Thompson was using an analogy of walking, bike, car and train for how companies get more efficient but less flexible as they grow. Applied to Nvidia, the missing governance and succession plan is the price of its speed.
Sharp Tech with Ben Thompson · 2026-11-13
Permalink →
Listen →
Sharp Tech with Ben Thompson
Around 28:33 into the episode
Ben Thompson
Right. You're such a city coastal elite. I think I've been using my head. So I'm outside. Just like a prairie or something. Going upstairs. Okay.
Andrew Sharp
Continue through the corporate progression. What's after bicycle?
Ben Thompson
Maybe like a motorcycle. Okay. Much more powerful, go much faster. Maybe not quite as flexible or weaving in out of traffic or whatever 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, 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. This is where the analogy of 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…
Ben Thompson
And well, you don't want to do an F1 cars crash. Right? So the. Well,
Andrew Sharp
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. Jensen clearly knew how to drive NVIDIA. And…
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
Thompson compares Nvidia to early Apple, where everything flowed up to one person. Apple became more resilient because Steve Jobs's illness forced it to prepare; he doesn't think Nvidia ever went through that.
Sharp Tech with Ben Thompson · 2026-11-13
Permalink →
Listen →
Sharp Tech with Ben Thompson
Around 24:40 into the episode
Ben Thompson
sort of company, you don't hire generalists. And Apple would tell you this. We don't hire, like, I remember what it was, 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.
Ben Thompson
And in retrospect, I was just talking with John 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 he died. And I think I suspect that was a real wake-up 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.
Ben Thompson
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
Andrew Sharp
master of public that
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
Thompson looks at what a finance-minded interim CEO means for Nvidia's role in funding the AI buildout. He thinks credit for the neoclouds, already tightening, gets tighter without Huang, and compares it to the railroad boom.
Sharp Tech with Ben Thompson · 2026-11-13
Permalink →
Listen →
Sharp Tech with Ben Thompson
Around 1:10:07 into the episode
Ben Thompson
…some policy issues, but then if you layer on a corrupt element on top of it, it just taints the whole thing. And obviously, this whole Jensen Huang,. I was going to say, I mean, like it's done now. I'm sure that we're going to have way stronger chip bans. You're wearing blacking them. Not just for
Andrew Sharp
Jensen, but for the end of your sell chips to China take at this point because it's just over.
Ben Thompson
This is for a chip industry. Like this is a disaster, in my opinion, in the very long run. And I also, you know, there's lots of things. It's bad for the whole AI bubble thing. We talked about in the context of NVIDIA's risky business. Like NVIDIA, we need to, you go from investing free cash flow to credit to equity. We're well in the equity stage. What comes after that? Well, what needs to come after that is you actually start making sufficient revenue so you can get back to free cash flow, not just paying off your debts, but then investing. And getting from
Andrew Sharp
here to there requires a bridge.
Ben Thompson
Requires a bridge. And NVIDIA is, and Jensen Huang, is out there backstopping companies, supporting companies, like being the bank for the bubble. And is a temporary CEO, particularly a CFO, is thinking about numbers, get away from the banks. And they get away with that. Yeah. Is a future CEO going…
Andrew Sharp
can I ask you something? How dependent on Southeast Asian chip sales was NVIDIA's business today? Because your point on the long-term interests, obviously, it makes sense to me. And I was a skeptic over the last several years. And eventually you wore me down and won me over. And I think your…
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
Thompson describes the logic of the chip-ban hardliners: recursive self-improvement is coming, superintelligence ends the game, so only the next few years count. He argues that belief lets people wave away long-run costs.
Sharp Tech with Ben Thompson · 2026-11-13
Permalink →
Listen →
Sharp Tech with Ben Thompson
Around 1:07:54 into the episode
Ben Thompson
mindset. What if you're wrong?
Andrew Sharp
What if you're wrong about the chip ban?
Ben Thompson
What if you're wrong that actually the world, it's not just the next two years that matter? Like there's such an excuse that comes with, actually, I don't need to make any trade-offs because there are no trade-offs because there's no long term. Yeah. All that matters is the next three years. And when you have this mindset, you throw away your chip industry, you throw away your ecosystem advantages in the fullness of time, like over the next or in the very, very long run, because it doesn't matter. You just have an excuse to do all these sorts of things. And meanwhile, you have, so I think you had a fusion. You had the EA Doomer anthropic sort of area, which believes this. Then you had the actual real national security apparatus is super anti-China and is like, you know, all they're thinking about is war with China.
Andrew Sharp
And members of the Trump administration, I would say the majority of the Trump issues. Well, this is totally different Jensen. Right.
