because Jensen just wanted to see all these people succeed. But there was a real self-interest in doing that. However, what I actually said on Sharp China was there's more diversity in the market as a result of Jensen at Nvidia over the last several years and what he's incentivized to bring to bear in the market. And without him there, I'm curious... like I'll read this note from Brian Chao on Twitter. He wrote, 'If we lost Sam, we have Dario. If we lost Dario, we have Demis.' Nvidia is singular. The gap between Nvidia and AMD would be wider than the gap between OpenAI and Mistral, Europe's top AI company whose models don't rank in the top 40. Like how dire does this become for the rest of the AI ecosystem?
So this is the whole tension. So the issue is these need to be done in systems. There's really still no... the only... the TPUs are probably the biggest competitor in terms of having large enough scale to do training, where because you need systems. You don't just need chips. Like you're not doing training, real training on AMD chips. Yeah. Because they're not a system. They're just a chip. This has been AMD forever. AMD's actually always been awesome at hardware and just sucky at software. Not even paying attention to it. It's like, here's a super awesome chip, go figure it out. Whereas like Nvidia, ATI, that was a company that AMD bought that became their graphics division. Whereas Nvidia, their software, they had better drivers. Like for example, even if their cards weren't as good, but then you know, plus the idea of shaders and then doing systems, like they're so much more early to this than AMD. And so they're still definitely the best option, will continue to be the best option for most companies as far as training goes. Again, TPUs I think really being the other alternative there. The challenge is, and in a world... the system approach arguably makes for a more efficient outcome. I believe today Nvidia chips are still the best in terms of tokens per watt. So in a world where we're electrically constrained, that is arguably a good world for Nvidia. At the same time, the extent to which Jensen was ruthless in seeking to preserve those big fat margins...
Right, and their primacy, sure.
Was, as he pushed the line in a lot of ways. Like the one example is, we're scarce, everyone needs chips. Well, who's going to get the chips? Like, how many of those decisions were made... were the correct economic decision in terms of the short run?
In terms of basically giving Core all kinds of chips way back when.
Right, and say like, oh, you know, I don't know if this is the case, but Amazon, you're going to get a few fewer... like we don't, you know, we don't like other stuff, right? Picking and choosing like winners and losers.
I think the way to think about Jensen is he is a very systematic thinker and he really does think in the long run and he makes big bets for those long-run outcomes. That's how CUDA gets there. That is a big long-term bet that is not going to pay off for a very long time until it pays off hugely. So you get to the AI era. He can see all the ways that the AI era right now is super great for Nvidia's financials and in the long run is probably bad for Nvidia's moat.
And was extremely aggressive in trying to ameliorate that and sort of fighting against that. And so some of the share allocation thing, playing hardball with that, and going to bat for selling chips to China.
Mhm. Over and over and over again. Well, we'll get to that. Nvidia without Jensen and with Colette Kress in charge.
Oh, so this gets to the governance issue. Always a little dicey going from a founder and an engineer to...
It's not just as a founder and engineer. So my exposure to these companies on a personal level is usually through interviews, right? And it's always kind of interesting. You get a sense of different companies by interacting with them. Even if it's a company like Apple who doesn't give me interviews, also fits. But their PR team is very, very professional. But yeah, you get...
No sit-downs with Tim Cook.
No. Maybe the Ternus era will be a different story, but okay.
I got takes, John. I don't... I have...
I've always said like the company that always impressed me is Meta. Meta, super professional, get back to you super quickly, like very straightforward. It's just like a pleasure to deal with their team in general. I mean, all the companies are good, but like they always struck me as like it's a very professionalized operation.
Nvidia, it felt like you were talking to a startup.
Like kind of seat of the pants, like make it up as they go along, somewhat like ramshackle in a very charming way.
I was going to say that sounds like a compliment for a four and a half trillion dollar company.
Ten years ago they were a gaming company.
Five years ago they were a gaming Bitcoin company.
