transcribe

The Information | TITV | November 18, 2026

The Information · 41m · transcribed 2d ago
More from The Information Business
𝕏 Share ▶ YouTube 📥 PDF 🤖 .md

Section Insights

# 0:00

Introduction and Upcoming Events

What is the focus of today's show and what events are upcoming?

The show will cover Nvidia's quarterly results, exclusive reporting on Nvidia's former sales chief, and Google's TPU software boom. There is also an upcoming live video summit on Anthropic's IPO.

  • The show will discuss Nvidia's performance without its CEO.
  • Exclusive insights on Nvidia's former sales chief will be shared.
  • The Anthropic IPO summit is scheduled for tomorrow.
# 8:10

Nvidia's Future Post-Jensen

How will Nvidia's relationship with Washington change after Jensen Huang's departure?

Nvidia's relationship with Washington is expected to evolve due to ongoing legislative efforts aimed at tightening controls on chip sales, particularly to China.

  • Legislation like the Match Act and AI Overwatch Act will impact Nvidia's operations.
  • Tighter controls on chip sales to China are being pushed by Congress.
  • The Jensen Huang ouster may accelerate these legislative changes.
# 16:21

Concerns Surrounding Supermicro

What are the concerns regarding Supermicro's relationship with Nvidia?

Supermicro, a key partner of Nvidia, has faced scrutiny due to past executive actions involving the illegal diversion of Nvidia chips to China, raising questions about preferential treatment.

  • Supermicro's past actions have led to scrutiny and concerns about compliance.
  • The relationship between Nvidia and Supermicro is under examination.
  • There are ongoing questions about the integrity of chip distribution.
# 24:32

Open Source and TPU Market Impact

How does the open-source nature of TPU software affect its market adoption?

The open-source TPU software aims to simplify usage for developers, allowing them to integrate more easily without relying on Google, which is expected to enhance market traction.

  • Open-source TPU software is designed for easier adoption by developers.
  • This approach encourages broader participation in the TPU ecosystem.
  • The shift towards self-service tools is significant in the AI landscape.
# 32:42

Greg's Return to OpenAI

What are the implications of Greg's return to OpenAI?

Greg's return is pivotal for OpenAI as the company prepares for a potential IPO, and he aims to take on a more significant management role to address the challenges faced by the organization.

  • Greg's leadership is crucial as OpenAI approaches an IPO.
  • The company is currently behind competitors like Anthropic.
  • His management style is expected to influence OpenAI's strategic direction.

Transcript

Speaker 1

0:58 Foreign. Welcome everyone to the information TI TV. My name is Akash Bhasricha. It is Wednesday, November 18th. A quick reminder before we get started, we have our Big Anthropic IPO Live Video Summit tomorrow starting at 11:30am Pacific, 2:30pm Eastern.

Speaker 1

1:30 It will be available to subscribers of the information. We of course are expecting to see shares get priced tonight. Tomorrow at the summit we will be talking through the market reaction in real time, how it will affect the IPO market and even Anthropic's playbook for drug discovery. If you are a subscriber, you can register@theinformation.com events. If you're not yet a subscriber, check out our website for more details. Today on the show, we are going to break down Nvidia's first quarterly results without Jensen Huang at the helm.

Speaker 1

2:03 We are then going to look at some exclusive reporting we have about Nvidia's former sales chief Jay Puri. We are then going to get to Google's TPU software boom with our Google reporter. And to close out the show, we are going to bring on our OpenAI reporter to discuss her profile and her interview with Greg Brockman. It's going to be a great show. So let's get right on into it. Nvidia reported its first quarterly results last night without Jensen Huang.

Speaker 1

2:31 The numbers were as expected, still pretty strong. Revenue doubled from the prior quarter. Company raised its top line forecast for next year to 75%. But there were few details still about the status of the board's investigation behind the scenes. Joining me now to break down the results and the call is Chris Miller, author of Chip Wars. Chris, welcome back to the show. It's great to have you here.

Speaker 2

2:56 Thanks for having me.

Speaker 1

2:57 Okay, so Chris, if you are an Nvidia shareholder, did last night's call do what you needed it to do?

Speaker 2

3:07 Well, I think it wasn't a surprise that Nvidia reported very strong results, but those are all backward looking. Tell us what happened last quarter. The question is what happens the next quarter and the quarter after that. It's no doubt going to take the board some time to figure out the leadership succession. But that is issue number one for the company. And until that's resolved, the second question of can Nvidia keep playing its current role in financing so much of the AI ecosystem will also remain unresolved.

Speaker 2

3:37 And so I think leadership is going to be the key issue going forward.

