Welcome everyone to The Information's TITV. My name is Rocket Drew. I'm in for Akos Pasicha. It is Tuesday, November 10th. Before we get into the show, Anthropic just set a lower-than-expected tentative valuation for its IPO between $1.26 to $1.4 trillion. According to an updated SEC filing just last month, Anthropic was expected to go public at a valuation of $1.5 trillion. They could start trading as early as November 19th, offering 825 million shares. We will keep you updated as the hotly anticipated debut nears. First up on the show today, The Information has exclusive reporting about how Nvidia CEO Jensen Huang's ties to a woman that was reportedly identified by the US government as a Chinese intelligence asset helped facilitate at least $7.5 billion in Nvidia chip shipments through Chinese companies. We'll also get into the risks to Nvidia, its investors, and Jensen himself with a tech analyst. We'll then shift gears to some new reporting about Google's search advertising business. We'll also talk about a new robotic startup creating autonomous catheters for surgery. And we'll wrap up with a great profile from our health and science reporter about the famous futurist Ray Kurzweil. It's going to be a great show. So, let's get right on into it.
New exclusive reporting from The Information reveals how Nvidia CEO Jensen Huang's relationship with Alice Huang, a woman reportedly identified by the US government as a Chinese intelligence asset, helped facilitate at least $7.5 billion in Nvidia chip shipments through Chinese companies. Jing Yang, our Asia bureau chief, joins me now to share what we learned. Welcome on the show, Jing.
Welcome on. Can you walk us through some of the biggest findings from the story?
Yeah, sure. So, one of the biggest findings is that the two Huangs, you know, Jensen and Alice, have known each other for at least several years. We, previous to our story, we saw social media was circulating footage and video of the two of them hanging out together once in 2024 and once in 2005, both in Taipei and in a large group setting, and we revealed that actually their relationship, they've known each other for much longer than that.
Do we have a sense of how they might have gotten to know each other?
Well, I was going to say, and the number two was that Alice Huang, by the way, I mean, I'm sure our readers know they are not related, they just happen to have the same surname. And then Alice Huang is behind a group of companies, and in the center of that group of companies is this company in Shanghai called Central Leon. And Central Leon has facilitated the imports of Nvidia chips worth at least $7.5 billion since 2023. And that whole network was built with, you know, by Alice Huang, her sister, and her husband.
So how did that network work? What was the role of that network in bringing chips into China?
Yes. So this company on paper looks very young. It was set up in early 2023, and then it mostly, you know, is in a business of procuring AI chips, and then also, you know, and then put some of the chips in data center spaces it has, as well as selling it to other customers. It looks just like a, you know, a very mundane, you know, kind of company that we have seen spring up since, you know, the AI boom.
So did Alice's relationship with Jensen start in this way? Do we know that they met through this business connection, or this business connection emerged later?
So what we were able to trace back is that Alice's husband, another Chinese guy named Kaneway, he, at least in the 2010s, he was a private equity investor in Shanghai. And in that period, the two of them had done a series of acquisitions of gaming companies in China. And if you will recall, back then, you know, Nvidia is not a very well-known company, and most you would have known that their GPUs were mostly serving gaming companies. And then this is how the two of them got to know each other.
Wow. Yeah. Back in the stone ages when Nvidia was still a gaming company pre-AI. So one of those companies in the network, Central Leon, it became a preferred Nvidia cloud partner. What does that mean?
It just means that you get preferential access to Nvidia chips. Again, this is all what's really interesting, that Central was set up in, as I said earlier, early 2023, and then just a few months later the company earned the credential to become an Nvidia preferred cloud partner. That is a status given to companies that can show, number one, that it has gone through enough training, it knows how to handle and run Nvidia chips. Number two, a company has to reach revenue at a certain threshold. And we don't know how that company got that status in such a short amount of time since the funding. But what we know is that that status has made the company, you know, essentially having a hotline to Nvidia, you know, if you may. At that time in '23, right, AI chips were running very hot, and then there was a lot of demand for it. And then it was also the time that the US expanded its export control on China, basically that the US government widened the scope for export control that essentially also barred the imports of some Nvidia chips that the company developed specifically for the Chinese market. However, in that short period of time, if there are chips that were already sitting in like a distributor's warehouse, then an Nvidia preferred cloud partner would be able to get allocation to it, and it still managed to sell them.
Apparently even over the export controls. So, you mentioned that some of these $7.5 billion of chips were supposedly banned. They were not supposed to be sold in China. Do we have a sense of how many or what kind of chips those were? Any more information on that?
Yes. So we do. According to our reporting, there was roughly about 240,000 Nvidia chips that were procured by Central Leon since 2023, and most of those chips were the A and H generations of Nvidia chips.