Ben Thompson
So it's bipartisan. Like, because usually that national security
Andrew Sharp
well, I mean, Jensen really torched his relationship with members of the intelligence apparatus on the left under Biden. Right. And then under Trump, it's been friendlier, but people like Marco Rubio, people like Scott Besson, I think everybody's been sort of tapping their foot saying, when are we…
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
Thompson traces Nvidia's feud with Washington to the first China chip rules, when Nvidia quickly shipped a cut-down chip that stayed just inside the law. Officials read it as circumvention, and Huang read their rules as foolish.
Sharp Tech with Ben Thompson · 2026-11-13
Permalink →
Listen →
Sharp Tech with Ben Thompson
Around 56:49 into the episode
Ben Thompson
…the benefits now, is thinking long-term. How do I preserve this? How do I keep what I have? This, and I think this, and that's why he wants to sell to China. It's super important for his moat to sell to China above and beyond. Of course, he wants to sell everywhere so he can make more money.
Andrew Sharp
More customers also sort of preserves the pricing power for NVIDIA. Question: So, I think,
Ben Thompson
so I think a really illustrative episode: NVIDIA had no presence in Washington, D.C. So, the China chip ban comes out and it gates the first version under Biden, and it gates what you can sell based on not on flops, but on like memory bandwidth, which was actually a relative. I actually thought the V1, I disagreed with it, but I thought it was that was the way to construct it. Because what you're constraining is the large systems you can use because those are limited by bandwidth. So, what did Jensen do? Within a month, they had come out with a new chip that had less bandwidth. And to be clear, that meant it was a worse chip. It was not as useful for training because it had less bandwidth. They literally go into chips and just like burn off the connections or like and so you have to link
Andrew Sharp
more of the chips together to get comparable power, correct? It's
Ben Thompson
not going to be comparable. Like, you're just fundamentally limited if you don't have a sufficient bandwidth and you're not going to be able to scale as much, do as large models, all these sorts of things. My view, my interpretation at the time is that immediately put NVIDIA behind the eight ball…
Andrew Sharp
yeah. I mean, it's funny because when you zoom out, there's a parallel here to Dario Amade and Anthropic saying, We've got a signed contract with the United States government that says we can dictate this, that, and the other thing, and then trying to dictate terms to the U.S. government how. Their…
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
Thompson argues against the hardline chip ban, which is now expected to tighten after Jensen Huang's exit from Nvidia. His point is that the US used to sell older technology abroad so other countries never had to build their own, and a ban pushes China to build a full alternative stack.
Sharp Tech with Ben Thompson · 2026-11-13
Permalink →
Listen →
Sharp Tech with Ben Thompson
Around 1:03:28 into the episode
Ben Thompson
…an ecosystem for Windows Phone. It didn't work. If Microsoft can't do it, is the Chinese government be able to do it? No, why? Because Android was right there because iPhones were right there. If NVIDIA chips are right there, the Chinese government is not going to bring an alternative into being.
Andrew Sharp
…What an own goal from the United States. And every six months, there are reports about Chinese breakthroughs on the manufacturing side, EUV, what have you, Huawei chips. But each cycle here, every six months, every 12 months, they are still so hooked on NVIDIA chips, particularly for training.
Ben Thompson
In the very long run, in the fullness of time, China is going to have their CUDA equivalent. They're going to have their manufacturing equivalent. And then they're going to take that ecosystem. They're going to take that equipment and they're going to go out from China and sell it all over the world. We have written the deaf. All of our semiconductors, we're getting to that for all our semiconductor equipment for CUDA, for everything. That is the cost of the chip ban. That is why the U.S. policy for ages has been to sell N minus two, N minus three versions to other countries. So they never have the imperative. If you can buy the older generation, it's still way better than what you have. Everyone who needs it isn't going to take the time to help this little startup making a chip. Good luck. I'm going to go buy the two-generational NVIDIA chip, which is a gazillion times better because I need it now for my business. We did this with CPUs. Like you go back. We've done this for all technology. We've done
Andrew Sharp
this with jet engines, with cars. It prevents
Ben Thompson
competition in the long run, and it makes a ton of money. So you can fuel RD and innovation going forward. And so you maintain, you own the technology. You still have a meaningful advantage because you're on the cutting edge and everyone else isn't. All the people that you're disadvantaging by…
Andrew Sharp
five years. Yeah. It
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
Ben Thompson is explaining why a data center bottleneck could actually help hyperscalers and chip makers, since an outside force slowing the build-out gives revenue time to catch up with spending. He argues the scarcity phase of AI will last longer than people assume.