It's a company that moves so fast that this idea that they accelerated to a yearly cadence of releases to support this, moves so fast, has been so innovative, turns on a dime, does things like the Mellanox acquisition, by the way, in the face of activist shareholder pressure at the time, is because it is a company that was... to go back to that F1 analogy from the Tae Kim book, so sorry we didn't even finish my story, it's an F1 car that's built around Jensen Huang. It is engineered that he is the decision maker for everything, which is how the only company that's ever been comparable to this is Apple.
One of my first articles on Stratechery was when Microsoft, Steve Ballmer wanted to reorganize Microsoft, and it was beautiful timing because I had resigned June 30th, and I think it was on June 30th or July 1st he released this company-wide memo about a reorganization where they wanted to shift from being a division organization to a functional organization. This was like peak Microsoft Apple envy, like Windows 8 was coming out, we're going to have touch and we're going to do it better than Apple, and we're going to also reorganize the company to be like Apple, who's a functional organization. This was insane. A company like Microsoft is like... so a divisional organization, you have different business units and they're almost like their own little companies within the company, and then they sort of all roll up to the CEO. That is how when you get large enough, that's the only way to function.
Apple always prided itself on being a functional organization where we have engineering, or we have core engineering, and we have hardware engineering, and we have design, and we have marketing, and no one owns the whole thing. Everyone specializes. We hire specialists who are best at what they do. We drive collaboration to make them work together to come up with products, and they all flow up to the top where the person at the top is the one who's the product manager for the company.
He keeps all the plates spinning.
This was one Steve Jobs could not be... he talked about this endlessly. Steve Jobs was like a student of organizational design, and he was 100% on a functional organization, which makes sense for someone like Steve Jobs.
I was going to say, if you're Steve Jobs, the whole company basically works for you, right?
It is the ultimate manifestation of the CEO as ultimate decision maker. He built the F1 car before Jensen did.
And when he came back, Apple had gotten all divisional when he was gone. That when he came back, it's not just that he came up with new products, you know, working with Jony Ive, whatever, the iMac. He completely reorganized the company to go back to being a functional organization with him at the top.
And this is very powerful and effective. It does not scale. You can only do a couple products. And Apple today, they still say they're a functional organization, they still are to a much greater extent than anyone else, but you can see the seams more, and you have like different... when stuff doesn't work together right and things like that. But no other company can really pull this off because it just doesn't scale to multiple products, things like that.
Nvidia is like that early Steve Jobs. Early Steve Jobs 2.0. Everyone there is a specialist. They do what they do, and it all flows up to Jensen. When he talks about having 60 direct reports, that's what it means. Like everything just flows through him, and he makes all the decisions. And that's not to say people below him aren't extremely competent, extremely capable, and are doing the actual details, right?
But it is designed that everything comes up to him. It's a company that is a manifest... And how can they do that? At the end of the day, Nvidia sells one product.
It made... the reason this is why the GTC things were so amazing. People say, 'How do they do so many products?' And I wrote this ages ago when I first wrote about CUDA. I'm like, 'It's not so many products. It's one product.' Like this is the magic of software. Software makes your one product do lots of things. Even back to Apple, when Steve Jobs comes back, all they have are our Macs.
They had, and he reduced it, the famous 2x2. We're going to have a consumer personal... a consumer Mac and a professional Mac, or laptop and desktop. We have four products. He drew it, and that's all we're going to do. That Steve Jobs reduction in products was part and parcel of reorienting Apple to be a functional organization.
Nvidia's like that 100%. The two companies are so similar. And is the issue just a lack of personal charisma in terms of managing that going forward for somebody who's not...
In that sort of company, no one... you don't hire generalists. And Apple will tell you that, we don't hire... like I remember Apple University, like there's one of their precepts, we don't hire generalists. We hire experts, specialists in their field. The only generalist is the CEO.