Speaker 1

3:41 Yeah. So let's talk about some of the specific things that analysts drilled down on last night. So gross Margins were kind of an interesting story here. We heard last quarter, Colette Kress, stay on the call that, hey, we're expecting gross margins to dip a bit in the second half of the year, which is what we saw in this quarter. Q4, they're going to dip a little bit more, I think, and then Q1 of fiscal 28 is when they're going to start coming back up again a little bit.

Speaker 1

4:11 And look, we know that we are in this memory crunch right now, and so that's part of it. Is that the whole story here or is there also a pricing factor here for the chips?

Speaker 2

4:25 Yeah, I think there is a pricing factor as well. And we've seen just a lot more competition, especially in the inference market for Nvidia over the last year or so, and even more competition expected over the next couple of years. It's all the Asics that the hyperscalers and now the AI labs are doing in house, but it's also a variety of new startups that are going to be coming to market over the next couple of years, which won't win major share immediately, but just lead to a more diverse market rather than one in which Nvidia is guaranteed to be the dominant player.

Speaker 2

5:00 And insofar as that's the case, that'll put pressure on Nvidia's pricing.

Speaker 1

5:04 So how big of a problem do you think this is for the company? And I ask this because you've studied these types of chip paradigm shifts through the decades. I mean, is this going to be a major issue for them?

Speaker 2

5:21 Yeah, I think you ask what's Nvidia's moat? You've got the Cuda moat, which is strongest in training, rather than inference. You've got this supply chain moat that they can lock down memory, lock down substrates, lock down packaging capacity to a greater degree than their competitors. Those moats are real. But in the inference market, there are other big players as well. If you're a hyperscaler, you also have the same balance sheet ability to lock down the supply chain if you want to.

Speaker 2

5:49 Now, many of them haven't been as aggressive as Nvidia, but they're starting to learn their lesson. If you want capacity in 2029, you got to put money on the table today to start securing that. And so that's something that Google can do. That's something that all the other hyperscalers can do and are starting to do to a greater degree. And insofar as they do that, that means Nvidia's supply chain moat is less of a Moat. And that's why you're seeing more competition in the space for inference chips.

Speaker 1

6:16 Right? Right. Now, the other question that was really interesting on the call was from Stacey Razgin. He asked about this statement in the CFO commentary on our commitments, talking about the commitments to the NEO cloud companies, the vendor financing commitments. Our commitments take effect only if the NEO cloud secure financing to purchase the data center infrastructure products. And I mean, I wondered if you could translate this for us. Is this basically Colette saying, hey, we know NEO clouds have gotten a lot of attention recently.

Speaker 1

6:50 We want you to know that the vendor financing is. It's not to build the facility, it's strictly for the chip purchases. Is that basically what they were trying to say?

Speaker 2

7:00 Yeah, I think that's right. And if so, it means Nvidia's got a more limited financial exposure than you might have thought, that that's a good thing in some ways. But it also means that Nvidia's backstop for the new clouds is, is. Is less significant than you might have thought as well. And so there's a push pull here. Nvidia shareholders don't want to be on the hook, but the rest of the ecosystem wants Nvidia's balance sheet to be deployed as broadly as possible.

Speaker 2

7:27 And I think this is going to be a area of immense focus. The next couple of months is under the new leadership. Will Nvidia pull back using its balance sheet to support the NEO Clouds, to support other newer companies that need the balance sheet support that Jensen was so active in providing.

Speaker 1

7:44 Do you think they will pull back?

Speaker 2

7:47 I think they will. I think any interim CEO is going to be more cautious, especially

Speaker 1

7:51 when they're the CFO of the company.

Speaker 2

7:54 That's exactly. That's exactly right. And I think the general questions that have been asked around the NEO clouds, the declines in the share price of Core Weave and Nubius, these are all reasons to be a bit more cautious in extending financing guarantees to these companies.

Speaker 1

8:10 Right. So let's talk about Nvidia in the post Jensen era, because you wrote this piece for the FT talking about all the different vectors on which this story could change. And you also wrote specifically about how Nvidia's relationship with Washington will change going forward. One of the parts that was interesting about your piece is you. You outlined the legislation that is trying to move its way through Congress to do with chips even before the Jensen ouster, and you outlined the Match act, the AI Overwatch act, and the Chip Security Act.

Speaker 1

8:43 Can you give us a flavor each of these Three pieces of legislation. What exactly are they trying to do and how does it ultimately relate to what Nvidia will be able to do?

Speaker 2

8:52 Yeah, I think this is a key part of the calculus. Even before the Jensen Gate scandal, Congress was pushing for tighter controls on chips and chip making equipment. The Match act would intensify restrictions on chip equipment sales to China. The AI Overwatch act would require congressional approval to if a president wanted to allow GPU sales to China. So it would limit Trump's discretion. And then third, the Chip Security act would mandate location tracking on GPUs to verify that they weren't being smuggled into China.