Sharp Tech with Ben Thompson · 2026-11-06
Permalink →
Listen →
Sharp Tech with Ben Thompson
Around 54:05 into the episode
Ben Thompson
…revenue come? Once the revenue comes, all these things unwind themselves in a very positive direction. And so if there is an external force that forces a slowdown such that revenue can get there in time to relieve all these capital pressures, everyone could win. This could actually be a diffusing.
Andrew Sharp
Win, win, win, win, win. Sure. But
Ben Thompson
especially the hyperscalers. And so, yeah, it's, yeah. Everyone thinking about AI jumps to the final state where intelligence is a commodity. And that state might take a very, that implies infinite compute. Yeah. And it just might be the longer it takes to get there, the more analysis predicated on intelligent being a commodity isn't actually useful to understanding the world as it is. In a world where there's infinite power, infinite compute, what is like, what does NVIDIA do? Like, there's going to be other chip products, right? In a world where you're limited in power, you have to repurpose your data centers. Suddenly, whatever margin you have to pay to get the most efficient, most powerful chips, it suddenly makes a lot more sense, right? And so it's this, the analysis in a market defined by scarcity versus a market defined by abundance is fundamentally different. Completely flips. And I think the time defined by scarcity has probably been elongated.
Andrew Sharp
Okay. Well, it'll be interesting to see whether there is ever a CapEx slowdown because at this point, I'll believe it when I see it. Whenever we get there, it will be fun to look back at some of the CapEx we've seen over the last couple of years and just how consistently insane it was. Like, I…
Ben Thompson
Well, in the short term, it's definitely a problem because they're selling chips. There's nowhere to put them. Their market being limited. On the other hand, overall, this has always been a counterintuitive point about NVIDIA, which is the one I just made. A world defined by insufficient power or…
Andrew Sharp
machine, the most efficient machine will be worth the upcharge.
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
Andrew Sharp had just raised the newly elected Texas governor's promised data center moratorium. Ben Thompson explains why a pause in Texas, with its own grid and cheap gas, matters more than any local ballot result.
Sharp Tech with Ben Thompson · 2026-11-06
Permalink →
Listen →
Sharp Tech with Ben Thompson
Around 06:42 into the episode
Ben Thompson
wow. This is a very, very educational episode. That's right. So many
Andrew Sharp
fun facts here. No, but
Ben Thompson
that's why that is the one that really matters because the presumed backstop for all data center opposition has always been, we'll just build it in Texas. Everybody's
Andrew Sharp
going to Texas. Yeah. I mean, they were on track to be home to 25% of U.S. data center capacity by 2030. Hundreds of people. If you're talking to anyone,
Ben Thompson
the assumption is it was going to be way more than that. Wow. Like just going forward, yeah, it's fine. New York's going to pass a moratorium. Even Ohio, which is a very significant portion of it, that's, you know, might end up being a lot of people. Well, we have 50 different states
Andrew Sharp
that can compete for these projects, but infrastructure-wise, Texas, obviously. Has all kinds of advantages and may not be replaceable. Well,
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
Ben Thompson says he can't stand how most of big tech talks about AI, and offers a line he says he'll keep repeating. His point is that AI companies fail to show consumers which problems in their own lives AI solves.
Sharp Tech with Ben Thompson · 2026-11-06
Permalink →
Listen →
Sharp Tech with Ben Thompson
Around 11:31 into the episode
Ben Thompson
Yes. Yeah. And why not, given the version of the future that's been sold? And this is where you go back to globalization. People didn't have a vote. Right. The factories just closed. In this case, once people had a vote, they're not too happy about it.
Andrew Sharp
…to the spring, I said then I still feel now it's an understandable human response given what AI portends, the way it's been marketed, and the ongoing centralization of wealth among a handful of big tech companies. People just want a chance to stick their finger in the eyes. And why not? I mean,
Ben Thompson
I can't stand most of big tech in the way they talk about AI. Like we, the, you know, I'm very proud of this one. I'm going to have to repeat this one every podcast for the next several. If you love advertising, you're optimistic about AI. Okay. Because why do I like advertising? Because advertising, good advertising helps the customer realize they have a problem in their life and then it gives them the solution to it. Now, sometimes that's very cynical. You're like creating a problem that doesn't exist and then saying you'll answer it. We've seen lots of political ads the last few months. We know exactly how that works, right? But sometimes it's true. It's like, I actually, yes, I could actually use that thing. There's a problem I did. didn't have, and that's going to help solve it. This is very analogous to the AI question in two ways.