And in retrospect, I was just talking with John at Dithering about this. We brought up the Apple comparison. One of the most fortunate things, it's hard to say fortune when someone died, right? One of the most fortunate things that happened to Apple is Steve Jobs' death was not sudden. He had to take a sabbatical for like six months or something because of his cancer, and Tim Cook took over, and then Jobs came back for like two, two and a half years before... before he died. And I think, I suspect that was a real wakeup call for Apple in terms of succession planning. Like, and succession planning isn't just... it's not identifying who's the next Steve Jobs.
What it is is professionalizing the organization. So it is an entity in its own right. It's not an exoskeleton that depends on the guy in the middle.
Sure. And so Apple, I think in that time period, 2008, 2011, became a much more professionalized organization.
More structure, more resilient in the anticipation we're not going to have the...
We can't find another Jordan, right?
That is going to be right, and so that's how they could seamlessly have someone like Tim Cook come in.
Not Steve Jobs, the exact opposite of Steve Jobs, but in many respects Tim Cook's in retrospect...
Makes me actually appreciate Tim Cook more.
It wasn't just all the supply chain stuff and, you know, all the things that he did there. He's the one that professionalized the organization. Like I go back to Apple University. Everyone talks about Apple University being a Steve Jobs thing. No, that was a Tim Cook thing. Like, I don't think it was worked out as useful as like a glorified HR training program now, but it was part of the overall how do we make Apple into a sustainable entity when Steve Jobs is gone.
And that's pretty interesting.
I don't think Nvidia's gone through any of... Well, I mean, that's one of the questions over the last week or so is like how...
He was talking like a month ago about working for 20 more years.
I saw an Information story where he was talking about working until he's 120 years old and running Nvidia until he's 120 years old. And it is pretty remarkable to step back and say, 'Oh boy, like what a weak board and no succession plan whatsoever for a $4.5 trillion company.' But Jensen was dynamic enough to make it all work.
No, that's arguably why Nvidia is so a four and a half to five trillion dollar company. Right. When I was at Apple University, part of the reason it was amazing, perfect internship. Very glad it was not my job. Learned so much about the company. Had very like high level access. It was amazing. But part of the, you know, my worry, like this whole concept seems fishy to me. If you're trying to define what makes Apple Apple when you write this stuff down...
It's like ossification, right? You're starting the process. But that's the case for any company. We talk about disruption. We talk about companies. We talk about Apple right now.
Like when you... the process of building a company, when you start out, when you're a startup, you're... I might have used this analogy before. I'm going to pretend I haven't and you're fresh.
You're going for a walk, right? Your eyes are open. You can go anywhere. You might see a path. You take the path, see where it goes. Suddenly you're like, you're not very efficient, but you can do anything. Then maybe you get a bicycle. Bicycle, you're much more efficient. You can move faster. You're a little more constrained in where you can go. You can't like hike...
up some boulders or whatever. Bike up a flight of stairs. Sure.
Right. You're such a city slicker. I think I've been using my head. Oh, I'm out mountain biking in the countryside. Just like walk like a prairie or something going up stairs.
Okay, continue through the corporate progression. What's after bicycle?
Uh, maybe like a motorcycle. Okay, much more powerful, go much faster. Maybe not quite as flexible or weaving in and out of traffic or whatever you're like as you can on a bike. Then you're in a car.
Kind of need a road. Might go off-road a little bit, but not too much.
Eventually. Then you're in a truck. Eventually, you're a train. You're like a bullet train where you move so fast and you're so efficient, but you can only go in one direction, right? And you see this with like these aggregators, right? They seem like these impenetrable business models, and then you ask them to do one thing slightly off of their business and they just utterly completely fail. They just like, cuz they're so tuned to do what they do and everyone in that company is a cog in the machine and then they just can't. This is what happens to companies. They get more efficient. When this is where the analogy, the F1 car goes further. If Nvidia never got to be a bullet train, they just built the fastest freaking car on the planet, right? Which meant no good governance, no succession plan, none of this. But that's also why they were such a dynamic company. Yeah.