Speaker 2

9:25 Congress has wanted to move this legislation forward for a while. The scandal is certainly going to facilitate that. And the question right now is how does this interact with the broader China US relationship? Trump and Xi meeting today in Shenzhen. The China US truce scheduled to expire in January, which would lead to new Chinese restrictions on rare earths and new US restrictions on chip making equipment. Unless there's an extension to be done, this is high stakes diplomacy right now.

Speaker 2

9:58 And AI supply chains are front and center in both how China and the US are trying to build leverage against each other. And Nvidia GPUs are perhaps the most prized commodity of all.

Speaker 1

10:09 The most prized commodity. But who else is in the crosshairs of this story? Because we saw Oracle had a chip deal with Tencent that fell through. And so I've been thinking about all the other companies amd. I mean, certainly maybe we put Google aside for a minute, but who else is in the crosshairs here?

Speaker 2

10:30 Well, if the Match act gets passed, that's bad news for asml, bad news for Tokyo Electronic. But actually for US firms, it doesn't really change much because they already weren't allowed to sell to China. So it's the non US equipment firms that get hit there. But if you look at the, the rules that now prohibit sending GPUs to Southeast Asia for renting onwards to Chinese customers, you know, that's certainly bad for Nvidia. It just reduces the number of customers that they have.

Speaker 2

11:00 It's good news though, probably for hyperscalers, certainly for US AI labs, there's now a bunch of data centers in Southeast Asia that as of tomorrow won't be able to serve their primary customers who are Chinese. And there'll be spare capacity that could be accessible to Neo clouds or to AI labs that swoop in and grab them. So there certainly has been a reshuffling of influence here and the moves of the last Couple of days, I think.

Speaker 2

11:24 Benefit the AI labs, benefit the hyperscalers, at the expense of chip makers and at the expense of Nvidia, above all.

Speaker 1

11:32 Right, Chris, given that you have studied so many of these chip companies and the history of these chip companies, the drama that we've seen in the last week or two, the market impact of it all, the market share impact that it could have, are there any historical parallels that come to mind in the chip sector or otherwise, given what you've studied?

Speaker 2

11:57 You know, it's hard to find good parallels. We've certainly had spy scandals before in the chip industry. Back in the Cold War, the Soviets and the East Germans sent spies to steal secrets, but those were secrets from small companies, not the world's largest company. And I don't think we've ever seen a spy scandal that also coincided with a customer of the world's largest chip company as well. So this is, I think, pretty unique, which is why it'll be fascinating to see how it plays out.

Speaker 1

12:24 Okay, let me ask you maybe one last question before you go. I think the consensus is that Nvidia may turn into a bit more of a boring company over the next year or two, given that the founder led personality of the business would go away here. Is there a world where being a more boring company could actually benefit Nvidia down the line?

Speaker 2

12:52 Well, I think Nvidia would have benefited had it been more boring starting six months ago or a year ago. I'd like to be more boring today, but I think there's a real risk for Nvidia that Jensen isn't easy to replace. And becoming

Speaker 1

13:05 a more hierarchical Jensen. What? Sorry.

Speaker 2

13:07 Is not easy to replace. The number of people that can step into his shoes is small. The role that he played in the ecosystem is huge. And Nvidia wishes it was more boring now. But there are costs to being boring as well.

Speaker 1

13:21 Great. Okay. Well, Chris, I want to thank you for coming on. That is Chris Miller, the author of Chip wars, here on TI tv. As the world re examines Jensen Huang's dramatic tenure at Nvidia. Jay Puri, the company's former top sales leader, is a name that has come up in a lot of reporting by my colleagues, Channel Liu and Phoebe Liu. I want to bring on Phoebe to share more about what they learned. Phoebe, welcome back to the show.

Speaker 1

13:47 It's great to have you here.

Speaker 3

13:48 Good to be here,

Speaker 1

13:50 Phoebe. Who is Jay Puri?

Speaker 3

13:52 Jay Puri is Nvidia's longtime head of sales At Nvidia, they called it executive vice president of Worldwide Field Operations, but he was the top sales guy. He was at Nvidia for I think 21 years, 17 of them. He reported directly to Jensen and was managing the whole company's sales operation. The people I talked to who had worked with him for a really long time kind of credit him for understanding how to interpret Jensen's directives and transitioning the way the company's sales operations worked from kind of oh like our product's tech is so cool.

Speaker 3

14:26 Like we'll offer you this discount, buy this much to like, if you buy our full thing at the full price, you will make more revenue. And that has really helped Nvidia get to the point where it is today.