Ben Thompson
Number one, everyone trying to create AI for consumers fails to do this. The issue isn't that you're not showing people, you're not showing people like what are problems you have in your life and how AI solves it. Like, that's why the AI show and tell thing we do. Like, we got tons of responses. I…
Andrew Sharp
Well, and I will also add that they're building a technology that is exceptionally expensive to build and doesn't have a clear application beyond saying, Well, this can actually just replace wide swaths of the workforce. No, exactly,
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
Ben Thompson is explaining why he thinks AI has terrible marketing to ordinary people, after arguing the industry treats humans as interchangeable cogs. He ties the data center backlash at the ballot box to how the industry talks about AI.
Sharp Tech with Ben Thompson · 2026-11-06
Permalink →
Listen →
Sharp Tech with Ben Thompson
Around 12:35 into the episode
Ben Thompson
…AI has terrible marketing in the like it can actually make your life better in part because it feels like no one in this industry has any sympathy or understanding for normal people, which why not? They're all like crazy rich, following around the coast, no conception of how normal people live.
Andrew Sharp
Well, and I will also add that they're building a technology that is exceptionally expensive to build and doesn't have a clear application beyond saying, Well, this can actually just replace wide swaths of the workforce. No, exactly,
Ben Thompson
…industry, but huge chunks of the industry are just, they have such a narrow view of the world. Humans fit in this column, and the measurement marks on that column are like IQ points. And that's how they rank the whole world. And they're like, oh, well, AI is climbing up the column, gonna wipe you all out. Think how much money you're putting in. And so, implicit in this isn't just the fact that the only message that is coming out of Silicon Valley is AI is coming for your jobs. Implied in that is all of you are worthless. All of you are uninteresting to us. And this is why I go back to
Andrew Sharp
all of us. I didn't vote against data centers, but all of Americans are raising their hand to the sky and flipping the bird to all these tech companies.
Ben Thompson
Absolutely. And why wouldn't they?
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
Andrew Sharp had just said some of the misinformation about data centers is easy to disprove. Ben Thompson argues the industry keeps fighting the details when the real driver is a broad distrust of technology and AI, days after voters backed local data center restrictions.
Sharp Tech with Ben Thompson · 2026-11-06
Permalink →
Listen →
Sharp Tech with Ben Thompson
Around 10:26 into the episode
Ben Thompson
And, but you understand why people come up with the people are stupid because some of the fake news stuff.
Andrew Sharp
Well, some of the misinformation surrounding Against Data Centers, is just easily disproven or refers to the way data centers operate. But the problem is you're
Ben Thompson
operating on the wrong level, right? You're trying to refute facts when what you're actually dealing with is sentiment. And it's the sentiment that is attracting stories. And that's why people will discard, okay, fine, that's not true. And they'll find another reason to oppose the data centers. They just oppose the data centers. And why do they oppose the data centers? Like there is just a fundamental distrust and unease, I think, with technology generally and certainly with AI specifically. That's why I do think this isn't, even though it was all about data centers and people are talking about the environmental costs and electricity. I think people
Andrew Sharp
are just voting against this version of the future.
Ben Thompson
Yes. Yeah. And why not, given the version of the future that's been sold? And this is where you go back to globalization. People didn't have a vote. Right. The factories just closed. In this case, once people had a vote, they're not too happy about it.
Andrew Sharp
Absolutely. I mean, if we go back to the spring, I said then I still feel now it's an understandable human response given what AI portends, the way it's been marketed, and the ongoing centralization of wealth among a handful of big tech companies. People just want a chance to stick their finger in…
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
Thompson is going through details from a 400-plus page S-1 filing and says Anthropic's business looks strong overall. Here he explains that about half of Anthropic's revenue comes through the three big cloud platforms. He describes how Anthropic records the full payment as revenue even though roughly a third goes to the platform.
Sharp Tech with Ben Thompson · 2026-10-30
Permalink →
Listen →
Sharp Tech with Ben Thompson
Around 12:58 into the episode
Ben Thompson
…on all the clouds. Yeah. Like they, they, you know, the, they're, they're always on Amazon and then Google, and now they're on Microsoft. And that was a huge, I think that's very underrated in why they took over the enterprise under OpenAI's nose. OpenAI was stuck on Microsoft. Like, and so what
Andrew Sharp
advantage did that give Anthropic with enterprises?