So the last question I have for you then is what did Nvidia know about what was going on here? I mean, there's a question of what Jensen knew, and then there's a question of what the broader company understood about where these chips were ending up.
Yeah, I wish I could get into Jensen's head just as much as you. We honestly, we do not know how much Jensen knew what Central and Alice was doing. We also don't know how much the broader Nvidia knows. What we can pinpoint from the story is just the two of them obviously have had a relationship that goes back longer than what we previously know. And I also want to stress that by the time that Alice set up her company in AI chips in 2023, she had no prior experiences working in tech or AI or anything related to that. But at that time she had already been boasting to her business associates and customers of her relationship with Jensen as a way why, you know, even though I'm a new company in the industry, you should buy chips from me. And she also, she does not hide the fact that she's close to Jensen. She flaunts how close they are by telling people, showing them photos of them hanging out together, as well as telling people that Jensen celebrates her birthday every year by either having dinner with her or sending her gifts.
Wow. Well, it's a fascinating look at both their relationship and the workings of the chip industry. So, thank you, Jing, for coming on and breaking it down for us.
Nvidia shares continued to fall Tuesday morning amid the news of an investigation from Nvidia's board probing whether CEO Jensen Huang may have been targeted by a foreign intelligence operation. Joining us to discuss the potential risks for Nvidia and its investors is Gil Luria, managing director and head of technology research at D.A. Davidson. Welcome on the show, Gil.
Well, it's been a wild couple of days. Maybe to start, you can just give us your broad thoughts on this story so far, because there's a lot that we could get into here.
Yeah, it introduces a lot of risks to Nvidia. Jensen Huang is who built Nvidia and to a large extent been the key orchestrator and facilitator for the whole AI ecosystem. That doesn't mean we need Jensen Huang going forward, but he's had a very major role so far, and this scandal creates a risk that he won't be able to continue. There's also a bigger risk that this could impact Nvidia's business depending on what these investigations find about selling into China, selling to companies that sell into China, selling compute to China, further restrictions that could happen. So the bigger risk is to the business, but there's also a risk that Mr. Huang won't be able to continue as CEO.
Right. As an individual, like you said, he's a key orchestrator of the ecosystem, kingmaker even, which I suppose is what makes him so important as a potential target for espionage operations. You want to be on his good side. What did you make of Nvidia's stock performance over the last day or so?
Yeah, I mean, it's lost more than half a trillion dollars of value, which I actually think is appropriate. The stock was very attractively valued before, but this is a very negative development. So it makes sense that the market will now assume that there's a couple of scenarios that are very bad. The baseline scenario is that Nvidia did nothing wrong. Jensen Huang continues to serve as CEO, and they'll be fine. And when I say fine, I mean they're going to grow very fast and do very well. But we've now introduced a couple pretty negative scenarios that the stock has to account for. And I think that at this point, after the last couple of days of trading, that's pretty much in the stock.
What are those negative scenarios you see?
Yeah. So again, it's that the investigation concludes that Nvidia had wrongdoing around China. That's one scenario. In a related scenario is if Jensen Huang has to leave.
Got it. So there could be additional regulatory or national security scrutiny, and in addition to or separate from that, there could be a scenario where Jensen himself is out.
That's right. And it's not just scrutiny. I'm more concerned the scrutiny is going to definitely happen. There'll be congressional testimony, investigations, especially with the Democratic Congress. I'm more worried about what they're going to find. There's a possibility that they do find that Nvidia improperly sold it to companies they knew were going to sell to Chinese companies, that they sold to companies they knew would serve compute to Chinese companies, and that may or may not be inside the lines and inside the regulation they were supposed to comply with. That's the bigger concern. The scrutiny is going to happen anyway.
What kind of outcomes would you see from a finding like that? We're talking about a big government fine or a serious change to Nvidia's stature in the industry.
Yeah, they'd be able to sell less. So maybe some of those Singapore connections that they have, some of those resellers in Singapore and Taiwan, maybe they can't sell to those companies at all. That could be a significant limit to their business. There's probably still enough demand outside of that, but that has been a meaningful part of their business over the last few years, selling into Asian, non-Chinese Asian countries. And the concern is that from those Asian countries, the GPUs or the compute was serving Chinese companies inside or outside of China. So that would be the biggest risk there is that they wouldn't be able to sell in broad regions of Asia due to these concerns about serving that compute to China.
Well, that's a pretty serious risk. I would think that something that significant explains most of the drop in the stock price that we've seen over the last couple of days. But what about Jensen himself? I mean, like you said, he is so synonymous with Nvidia and Nvidia's success. If Jensen is pushed out, but there is not a negative finding on the regulatory front, what does that kind of outcome look like for Nvidia?