Speaker 1

14:36 And when did he leave the company?

Speaker 3

14:39 So he announced his intention to retire in June, which was coincidentally when I was trying to work on a profile with him, and then officially stepped down for his successor, alongtime Microsoft guy Nick Parker, to take the role at the end of August.

Speaker 1

14:55 At the end of August. Okay, so it's been a couple months here. Yes.

Speaker 3

15:00 But anything that might come up in the investigation about potential chip diversion to China, Jay probably would have been in his role at the time.

Speaker 1

15:08 Right, right. So let's talk about them because the root of the story that you and Channer published really looks at Jay's working relationship with Alice Wong and with the customers that have caused so much concern for Nvidia. What was the story there?

Speaker 3

15:33 Yeah, so we found by talking to people who had worked with them that Jay knew Alice. They had met, I think in normal business conversations because like Jay's the head of sales, of course he would talk to customers, but I think still notable that he did know her and all the other customers. In addition to that, he is now married to a former high up sales executive at Supermicro as well.

Speaker 1

16:00 And Supermicro is what remind us

Speaker 3

16:02 Super Micro is one of Nvidia's biggest customers. So Nvidia usually sells its chips to equipment manufacturers like Dell or Supermicro. Those are I think the two biggest ones still. And they kind of assemble the chips into the full server racks that you see like lots of fancy pictures of. And then in turn those get sold to customers like a Microsoft or like an OpenAI, soon to be something like that. So Super Micro is one of Nvidia's most important customers and partners.

Speaker 3

16:31 And over the past couple of years they've come under some scrutiny for some executives who are no longer at the company, helping illegally divert Nvidia chips to China as well. To be clear, Jpuri's wife, Sen Li Chen, who also left Supermicro earlier this year, has not been implicated of anything like that. But it is still interesting, I think, that they're married.

Speaker 1

16:56 Okay. And so, I mean, give us sort of an overview here of some of the concerns here. Is it basically the same flavor of concern as Mega Speed, that Supermicro, what, got some. Got preferential treatment, or that? Like, what exactly is the concern here?

Speaker 3

17:19 Yeah, I mean, I think it's more raising questions and planting a flag. Because I think the big question here is, like, as Nvidia continues to come under scrutiny from government and other investigators for their role in facilitating or turning a blind eye to chip servers that end up in China, the question is, who, besides the one Nvidia employee who was arrested in Taiwan in August, knew? And there's no proof that, like, Jay or Senli or any of these people did know, but they would likely be someone that prosecutors or other investigators would want to interview.

Speaker 3

18:02 There were some people that were saying that basically Jay was the person at Nvidia, besides Jensen, who had the most power in deciding which customers got how many servers. And Stanley had quite a bit of power within Supermicro in deciding how those servers were allocated to Supermicro's customers. So one, maybe there is a little bit of a conflict of interest there. I'm not sure exactly when they got married, but I think it was around the time that Sendly stepped down.

Speaker 3

18:31 Otherwise it would have needed to be disclosed in SEC filings. But just given that relationship, and for context, Sendly reported directly to Supermicro's CEO, Charles Liang, and there were some legal complaints that were saying that she signed off on GPU orders, the serial numbers of which ended up in China, and that was the reason she stepped down. Obviously, Supermicro has denied that she's now CEO of a different publicly traded company in the space, but I think it just raises some questions.

Speaker 3

19:08 And I've heard that there's more reporting to be done along these lines, but don't have all the answers yet.

Speaker 1

19:15 So, I mean, I think the broader takeaway that I sort of was left with from your story is we've had so much attention on Jensen, and there has been a question around, well, who else in Jensen's orbit may have known anything about what was happening behind the scenes? And we focused so much on the board. Obviously, we focused on the government officials that. That knew anything about the situation. J.

Speaker 1

19:46 Puri's, you know, working relationships here are sort of another data point, and it sort of got me thinking around. Okay, so as the board does its investigation. You know, who else in the Nvidia executive circle become targets here in the investigation? And the reason I'm thinking about that, Phoebe, is because the question of succession, I mean, maybe this is actually the argument to bring somebody from the outside in. It was because maybe. I

Speaker 3

20:13 think so, yeah.

Speaker 1

20:14 The people inside Nvidia are like, maybe they all knew about it. I don't know.

Speaker 3

20:18 Right. And I mean, the two external candidates, the two candidates that have been reported to have been interviewed for the job are both external as well. Yeah, I think so. There are a few names of Nvidia executives that have come up that might know about this. And it's still unclear to me exactly what the scope of the board investigation is. But obviously Jensen and Tim Teter, the general counsel, are testifying to the government early December. I don't know.