Ben Thompson
Because if you have all your data on AWS, you would rather use the service on AWS. And like, so, so there's another thing. Sorry, there's, it's an S1. It's like, I've seen 400 plus pages. That's right. So a lot of threads. So about 50% of their revenue is coming from the big three via those services. So it's coming through AWS, coming through Azure, coming through Google. And it is confirmed that they're doing kind of sketchy accounting, depends how you want to define it, which is if you pay Amazon $100, that their Anthropic is recording that $100 as revenue. Anthropic revenue. And then in reality, like 30, a third of that or so depends, it sort of varies. I think people were saying about 37% goes to the provider, the platform provider, the AWS, the Azure, the Google. So they are putting that 30% into sales and marketing.
Andrew Sharp
I was wondering what the sales and marketing costs are. So that's part of it. Okay.
Ben Thompson
A little bit of BS accounting, in my opinion. So for two factors. Number one, I think it's fair to argue that they should not be recognized at 30% at all. They should only be recognizing the portion that is going to them. And then their revenues would be lower.
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
Thompson is talking about how companies use AI. In the lines just before, he says improving the bottom line means replacing humans, and the same turn then turns to the top line, where AI multiplies what people can do. He argues that this second kind of gain usually takes longer and comes from new AI-native companies, which makes the timing of the AI spending a problem.
Sharp Tech with Ben Thompson · 2026-10-30
Permalink →
Listen →
Sharp Tech with Ben Thompson
Around 43:36 into the episode
Ben Thompson
…reason they're worried about Anthropic is like, we're the right relationship. We're the wrong relationship. That's right. But is what they see with companies is companies who are focused on using AI to improve the bottom line struggle. Because how do you improve the bottom line? Replacing humans.
Ben Thompson
How do you improve the top line? Multiplying humans, right? There's this difference between multiplying versus replacing. And when you're multiplying humans, you're basically creating entirely new things to do, new workflows, new opportunities. The challenge for this whole trade in general is that usually number one takes longer. And number two, usually comes from new companies that they come in, they disrupt the incumbents by virtue of the AI first company taking on the old school SaaS company. The SaaS company is trying to shove AI and use it as much as they can, but they're just fundamentally not constructed properly to do it. And so, in the fullness of time, those companies go away. The AI native companies take their place, and that's how AI takes over the world. I think there's a very decent chance that that's how it happens. The problem is it takes a long time. And timing is the issue. All the money that's being spent on CapEx is being spent now. And that money needs to circle back into because when you go from paying for free cash flows, paying with debt, paying with equity, the idea is it circles back around to your cash flows. Right. Right. And so the question is, how long is that going to take? It's all about timing. I 100% believe in AI. Like, I think that is pretty…
Andrew Sharp
how to monetize it, the new businesses that it enables, the
Ben Thompson
interface of it, the business model around it. Also, while we're here, if this whole interview, them open AI is evil because they include advertising.
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
Early in the episode, Thompson is asked about an AI bubble. He gives his view that a painful pop is inevitable, while granting that this era differs from the dot-com era in several ways. He says he doesn't know when it will happen and points to financing, including trouble for the newer AI cloud companies, as the gate.
Sharp Tech with Ben Thompson · 2026-10-30
Permalink →
Listen →
Sharp Tech with Ben Thompson
Around 07:18 into the episode
Ben Thompson
I don't know. I mean, so there's no question. No, there is a question.
Ben Thompson
I think it's inevitable that we are right now, will be whatever, sort of in a bubble, and it's going to pop and it's going to be painful. And then we'll come back. That's just the inevitable nature of these things. I think there is an extent to which AI is different in lots of areas. The sort of mechanics of stuff piling in too quickly, there being a mismatch in timing between when it's absorbed or not. Like there's lots of differences that we've talked about between the dot-com era and this era. But that dynamic, I think, where dot-com had lots of good ideas, just didn't make sense for a while or didn't sort of get penetration for a while. I think that is almost inevitably going to be the case. Now, when does that happen? I don't know. I mean, you have all the, there's been a lot of issues with the Neo scalers, Neo, NeoClouds. Neo clouds, yeah. Right. Can they get financing? All of which goes back to, I think Trotectary was pretty on the ball with this. We've been talking about the concern about the credit mismatch, not like just, are we going to run out of money? This is when we had the discussion about railroads. We had, you know, we talked about Google. I know we're going to get more into some of the financing questions sort of later on, but I think it is likely there's going to be pain. But so financing is the
Andrew Sharp
gate and people will experience pain. So my experience reading your coverage of the S1 this week, though, was turns out Anthropic's actually like a rock-solid business.