And well, you don't want to do an F1 car's crash, right? So the...
Well, I mean we've seen this in literal F1 cars where the Red Bull is tailored to Max for stopping and historically has been basically impossible for any teammates to drive and Max can push the boundaries in all sorts of great ways and knows how to drive it. Uh, Jensen clearly knew how to drive Nvidia and going forward, you talk about his bets, like I think about how much money they've sunk into open-source AI and stuff like that. I worry that there are just going to be fewer of those bets going forward. Uh, and I worry what that means for the diversity in the ecosystem. But you said that this could be a good thing.
Well, it depends on who you're talking about. For I think startups, uh, this is a disaster.
There was one entity with tremendous financial resources whose incentives were aligned well, no, but their incentives were aligned with open models, uh, with, you know, more customers. Also, if you started out and you built on Nvidia, you had a whole, there's a bunch of developers that could work for you, like, and you could, there was lots of places to buy Nvidia chips like, or Nvidia servers like the whole Neocloud, Neolab, whatever it is. Um, all Nvidia was very aligned with that world and they basically, you know, Nvidia still is, right? But they've lost the guy who got it, who was, you know, the patron of that world. And uh...
Well, and he had the personal cachet within that company to rationalize any of those bets.
It didn't mean cachet, the company did what he wanted. Yeah.
Yeah. And with the board, with shareholders, in Jensen we trust, and everybody got rich following that plan.
Yes. Now, it will be interesting to see whether any of that trust can be established under a new regime. Um, I'm not terribly worried about them like in the immediate future, but it's murky for everybody right now. So...
Well, well, the, I mean though the long term, you know, in the long run, the other threat, we should get to the China point. Yeah. Um...
That's coming after the ad read for sure.
Uh, well, just to let me sort of preview the discussion because we've talked about it before. The reason why he was so obsessed with China is because who's going to create an alternative to CUDA? Uh...
Somebody who can't access it.
That's right. And who has the resources to do that?
Okay. The Chinese government is not going to create a CUDA alternative. Mhm. People actually using AI, actual engineers, actual companies, that this is where everyone in this debate from Washington DC drives me bonkers.
Oh boy. I am going to be playing the role of a deep state DC guy on this show.
Don't create an ecosystem like CUDA with a top-down dictate. Mhm. You can start the project. You can say, go do this. Ecosystems are pulled into existence by the developers who use them and want them and leverage them. Yeah. Those developers, those people who would pull a CUDA alternative into existence, all would prefer to use CUDA. All would prefer to use Nvidia chips. As long as there were Nvidia chips in China, the government can want everything they want. They don't get to dictate the choices of like individual developers using like...
Were they going to look over your shoulder what programming language you use?
Sure. There was never going to be the necessary pull, the necessary demand for an alternative as long as Nvidia chips were available. Mhm. That's why you wanted to make sure Nvidia chips were available.
Okay. Well, we will come back to that point. For now, today's show, as I said earlier, is presented to you by Stripe. Today we are focused on Stripe's digital wallet, which is Link, and which can be used for agentic purchases across Muse, Intellect, and other consumer-facing agents. Agentic purchases with Link were up 38x in September. So clearly this is a use case that will continue to be relevant for all of us. Ben, I have seen people say that agentic commerce was solved with the introduction of Link. Uh, curious what...
No, it was not solved. Well, it's funny. There was a bit of a controversy about this last week. Like there was a tweet like someone who, uh, they made a purchase from Best Buy that was like rejected because one was like a third party, and they're like, 'What's going on?' And they looked out like the Link thing. So, I actually like did some research on this.
Okay. Um, so what needs to be solved?
Well, it's sort of getting solved now. So, you have this issue of do you want to actually like give your credit card to your agent and that agent gives your credit card to a merchant?
Sure. I'm sort of fine with it for the record.