Speaker 3

20:49 Colette's name has come up because she obviously, in light of the SEC also investigating Nvidia, oversees financial statements and things like that. But, yeah, not making any accusations here, but these are the people who, like, would have overseen a lot of the things that are now under scrutiny.

Speaker 1

21:06 Right. Okay. Well, Phoebe, some great reporting. I know there's more to come, and so I will leave you to it. That is Phoebe Liu, our Nvidia reporter here at the Information. Yesterday on the show, we talked about the traction Google's TPU has been getting in the market. Today we are double clicking on one important element of that story. Google's TPU software deserves a lot of credit for the newfound traction. My colleague Ryan Barwick, who covers Google, wrote about that.

Speaker 1

21:36 I want to bring him on to talk us through his reporting. Ryan, welcome to the show. It's great to have you here.

Speaker 4

21:41 Hey, man, what's up? How's it going? Excited to be here?

Speaker 1

21:43 I'm doing good. Okay, so let's talk. We have talked on the show extensively about Cuda and we just heard from Chris Miller about software being such a moat for Nvidia. There's another software out there called Pytorch. And, you know, I guess we haven't talked about it so much. So what is Pytorch? Is it a rival software?

Speaker 4

22:06 It's not exactly a rival software. It was developed by Meta in, I believe, 2016, 2017. And it is essentially this framework to put these chips into action. Cuda is specific to Nvidia, obviously, but Pytorch is used to, I guess, kind of speak to CUDA to get things rolling there. And it has been widely adopted throughout the industry.

Speaker 1

22:36 And so help me Understand this, you could conceivably be using, if you're a chip customer, both Cuda and Pytorch, because Pytorch has to connect to Cuda. Is that the idea?

Speaker 4

22:46 I believe so. I'm still relatively new to the beat, but as I understand it and as I've been kind of focusing on these tools that kind of get these chips integrated into obviously these data centers and with the companies that are using them. Yes, that's how I understand it.

Speaker 1

23:03 Okay, so now that's Pytorch. Now, your story centered on a different software called Torch tpu. So is this a new version of the software? Is it related? Is it, is it different? What is it?

Speaker 4

23:16 So it is a software layer. That is the terminology that Google and I went back and forth on using. Essentially, Google has these chips, these TPU chips, and it was traditionally built with jax, which is the kind of Google framework it wants its customers to use. The challenge with JAX is I actually source describe it as alien technology to me. So it's very complex to use and there wasn't necessarily.

Speaker 4

23:51 It was very clunky, so there wasn't really a lot of adoption. So converting your entire workflow to TPUs took a lot of work. Google has obviously been well aware of that. And they roll out Torch tpu. They rolled this out in late October. And the general idea is that it will take Pytorch, which is what the industry was using, and convert that into TPUs. So the idea is to make it as efficient as possible to use Google's chips.

Speaker 4

24:23 And already you're starting to see the traction, I think, in Google's third quarter earnings report. Thomas Kurian, who's leading Google Cloud, he specifically says, we think this software update is actually going to really open up the market to TPUs.

Speaker 1

24:39 And torch TPU is open source, right?

Speaker 4

24:41 Yes. Yeah.

Speaker 1

24:43 And is that a big reason why it's gotten so much traction?

Speaker 4

24:47 Well, I think the idea is they want to make it as easy as possible to use these chips instead of having to go back and forth with Google folks. People can just do it themselves. Of course, this is even more kind of within the context of AI getting really good at writing software and kind of converting these workflows. The fact that it's open source is just trying to get as many people on the boat as possible.

Speaker 1

25:13 Right. So we had a chip investor, Umesh Podval, on the show a couple weeks ago. He was talking about how vibe coding has really helped Google chip customers really better integrate into that ecosystem effectively. Does that resonate with what you were hearing?

Speaker 4

25:34 Yeah, absolutely. I mean that, that is certainly. I think this was reported late last week, but Anthropics own executives have told investors they were able to get, I think it was AMD chip, AMD GPUs. This was a migration that would have taken six to 12 months. They were able to do it in three days. So what you will see is it's going to become so much easier to change chips, to diversify your chip strategy. And Google obviously wants a piece of that.

Speaker 4

26:06 And this is all happening, I guess in the light of this Nvidia drama. Right. So these two kind of things are happening parallel chips are. Nvidia's facing a lot of pressure, but also chips are getting a lot easier to use and a lot easier to diversify and go back and forth between chips.

Speaker 1

26:26 And so just to. Again, I've got some clarification questions here, so bear with

Speaker 4

26:30 me.