Ben Thompson
So yes, you're totally right. Like the core business looks amazing. The subscription for sure is profitable. It's less profitable than the API, but the subscription is a great product. People are a lot like they've already paid for it, so they're going to use it. They're going to get used to your…
Andrew Sharp
clawed people. That's right. Exactly. So once
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
Thompson is arguing about the U.S. defense industrial base. Just before, he said the World War II arsenal of democracy worked because civilian car factories were repurposed for military production, and that today's military contractors serve only the government. This passage is his explanation of why that single-customer model pushes toward low-volume, high-cost weapons.
Sharp Tech with Ben Thompson · 2026-10-30
Permalink →
Listen →
Sharp Tech with Ben Thompson
Around 1:08:35 into the episode
Ben Thompson
…we don't have that capability. What do we switch to making all this military stuff? All our military contractors, because they only serve the U.S. government, they end up, number one, the U.S. government is stuck with these insane entities that all they exist to is basically the U.S. government.
Ben Thompson
But then also you tend towards this very low volume, very high-cost weapon systems because you're just serving one customer. It's like the subscription business. In the subscription business, you want to maximize your revenue per user because you don't have as many users. If you're selling to only the government, you want to maximize your revenue per item because you only have one customer. If you're serving the entire market, then you make it up in volume. And so the future of war is, in my estimation, clearly things like drones, missiles, unmanned things, which is a volume play. All of our defensive primes are not built around volume plays. The consumer market is how you get sufficient volume. Do we
Andrew Sharp
need to fix the regulatory regime to enable the flowering of a domestic drone industry for the sake of national security?
Andrew Sharp
If we're going six-dimensional chess. Yeah. Maybe we do. No,
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
Thompson is closing a segment on power. Just before, he said that Ukraine, relative to the U.S. government or even Starlink, does not have power. He then uses that framing to explain what his article on Anthropic's dispute with DOW was really about: who holds power, not legal technicalities or surveillance.
Sharp Tech with Ben Thompson · 2026-10-30
Permalink →
Listen →
Sharp Tech with Ben Thompson
Around 1:15:11 into the episode
Ben Thompson
to Ukarsh's overall point. We're talking about power.
Ben Thompson
Like, this is the whole thing. Who has power? Ukraine, at the end of the day, relative to the U.S. government or even relative to Starlink, does not have power. Once you grok, no pun intended, that we're talking about power dynamics, you get what I was going for with the Anthropic versus DOW article. It wasn't about legalism niceties. It wasn't about mass surveillance. It was about who, at the end of the day, has power. And if Anthropic keeps saying and insisting, perhaps accurately, we are accruing power to an extent that rivals nation states, the nation states are going to have a say in it.
Andrew Sharp
Absolutely. And if Anthropic's not going to jump or not going to say how high when the government says jump, then the government is going to investigate other options and ensure that its suppliers are not relying on Anthropic. It is going to continue to be an issue, though, because particularly as…
Ben Thompson
This is unbelievable. We're an hour and a half in. I wanted to. No,
Andrew Sharp
no, no. First of all, we had a little interruption about 10 minutes in. So we have an extra five to seven minutes here. The text message that you sent me 30 minutes before recording was you looked at the rundown and you said, why no, you can see everything segment with multiple question marks. So
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
Thompson says the anthropomorphizing of chatbots bothers him and contrasts how the two leading labs present their models. He calls it a fundamental divide, and argues that models should be treated, and controlled, as tools.
Sharp Tech with Ben Thompson · 2026-10-23
Permalink →
Listen →
Sharp Tech with Ben Thompson
Around 1:08:38 into the episode
Ben Thompson
…you right now. I wonder if that is an accident. And yeah, it's like a scene. And it does bug me because Quad is a it's not just that your experience with Quad is different than my experience with Quad. Literally every token is like a fresh experience, right? That's the nature of LMs is everything
Andrew Sharp
because it personifies the technology, the machine intelligence.
Ben Thompson
…they actually come back down? What makes us human is our more animalistic bits, our instincts, our desires, our drives, which are linked to our cerebellum, like our higher level intelligence. And that link, the AI is all on the top. It's all like the top part of our brain. It doesn't actually have the brainstem. Like that actually is the drive to do it. And there's an aspect of, it's actually ChatGPT was actually the greatest name because it's such a temptation and it's so easy for people to get caught up in feeling like they're talking to something real that isn't real. And it's actually responsible to do everything in your power to continue remind people that this isn't real. Anthropic does the exact opposite. They try to make it real. Real. They try to teach it like it's real. Their whole view of alignment is not that they can control Quad, it's that they teach Quad to be good. Yeah. So it will control itself. And I just, it's a fundamental divide.
Andrew Sharp
A corporate misalignment is what you're describing.