Well, it's getting a little dodgy there. So, let me go back actually a few years. Uh, when I was in Taiwan, I was at the equivalent of a Home Depot there, which is called, uh, there's a B in it, but it's also orange. Okay. So, it's very confusing. So, but um, I think...
Signature Home Depot or is this off-brand Taiwanese imitation?
So I bought something and I go to return it, which by the way, in general, returns aren't generally a thing in Taiwan. Uh, but in this case, the hardware store accepts returns, but given returns are a thing, it wasn't very easy. And so, I had to like give them my credit card number for the return, but I had paid with Apple Pay. Now, this is relatively early on with Apple Pay. And I didn't realize at the time when you pay with Apple Pay, the merchant does not get your credit card number. Okay. What happens is Apple, there's like a, in conjunction with the issuer, they basically get a token that is like a unique number that's used just for that purchase. So if you actually go to your wallet and you look on a purchase under a card and you like, there's like the info button, it will tell you the number or the last four digits of the number that was actually used for that purchase.
Ah. The purchase is different than your number.
And I was actually able to find it in line. I go to this on the spot. So I was like trying to figure out, I paid with Apple, because I gave them my number, like that doesn't match, and I, oh, I gave them the number. Oh, there it is. My return got processed. So there's, so in general, there's this concept of, because credit card security is obviously a big deal. I mean, it's kind of crazy. You used to do the, do you remember that?
Do you really remember that?
Of course, I'm not that much younger than you are.
But it is pretty funny because there was, I mean, the first 15 years of digital life, everybody was terrified of putting their credit card online and what might happen, right? And now we just take security for granted.
Well, you also didn't realize the extent to which companies were solving these problems and protecting you. Like the Apple Pay one is a good example. I actually didn't even realize that's what was happening.
When I thought about it, it's like, oh yeah, of course that makes sense. But and it's a much safer transaction. It's always funny when this would happen in Taiwan sometimes. I don't know if it happens in the US. Uh, the super large purchases you would have to use your card, not Apple Pay.
That's actually interesting. That's actually backwards. Apple Pay is much more secure than a card. Okay.
Uh, anyhow, neither here nor there. So you, but you have this challenge that is being solved or that is presented to itself, like how do you not just have these 16-digit numbers? If someone has the numbers, why can't they just do what they want with them, right? So, when the agentic Link thing comes along, there's no, you just kind of have to hack it together. So, Stripe is a credit card issuer.
And so, what they do is they basically issue a credit card on the spot for your purchase called a virtual card. So if you're making a purchase at Best Buy for $150.28, they will issue an anonymized...
In like, your number actually goes through. They issue, like, they actually generate a credit card that is only useful for $150.82. Use that credit card to pay and then you bought the credit card from Stripe.
That was like a hack to get this working, right? But there's lots of problems. Number one, you had this issue where actually Best Buy was trying to charge multiple entities because it was a third-party marketplace plus Best Buy. If you make this purchase on your credit card statement, it shows up that you like bought like computer software because like Stripe's, you're actually buying from Stripe the card. Okay? So, you don't get like your rewards, like some cards, like if you buy from a restaurant, you get 2x points or 3x points. So it's like a brute force substituted in, but it is not ideal in the long run. What's actually happening now, the solution that is now coming online over the last few months is Visa and Mastercard and American Express. I think American Express just launched a little bit ago, is the whole thing like Apple Pay. They generate the token. They do it. Then they know what the card is. They can do the reward thing. They can divvy up the payment because they're the ones who are holding the actual card. Whereas Stripe was like, they didn't hold the actual card. They were in the middle of a transaction but they didn't have like ownership on the other sides. So the answer is to have these tokens that actually go out, they handle the purchase, your card is completely obfuscated or not seen by the other side, but then they can translate, oh, this was a restaurant purchase, give him points, or they can split up the purchase or whatever it might be. But actually getting this out there depends on the issuer, like the bank has to then be signed up just like Apple Pay. Visa and Mastercard get Apple ready to go. It took a while for all the banks to support it, even now because I just changed phones. We have to have another phone discussion.