I'd argue that's not as bad because Nvidia is so well situated at this point, and Jensen has built the most important company in AI, really, and so it's a very resilient business. And now that there is scrutiny, now that there's a board investigation, the possibility of Mr. Huang leaving, this is the time to think about succession and do so quickly. So internally, who are the candidates? Start vetting them, start figuring out if those are the right people, or maybe externally, maybe there are people in other companies that would be able to handle such a big role and move Nvidia forward even without Mr. Huang. This is the time the board needs to do that right away to have options. So if they see that they're about to approach an outcome where Mr. Huang is no longer at Nvidia, they need to have immediate ability to act upon that and have somebody step in, not just as interim CEO, but as the CEO that carries this forward. It's important for Nvidia's stability. It's important for the whole ecosystem.
Is there anyone either internally or externally that you think they are considering or should consider?
Let me just throw one name out there. I don't know who they are considering, but I'd say Lip-Bu Tan probably would be my choice. This is a really good CEO at a really bad company. There's only so much a good CEO can do with Intel. Intel is so broken six ways to Sunday. And so he's done a lot. He's turned it around. He's gotten the stock up significantly even after the big correction. But there's only so much he can do. He understands the ecosystem very, very well. He understands all the players. He has good connections in Asia, so there'd be continuity there. He could be an interesting candidate for this job.
Well, that would be a fascinating development. If you hear that moving forward, give me a call. I would love to know. So, more broadly, do you have any other thoughts on how this will affect Nvidia's relationships with its major customers, with hyperscalers, with the US government?
It should. Even if there are findings about selling into China, then we just need to stop doing that. Nvidia is a remarkable American company that has fantastic relationships with all of its customers. You don't hear, I mean, Microsoft, Amazon, and Google, they complain about how much they have to pay for the GPUs, but they're also very grateful to Nvidia and have been relying on Nvidia in growing their business, which in turn has served Anthropic and OpenAI. It's made TSMC's business bigger, ASML's business bigger, the whole semiconductor supply chain. Nvidia has been, again, it has led the AI ecosystem. It's very much respected and appreciated by its customers and suppliers, and that should continue, and that should transcend Jensen Huang. But again, there would need to be a clean handoff for there not to be a disruption here.
Great. Well, thanks very much for coming on the show, Gil, and giving us your perspective on this.
Google's search advertising business is showing signs of slowing down. Ad revenue growth fell from 17% in the second quarter to 12% in the third. The company partly attributes this to making less money from ads in AI Mode, its ChatGPT-like search tool. Our digital advertising reporter, Alix Coutures, has been looking into how AI is disrupting Google's ad business. She joins us now to discuss. Welcome on the show, Alix.
Hi, Rocket. Great to be here.
Great to have you. So, your story opens with an anecdote from the CEO of this business in LA. He noticed something was amiss with his ad budget. What happened?
So, yeah, it's a sales company, and basically he allocated a budget to Google as he always does, and a few months ago he started noticing that the money was going unspent, and so the money kept returning to him, which means that Google wasn't using that budget. And so that's because people were not clicking on his ads and not seeing his ads, and something was wrong with Google's advertising business because he wanted to spend on Google, but the money was just unspent. So that's when he started thinking, okay, what am I going to do with that money? And not necessarily his, but some advertisers have been reallocating the budget to other platforms such as Meta.
Wow. He's literally like, take my money, and there's just, there aren't ads to make it happen. Can you explain to us how Google is structuring its search business right now? Because there's still traditional search, but now there's AI Mode, but that's different from AI summaries.
Yes. So basically this traditional search where you search keywords and then you get a result, and then in '24 and '25 it started rolling out AI Overviews and then AI Mode. So AI Overviews is basically a summary at the top of the page of your query. So whatever you ask, Google is going to summarize it. That's AI Overviews. It's going to use AI to do that. And then AI Mode is powered by Gemini, which we all know, and it's basically a follow-up question that you have, and you can have a conversation with Google in AI Mode just like you would have in Gemini or on ChatGPT. So those are the two AI-powered features. And so at the beginning, Google did not roll out any ads in AI Overviews and AI Mode. It did it progressively, and it did it a few months after launching AI Overviews and AI Mode, all of that to avoid disrupting users' experience.
So there are ads in AI Overviews then at this point. So why should the ad revenue be any different? Why are these revenue streams different?