Speaker 1

26:31 So we talked about how Pytorch was sort of not really a competitor to Cuda because you kind of have to use them at the same time. Torch tpu, that is a rival to Cuda.

Speaker 4

26:46 It's not so much, I guess it's a rival in the sense that Google as a chip maker is a rival to Nvidia.

Speaker 1

26:54 Yeah.

Speaker 4

26:54 And torch TPU is the language that will help people use TPU chips. So I guess in that sense it is a rival. But ultimately these are just frameworks to get these chips working efficiently.

Speaker 1

27:08 Right. Okay. And I mean, I guess higher level question here. Is Nvidia doing anything about this from, from what you're here or you know, what's been the reaction in the market?

Speaker 4

27:22 Yeah, it's funny, it's like, so when obviously we're chasing this TPU story and we're hearing, we're trying to get a sense of how many customers there are for these chips and they're still. So Google is expected to make Morgan Stanley estimates they'll make $81 billion on TPU sales. So obviously this is a very large business. But that adoption hasn't exactly, we haven't yet seen the full adoption. I was talking to one cloud software provider yesterday and he said of his 5,000 customers, only 16 of them are using TPUs at any real volume right now.

Speaker 4

28:01 So I think it's a little too early to. Obviously this is a significant business. We've reported that Oracle and OpenAI are buying TPUs. But, but I think there's still obviously a small part compared to Nvidia's chip business.

Speaker 1

28:19 Great. Okay. Well, Ryan, I Want to thank you for coming on and it's your first time on the show. We're going to have you on more often, my friend, because Google is getting more and more important in the AI story. Always. That is Ryan Barwick, our Google reporter here at the Information. Greg Brockman, The President of OpenAI is in full management mode. That is what he told my colleague Jemima McAvoy in an interview recently as she set out to profile Brockman, who has become operationally perhaps the most important person at OpenAI.

Speaker 1

28:56 I want to bring on Jemima to share more about what she wrote in this story. Jemima, welcome back to the show. It's great to have you here.

Speaker 5

29:02 Thanks for having me.

Speaker 1

29:04 So this was a great profile of Brockman and I want to get to your conversation with him. I figure maybe a good place to start here is what at OpenAI does he actually oversee right now, tactically?

Speaker 5

29:19 Yeah. In short, a lot expanding on that. He is overseeing compute, he's overseeing communications, he's overseeing the sales team, the product team, marketing. And then people that I spoke with said he's kind of all over the place, also working on advertising, robotics, and then parts of the research team. So he's basically ubiquitous across the company at this point.

Speaker 1

29:47 And so what is Sam Altman doing that? What's the division of labor between them?

Speaker 5

29:52 Yeah, that was a key question I had. If Greg's doing all this stuff, what is Sam doing? You know, speaking with Greg and other leaders? The key message they were sending was that Sam is still very much involved. He hasn't gone anywhere. Just more big picture, you know, coming up with these visions for things like, you know, we want to build a chip, jalapeno chip, and then Greg is the one who's operationally executing on a lot of that stuff.

Speaker 5

30:19 So they're really tag teaming on on a day to day basis. And Sam also is very focused on fundraising and speaking with governments and government officials, that type of thing. So a little bit more external stuff.

Speaker 1

30:31 Right. And I mean, if we think about how Brockman's rule has changed, I mean, you go through his extended backstory in the story and I encourage everyone to read it because it's really fascinating to sort of see his story in the early days of OpenAI and how it's progressed, but focusing on the last year, what has changed about the role in that time period?

Speaker 5

30:54 Yeah, and I'll just briefly bring in a tiny bit of the backstory. Not as long as I go into it in the story, just because it helps explain things but for much of his time at OpenAI, he's acknowledged that he was mostly, as he says, IC individual contributor is what I realized that meant. So he was just hopping between a lot of different projects, a lot of different teams. He would come in with ideas. He was very in the weeds, spending 80% of his time, you know, doing these types of projects and a lot of coding and just working on what he felt were the most important things that OpenAI was doing.

Speaker 5

31:27 But he wasn't really managing anybody. He went on his sabbatical in 2024, came back and then took over Compute. And what's really changed over the past year is, you know, Fiji, who was kind of the product czar and leading a lot of that stuff at OpenAI. She had to go out on medical leave and Greg stepped in to take over temporarily the product team. And after she left, the company was deciding whether to hire someone external to replace Fiji, in the end, decided to just divvy up responsibilities between a lot of executives.