Um, like, my Tom and Cracker are such a pain to get on to Apple Pay. Like, it's just a big problem. So the long and short of it is there was this controversial tweet last year, or last week, about complaining about this is not going to work.
He needs to go to his bank. He needs to get on the agentic token. Um, and like, cuz probably just his card isn't yet supported to do this. So, he's getting the hack, which is a virtual card, but Link has most of the major cards already cooperating.
The whole point is this is really difficult and hard.
So, there are clients that can do it.
Well, that's like the Stripe value proposition, right? Like this stuff is so messy. There's a new term I'll use on you. Everyone wants to stay outside of PCI scope. Uh, PCI scope is a set of rules that govern how payment information can be used. No one wants to mess with that, right? And so that's why the whole reason the virtual card, like Meta doesn't want to get into that business. Sure, it's a messy bit and it's not just that, you want to scale the whole world.
So I remember I talked to Netflix ages ago about this where like they need to support gift cards or they need to support like local payment methods, like all sorts of weird ways people pay for stuff. Credit cards are a thing, not everywhere. And you want that all abstracted away. You want someone to do all that and you just have one...
Someone else's problem. Sure.
So that is the idea of Link.
That's right. So, uh, that's the value. So Link is now becoming a thing for agents from a user perspective. The idea is you could, it's safe. You can set rules. How much do you want to spend on XYZ? It gives control and visibility to the human owner on what the agent is executing and you can approve what it spends. It can come back to you if the final price changes. Uh, and eligible purchases can even come with protection if the agent makes a mistake. So, there's Link. That's a longer ad than normal, but that was a pretty interesting conversation.
Well, it was funny cuz we were talking about Link this week and I saw that tweet and there's a lot of discussion about it and so no, I actually, I called Stripe up. I'm like, I'm going to leverage this here. Can you explain all this? Like, this seems like a bad situation. He's like, well, because this particular person, it's like a hack. Okay,
Because, but in the long run agentic tokens are actually a thing. So there you go.
Ah, fair enough. All right, well back to the show, back to Jensen and the aerial vibes here. We talked about Jensen's greatness. I am going to skip to this question we have from Samuel. He says, 'What's the Vegas line on the over/under for Jensen affairs? Personally, I find it naive to think that an affair with a Chinese spy whose company paid Nvidia billions of dollars had no impact on Jensen's decision-making. Sure, it's pocket change to Nvidia, but studies show that doctors prescribe significantly more of a drug when a pharmaceutical rep buys them a single lunch. Money plus sex equals influence.' Uh, so setting aside that second sentiment there, Vegas line on the over/under for Jensen affairs. While we were recording this, Reuters did come out with reports of another Jensen Huang affair with an Nvidia employee. So over/under, I'm going to set it at five and a half in terms of how dark this could get. But I don't know whether you have any thoughts.
It's not really a conversation I want to get into.
I'm sure, I would guess it's much higher than that. And there's a...
Not to get like all like...
Sociologist on you or whatever.
Why do great men do great things?
That's one of the traditional reasons. You're strong, powerful, you get access to lots of women. Like, uh, I'm not excusing it obviously in our sort of moral context, that's not something that's allowed. It's also not something that particularly shocks me.
No. And I don't think that the...
And I'm not condoning particularly the employee thing to be clear.
We said it was a red... we're moving into the we're not condoning area to be clear.
And I only read that question because I literally got a text message about an affair that Reuters is reporting on as we're recording this. Here's the question and there are probably others. Um, Samuel takes as a fact that Alice Huang is a Chinese spy.