Well, it's very different because the way people search in AI Overviews and AI Mode is just very different. So it's difficult for Google to basically figure out where they're going to show these ads because before it was pretty easy. You would ask, okay, I'm going on a camping trip to Utah. I want a tent. So you would search tent, and then you would get a tent, and you would buy it. That was easy. Now you're going to ask a richer and deeper question. So now you're going to ask, "How should I prepare my trip to Utah?" And then you don't necessarily want to buy immediately when you do that. You just want some tips on how to prepare your trip. So then where is Google going to show the ads? They're going to say, "Oh, you would need probably a tent, and you would need a pillow and a sleeping bag." And so it's more difficult for Google to figure out where to actually show these ads because the queries are just very different. And also when I actually make that search, I don't necessarily want to buy. So I'm going to see those ads, but am I going to click? Not necessarily. But it's still going to have an impact on what I'm going to do down the line. So that's the question of how to monetize it because it's going to have an impact on my decisions down the line, and I might buy down the line, but just not now. So how do you charge for that? Where do you show it? So that's a whole thing, a big question that Google is trying to answer right now, and Google is not the only one, obviously, trying to answer that question. A lot of other chatbots are actually struggling with the same issue, such as OpenAI.
Right, right. You make a good point. They have a lot of data. Presumably, they'll find some way to leverage that into better serving ads. But yeah, like when I want to do a deep shopping dive, I actually tend to go to ChatGPT over Google. And when I know exactly what I want to buy already, I just go right to Amazon. So there's not a lot of shopping that I'm left doing on Google anymore these days. So I can see how they wind up in that position.
Yeah, yeah. Google is really squeezed between now Amazon and OpenAI. OpenAI for the deep research, and then Amazon for the products.
So you say that Google is still in the early stages of this experiment with AI search, and so they're trying not to alienate consumers by overloading them with ads in AI Mode. How long do you think that lasts? Are they going to ramp up the advertising that we see?
Yes. So they've started actually ramping up already. So what they did on AI Mode and AI Overviews, which was not showing too many ads and doing it slowly, it was very intentional. And people tend to forget that. People say it was a panic attack. Look, they're in panic mode. They're not really in panic mode, actually. I think it was very intentional. They just don't want to lose users because that would be a real disaster for them and a real hit. So they want to do it slowly. And so now they're starting to ramp up. To ramp up on formats, to ramp up on the number of ads they're showing. So everything is scaling up, and I think they're going to make revenue. Well, more revenue. I think the growth is going to pick up pretty soon in a few quarters. Maybe next quarter, the following one. At one point it's going to pick up. And it happened to Meta also a few years ago when Meta started showing Facebook Reels on Facebook. Its advertising business took a hit also because it's just new formats, and you need to do it slowly, and then it picked up a few years later or a few months later.
Yeah, OpenAI similarly ramped up advertising in ChatGPT pretty slowly, and there are other chatbots that still don't advertise. How is Google thinking about acquisitions related to its ad business? Is that something you see on the horizon?
Yes. So I did hear that Google was actually hunting for creative startups. So Google is always going to be looking at very mature companies. Usually, that's what they do when they want to acquire a startup. So creative startups would help them actually innovate in terms of like ad formats, and that would help them improve ad quality on search, and that would be a good thing for them because all they want to do is keep the advertisers on their platform, and they want to do retention. So having more creative ads that would be more aligned with AI or AI formats would make sense for them.
So we've talked a lot about Google's position here. I just want to come back to the point of view of the advertisers. We started with this person with this LA business who can't spend their money on advertising anymore. How are advertisers changing their strategy now?
Well, now advertisers are trying to basically shift from clicks to impressions. So that's a bit of, you know, jargon. Basically, before they were thinking about how many people were going to click on their ads, and now that the way people search is changing, they're thinking about how many people are going to see their ads, and that's what you call brand awareness. It's basically when you're going to see the ad and you're going to think about the ads. So you do brand awareness instead of clicking with a commercial intent. So advertisers are trying to change their strategy based on that. And so what they do is that they give much more context to their advertising prompts. And when they write the ad, they think about, they give a lot more information about their company, about the websites they want the users to land on when they click. So they're just rethinking their strategy based on that. And I think Google is helping them do that by releasing a lot of tools, creative tools, and yeah, reporting tools, things that are going to help them actually transition.
Right. Tools based on AI. So they're saying there's a silver lining there, we'll help facilitate that transition for you.
Great. Well, thanks very much, Alix, for coming on and breaking it down for us.
Robots have been helping surgeons perform operations for a long time. But a new generation of robotics companies is taking advantage of all the latest AI technology to make surgical robots more autonomous than ever. One such company is Mosaic Intelligence Labs, which raised $47 million for the semi-autonomous catheters it's developing. Co-founder and CEO Ileana Pirozzi joins us to tell us about it. All right, welcome on the show, Ileana.
Hi, Rocket. Thanks for having me.