Speaker 5

32:00 And so Greg's responsibility was product from that group that Fiji was handling. And over time, he's just expanded and expanded and expanded his role. And people say it's been very organic, but it's really been a pretty amazing transformation over the past year. And just comparing to late 2024, when he was. He had, you know, no direct reports. He's now has so many of these most important people at OpenAI reporting to him, and he's handling a lot

Speaker 1

32:31 well. And I totally forgot he went on leave for a while in there. I completely. I mean, it was very vivid in the moment, but, I mean, this was a big deal at the time, right?

Speaker 5

32:44 Yeah, it was. And we ran a story at the time being like, what is it going to look like for Greg when he gets back from OpenAI? What's his place going to be there? And, I mean, he said, we talked about this, that he had this realization when he was on sabbatical that he needed to change things up and that he wanted to have a bit more of a management role at OpenAI. And he's really taken that too, the extreme now.

Speaker 1

33:07 Right. So you had a chance to speak with him. And I. I want to get to what he said ultimately. But I'm curious, what were the big questions that you went into that interview with?

Speaker 5

33:19 Yeah, I mean, this is obviously a very important time for OpenAI. They're preparing for hopefully an IPO next year. They have fallen behind, as we, as we talked about on a recent show, fallen behind anthropic and in some of the financial metrics, and generally there's a lot that needs to get done ahead of the ipo. And Greg is, in his new role, responsible for a lot of that stuff. So I was very curious about what his strategy is, where this additional revenue is going to come from, how the margins are going to improve, all of those things that we've been talking about a lot.

Speaker 5

33:56 And also how he divvies things up with Sam on. On a regular basis. How does that relationship work? So. So that was a lot of what I was looking to explore in the interview.

Speaker 1

34:06 Okay, and. And so what parts of that did you get clear answers to and what parts still remain? Questions to us?

Speaker 5

34:16 Yeah, I think for me, one of the most interesting things was, you know, I spoke to Greg. I also spoke to Sarah Fryer and a bunch of other execs at OpenAI for the story. And Greg and Sarah both said point blank, we think that cybersecurity is going to be a bigger opportunity for the company than coding. And remember, the enterprise business at OpenAI is now its biggest source of revenue, and that is in large part coding right now.

Speaker 5

34:46 So that's a huge revenue driver for OpenAI. And obviously anthropic has made that a huge business too. So cyber being even bigger than that, I thought was very interesting. I think maybe something that I didn't get answered as clearly was how they're going to get there. You know, there's a lot of competition in the space and they are making traction, but it's, it's, you know, going to be a big jump for them to, to have cyber be bigger than encoding.

Speaker 5

35:16 But everyone I was speaking with was saying that Greg is super passionate about cyber right now, and this is the area that he's perhaps spending the most of his time.

Speaker 1

35:25 Right. Well, that's super interesting because what was it, Forget what the name was for the month of October that we had all these hacks and everyone was like, oh my God, this will be the big use case for AI beyond coding is discovering these vulnerabilities and fixing them. And so that's certainly jives with the headlines that we're seeing, not the least of which is this week we had the Java vulnerability that was discovered. And I mean, I guess Brockman is saying this, which I think is a great strategic move.

Speaker 1

36:00 I wonder what people in the industry, maybe customers, customers of OpenAI, think about this, because there is certainly anthropic with Claude and they sort of brand themselves with security with Project Glasswing and Mythos. You also have Google that acquired Wiz, and so they already have a pretty strong cyber business there. So what confidence do the sources that you talk to have in OpenAI's ability to stand out here?

Speaker 5

36:28 Yeah, I mean, and one thing I'll note that in his role, you know, leading sales, one of the big things that we get into the story that Greg has done is switching Denise Dresser, who, you know, the former Slack CEO who was there for eight months with Dali Rajek, who was the president of Wiz, the cybersecurity company. So obviously that shows that gives some inkling into the fact that they're really serious about doing this.

Speaker 5

36:59 They brought in someone who's much more specialized in that. So as for what sources are saying about whether it's actually going to be possible for OpenAI to live up to its expectations here, that's definitely something I need to be exploring more in depth. I don't think we fully answered that question in the piece, but the customers that I spoke with for the story were one of them. I asked them about even just competition with Anthropic and they said they're using both companies cybersecurity products because they're good at both things.

Speaker 5

37:33 I think it's kind of, I'm curious to see if this is how it plays out bigger picture, but it's kind of like how a lot of companies will be using multiple models because they're good at different things. I could see that being how things go with cyber. And so maybe it isn't just a winner takes all situation, but it's another area where Anthropic and OpenAI are obviously competing very closely. And it's not just them. There's a lot of cyber companies booming right now that they're going to have to.

Speaker 1

38:02 Did you by any chance ask Brockman if he wanted to be CEO of OpenAI or did people that you spoke with expect that he will become CEO of OpenAI? What's the consensus?