Ben Thompson
I think so. I think we should be treating these and seeking to control them as tools. And there's good news in this regard. It turns out LMs are more understandable than we thought. There's some really interesting paper about interpretability and just being able to analyze it and understand what…
Andrew Sharp
That is great news. It is great
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
Thompson is explaining why he thinks AI safety debates end up in "crazy places." He argues for treating models as tools to make more understandable and controllable, and says Anthropic goes the opposite way on each of those points.
Sharp Tech with Ben Thompson · 2026-10-23
Permalink →
Listen →
Sharp Tech with Ben Thompson
Around 1:10:00 into the episode
Ben Thompson
news. And it's so funny. Like so many safety people were so dismissive of this, of this paper immediately because it gives a real blow to their view. Sure.
Andrew Sharp
It undermines the narrative. And it turns out this paper is
Ben Thompson
…than my fist. This idea of there's something that exceeds some of our characteristics is not a strange concept. Right. It's just like we're, and part of this is, I think, Silicon Valley, the sort of person in Silicon Valley, is so, their sense of self-worth and value is so wrapped up in being smart. Everyone in Silicon Valley has always been the smart one. And like, that's how they, there's, and so this idea that something's smarter than me is so overwhelming. It's such a blow to the psyche that you have one side is like descending into nihilism. It's all over. And the other side is like, it's okay if it's God. God can be smarter than me. Alignment, You end up in like these crazy places. And I just, I just think it's wrong. I think we need to be pushing towards making these tools, making them controllable. And part of that is helping people think about them in this way. Okay. Anthropic on every single those points goes in the opposite direction.
Andrew Sharp
Well, a counter to this conversation and that. Well, the part of this conversation
Ben Thompson
is that Opus 5.5 is unbelievable. It's not just time for everything. And there's an aspect of Quad is so much better at agentic stuff in my experience. Yeah. Because it just, it can multitask better. It can keep track of lots of different things, all the stuff that's going on. Like ChatGPT gets so…
Andrew Sharp
I think that people should think about Anthropic Alignment, whether they are aligned with humanity's best interests in terms of how they view these tools and how they're framing these tools for the masses to understand. I will note, though, that their Claude branding seems a lot more coherent than…
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
After a week of PHP exploits and AI-assisted attacks, Thompson says security pressure will keep closing off the open web. Self-hosted sites and email servers are giving way to a few big, defended platforms.
Sharp Tech with Ben Thompson · 2026-10-23
Permalink →
Listen →
Sharp Tech with Ben Thompson
Around 47:53 into the episode
Ben Thompson
coming. It's going to get worse and worse and worse.
Ben Thompson
Until it feels like the end of the internet. Like the internet, the old internet of I'm going to throw up a WordPress site. I used to do it all the time. I've had a bunch of WordPress sites over the years. I'm going to run my own email server. That's dead. That's been dead for a long time. No email gets delivered if it's not sent by one of the big ones. This, this, the whole promise of the internet of being openness, a million nodes. You Can See and be connected to anyone else. BBS is hosted on someone's like draw, you know, server in their basement. We are fortunate. Everyone who's listening to this who knows anything of what I just said, you should consider yourself fortunate because it was a moment in time. It's a golden era. It will never return. It will never wait. It's impossible to return.
Andrew Sharp
What do you mean by that? Like, what would a new version look like? There
Ben Thompson
will be no new version. We can't patch our way around. Absolutely not. Everything is so the one of my favorite. Oh, what's the
Andrew Sharp
honestly? I won't miss it too much at this point. The internet has been kind of a cesspool for a while, so it's fine.
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
Thompson is talking about tech's habit of mistaking more data for certainty. He calls LLMs consensus mechanisms that sample from the middle of a probability distribution, and says they avoid the edges because the edge that holds brilliance also holds the crazy.
Sharp Tech with Ben Thompson · 2026-09-18
Permalink →
Listen →
Sharp Tech with Ben Thompson
Around 20:23 into the episode
Ben Thompson
But there's a reason those rockets never ran up. By the way, all things considered, they are underrated. We're a great team. They ran up against one of the greatest teams of all time. And this is why it was so devastating. That's a round one. That was the
Andrew Sharp
ultimate problem for Daryl, but it is what it is.
Ben Thompson
…that arises from more data. Just because you have more data doesn't mean you have all the data. And tech has always been susceptible to falling in love with data. Data definitionally is backwards looking. Like the LLMs want me to cite something when I'm saying something that's happening right now or prediction about the future because that's how they work. They are consensus mechanisms. Like LLMs are literally the midwit of the center of the bell curve. Like that's what they are. It is a probability distribution. They're going to pull from the center of the distribution. They're not going to pull from the edges. And why? If you pull from the edges, you might pull from the crazy side as much as you pull from the brilliant side.