That is a point of contention at least, uh, in your coverage over the last couple days. What do you think as we look at what happened this week? What's your mental model for how we landed here? It is unbelievable. We shouldn't take for granted how shocking it is that two weeks after the initial New York Times reporting, Jensen Huang is no longer the CEO of Nvidia, the most important AI company in the entire world. And even Monday, that seemed inconceivable. So, here we are. How did we get here?
Uh, two weeks or several years. Mega Speed has been a topic. The New York Times wrote an article about them a year ago. Bloomberg wrote an article about them a year ago. Uh, these questions about like, is this a front? Is it being funneled to China? This is an issue that has come up. So yes, what changed this week? Oh, Chinese spy, that changes things. What is ByteDance, a Chinese state company?
Uh, I don't know whether they are classified as such,
Right? But yet, why? What's the objection to ByteDance owning TikTok?
Uh, military-civil fusion in China.
Well, no, I'd say it's more just that any company in China ultimately serves at the pleasure of the state.
And to the point, like that's actually like a law. Like basically if the Chinese government tells you jump, no matter what company you are, you have to answer how high.
That's right. Now, and why this episode, we can ask to what extent does that apply to the US as well, but yes, so the issue is at the end of the day, everything is the Chinese government and nothing is the Chinese government.
Everyone is a Chinese spy and no one is a Chinese spy.
What I, the way this was framed is like Alice Huang was recruited from a young age, trained in the dark arts, which by the way...
That does happen to be clear.
That's a Chinese move. Sexual blackmail is definitely an MSS specialty.
Yeah, it's not just, I believe it's used by spy services everywhere. Um, it's not exactly a new tactic.
I don't think that was the case here. Now, was she in contact with, by virtue, like her husband was originally in this game. I think he ended up in prison for something or other.
Uh, but I think he introduced her to Jensen way back when.
There's this whole world of like smuggling, moving stuff around. This is a large world, okay? Like there's like, whether it be because of sanctions or like just, it's a big market. There's people who are in this market. If you're in this market and you're suddenly moving very important equipment for billions of dollars worth, is it likely that the Chinese state apparatus knows who you are and may be in touch?
For sure. But so is that a spy?
Right. Well, I do think that's a useful distinction to the extent that we have normies listening to Sharp Tech. I think we have this note from Nerra who's talking about his family talking about Alice Huang and how Jensen fell for this beautiful Chinese honeypot. Uh, I can confirm that this has crossed the mom and wife threshold. I've been getting text messages all week about the honeypot scheme that Jensen fell for. How could this ever happen? And I'm not sure that's really an accurate way to model what happened here.
No, but it's an accurate way to make people care about everything else.
It's an accurate way to initiate the endgame.
Yeah. And I think that is, um, and this week was the endgame.
If Alice Huang was such a threat, why was this not addressed months ago when she was identified as an issue? Right. Um, or since she was...
It's Alice Huang, just for the record. Same as my wife. You said Alicia.
Oh, I've been saying the whole time. I apologize.
Delights me to no end. Her name is Alice. Yes.
If Alice was a threat. I mean, that's one of the number one questions when you look at the timeline here.
But Chinese honeypot, very salacious, gets a lot of attention, amps up the pressure.
All eyes on. Next day, boom, employee affair.
Next day, boom, another affair. And the strong hint from everyone in DC, more is coming and it's actually way worse.
Mhm. What could be way worse?
Uh, I don't know. You want to guess? I have some guesses.
That the chips flowing through Alice Huang were not going to Alibaba. We're ultimately ending up with the Chinese military.
Yeah. So yes, duh. There's an aspect of the China conversation, which by the way, Jensen Huang is very much guilty of, that is very much treats its listeners like they're stupid. If GPUs are so important and valuable, if they're going into China, the military is going to get their hands on them to the extent they want them.
Sure, of course they are.
I mean that's why Alibaba versus PLA is a distinction without a difference as far as I'm concerned.
And that applies to the US as well. Like, does the US have civil-military fusion? Of course we do. Right. Like...
Is Microsoft a military company? Is Amazon a military company?