Speaker 5

38:16 You know what? I wish I. I wish I would have asked him that. I didn't. I

Speaker 1

38:20 think, I think he probably would have said, I think Sam's doing a great job. I think he would have dodged the

Speaker 5

38:24 question.

Speaker 1

38:24 Honestly,

Speaker 5

38:25 to be honest, I don't think I would have gotten. I mean, I definitely, you know, what is Sam actually doing? Was a question that I was bringing to everyone when Brockman's doing so much at this point, everyone was very insistent that Sam is the CEO. Like, that is still the situation. Greg hasn't, at least from what I've heard privately expressed desire to take that job from Sam. But, you know, it does make you what he's doing right now looks very much like what many people would consider a CEO role to be

Speaker 1

38:59 or, you know, or coo. I mean, you know, let's, let's not, you know, there are many companies out there. The COO does play certainly an outsized role here. I am just thinking though, a little bit of again, we were talking in the show earlier about Nvidia being a quote unquote, more boring company in the post Jensen era. And I'm thinking about OpenAI as a public company and I'm thinking about maybe, I don't know what, what an operator like CEO would mean for the business.

Speaker 1

39:31 Maybe it would mean a little more predictability and maybe that's the argument for making Brockman CEO. You know, it's just,

Speaker 5

39:39 yeah, people were

Speaker 1

39:40 thought, I have.

Speaker 5

39:41 People were very emphatic that Sam and Greg are running the company together and that OpenAI is working on so many things that it's really a two person job. Like they need to tag team these responsibilities. So I don't think I would predict a Greg, you know, completely Greg run situation anytime soon. But you never know. I guess. We didn't expect Jensen,

Speaker 1

40:06 right?

Speaker 5

40:07 Yeah.

Speaker 1

40:08 Right. Okay. Well, Jemima, I want to thank you for coming on. It was a great piece. That is Jemima McAvoy, our OpenAI reporter here at the information that does it for today's show. A reminder, we are on the stream Monday through Friday at 10am Pacific, 1pm Eastern. If you can't make it, then episodes are available on the information.com on our YouTube channel or wherever you get your podcasts. Make sure to follow us on social media, on X, on Instagram, on TikTok and on LinkedIn.

Speaker 1

40:35 I am already excited for our next show tomorrow and I am even more excited for our live stream, our live video summit happening tomorrow at 11:30am Pacific, 2:30pm Eastern. It's going to be a great show all around, a full day of programming. We will see you then. Bye bye for now.

Summary

Nvidia's recent quarterly results, the first without CEO Jensen Huang, showed strong revenue growth but raised concerns about leadership succession and competitive pressures in the AI chip market. The discussion also highlighted Nvidia's challenges with gross margins and the implications of ongoing investigations related to chip sales, particularly in relation to China.

- Nvidia reported a revenue doubling from the previous quarter and raised its forecast for next year.
- Leadership succession is a critical concern for Nvidia following Huang's departure, affecting future strategic direction.
- Gross margins are expected to decline further, influenced by competition in the inference chip market.
- Nvidia's pricing power is under pressure from new competitors and in-house chip development by hyperscalers.
- Legislative measures in the U.S. could tighten controls on chip sales to China, impacting Nvidia's market.
- Jay Puri, Nvidia's former sales chief, is under scrutiny due to his connections with Supermicro amid investigations into chip diversion.
- Google's TPU software is gaining traction, aided by the introduction of Torch TPU, which simplifies integration for users.
- OpenAI's Greg Brockman is taking on a more operational role, focusing on cybersecurity as a significant growth area, potentially surpassing coding in revenue.

Questions Answered

What is the focus of today's show and what events are upcoming?

The show will cover Nvidia's quarterly results, exclusive reporting on Nvidia's former sales chief, and Google's TPU software boom. There is also an upcoming live video summit on Anthropic's IPO.

How will Nvidia's relationship with Washington change after Jensen Huang's departure?

Nvidia's relationship with Washington is expected to evolve due to ongoing legislative efforts aimed at tightening controls on chip sales, particularly to China.

What are the concerns regarding Supermicro's relationship with Nvidia?

Supermicro, a key partner of Nvidia, has faced scrutiny due to past executive actions involving the illegal diversion of Nvidia chips to China, raising questions about preferential treatment.

How does the open-source nature of TPU software affect its market adoption?

The open-source TPU software aims to simplify usage for developers, allowing them to integrate more easily without relying on Google, which is expected to enhance market traction.

What are the implications of Greg's return to OpenAI?

Greg's return is pivotal for OpenAI as the company prepares for a potential IPO, and he aims to take on a more significant management role to address the challenges faced by the organization.

© transcribe · For agents Built with care and craft by Gokul Rajaram