Andrew Sharp
You can get things wrong and you're sort of engineered to be accurate.
Ben Thompson
So I said in that article, like this is clearly like half of the political divide. And it didn't like that because it's like, you're going to make people upset. Like, but look, that's literally what's happening. And the reason it's happening is because this is unfalsifiable. It's something that…
Andrew Sharp
Not only that, it's going to be a reality that informs how policy is crafted here. And I do have some insight because I'm in Washington, D.C. and have a bunch of friends who are in politics. And I remember, I think it was like nine months ago, I met with somebody who is consulting with a potential…
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
Ranking the frontier AI labs by their setups, Ben Thompson argues OpenAI and Anthropic have the biggest upside because of how completely their teams believe in the mission, comparing the two companies to different strands of religious devotion.
Invest Like the Best (Patrick O'Shaughnessy) · 2026-08-18
Permalink →
Listen →
Invest Like the Best (Patrick O'Shaughnessy)
Around 56:35 into the episode
Ben Thompson
…in terms of your supply chain and making hard decisions is very, very different than everything that goes into like making great AI. I'm generally prefer companies to do what they're good at. So from my perspective, I'm fine with Apple not doing AI. I want them to keep making great devices. Of
Patrick O'Shaughnessy
the five potential frontier AI winners, so OpenAI Anthropic, Gemini, SpaceX AI, Grok, and Meta, which of those firms do you think has the most interesting setup?
Ben Thompson
Open AI and Anthropic obviously are the riskiest, but also have the biggest upside. Never discount, number one, the power of belief. They think they're creating God. The most impactful things in history have usually been fueled by religion. The two religious organizations in Silicon Valley are OpenAI is kind of like mainline. They go to church every Sunday. They're sort of like evangelicals. That's Anthropic. They're all in. It is core to their belief. That goes a long way. The fact you need to make a business work for you to survive goes a very long way. Google just needs search to not die too quickly. Meta has the huge advertising business. In a world where Meta was run by anyone other than Mark Zuckerberg, they would not be on the leading edge. That is one of the purest manifestations of founder…
Patrick O'Shaughnessy
You've written a lot about this recently.
Ben Thompson
I think there's a very good case to make that it is more reckless to not be on the frontier if you're a digital company. The counter to meta is actually Microsoft. Microsoft is not on the frontier. The reason why Microsoft has $20 million of free cash flow last quarter, Microsoft paid a $10 billion…
Patrick O'Shaughnessy
did IBM do? What's the analogy? Well,
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
Ben Thompson is explaining why no one has cracked a business model for the low-cost, casual side of AI chatbot usage. He argues Silicon Valley keeps forgetting a basic truth about consumer behavior that free-to-use apps like ChatGPT are running into again.
Invest Like the Best (Patrick O'Shaughnessy) · 2026-08-18
Permalink →
Listen →
Invest Like the Best (Patrick O'Shaughnessy)
Around 33:38 into the episode
Ben Thompson
…but they want to hold on to this set cost for the vast majority of employees who can fit in that because they need to ask their customers to think a little bit for those extreme employees, but they don't want them to think too much because that breaks the model in very surprising ways. Are you
Patrick O'Shaughnessy
surprised at all that the recipe builder user that is very low cost to serve, that there hasn't been a great business model that's emerged around them just yet? Google and Facebook are sort of business perfected in this prior era. They haven't seemed to figure this out at all.
Ben Thompson
I am frustrated, but not surprised. This is obviously a market that should be supported by advertising. That is why advertising is always the consumer business model. Consumers don't want to pay. There's two things to understand about consumers that Silicon Valley has to relearn about every 10 years. Number one, consumers do not want to pay for software. And number two, consumers do not care about being productive. We went through this in early SaaS. The canonical company for this in my mind is Dropbox. So Dropbox, unbelievable product, like especially when it first came out in business school, I was one of the first people to use Dropbox. And that went off like crazy. I have so much storage still, like my free Dropbox because I gave out my code to like so many people. So Drew Houston makes this amazing…
Patrick O'Shaughnessy
best engineers don't want to go work on this problem.
Ben Thompson
And so you literally had OpenAI replaying the Dropbox story, but at like 100 exercise, being like, no, we're going to sell subscriptions to consumers. They did. They sold a lot, but they didn't sell enough. If you're going to be in the consumer market, you have to be doing advertising. They're…
Patrick O'Shaughnessy
goes up as the bottom line.
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.