Robotics for Logistics Startup, Harvard Professor on AI Chip Wars, OpenAI Commerce | Nov 5, 2025

0:13Welcome, everyone, to the Informations TI TV. My name is Akash Prashreech. It is Wednesday, November 5th. We have got another big show planned for you today with another big slate of earnings that we're going to break down for you. Today, we're talking about AMD and Pinterest earnings, two very different stories, but very timely with all the AMD announcements as of late and Pinterest shares that were moving down big time this morning. We're then going to talk about a story that we published this morning about how OpenAI's applications and commerce strategy is sitting with consumer companies. And we're also going to talk all things chips with a longtime board member from Intel.

0:51Finally, we're bringing on another company from our 50 most promising startups list from this year, this time from the robotics category. It's going to be a busy show, and so let's get right on into it. AMD reported earnings last night. It was a closely watched quarter with the company announcing its big deal with OpenAI and with Oracle in recent months. Although shares were slightly sinking after hours, they were about flat this morning. Joining me now to discuss the chip company's quarter is John Vinn, Equity Research Analyst at KeyBank Capital Markets. John, it's great to have you here. Welcome to the show.

1:27Great. Thanks for having me, Akash. So what stood out to you from the quarter, this quarter for AMD? I think notably and what was the most watched metric for AMD was their data center revenues. It did not disappoint. Revenues were up over 20 % on a year-over-year basis. There were two things that was driving that. One was the ramp of the latest MI350 5G AI GPUs, which grew substantially on a year-over-year basis, and also service CPUs. They've been launching a new server CPU. Codename is Genoa, and it's been doing extremely well. It continues to gain share at the expense of Intel. And those are the key drivers of the results and guidance for them.

2:15Can you break down each of the segments for us? I mean, we think of AMD as this Challenger 2 NVIDIA. It obviously has a very complicated business. They do a good job of at least breaking out their revenue into the segments. Can you talk about what is in the data center revenue? What is in the client and gaming revenue category? And then the embedded segment is another one that we don't pay attention to that much. Yeah, the data center segment is made up of two major subsegments, right? You've got server CPUs, right? This historically has been driven by traditional data center demand. And then obviously the AI GPU business, they are an up and comer in this segment of the market.

2:55obviously beginning from like virtually zero share they're probably going to do close to seven billion dollars in revenues uh this year you know obviously still a distant second to nvidia which is probably going to do uh well over 200 billion dollars in ai gpu revenues but that's the that's the data set segment client and gaming segment is a combination of their client pc notebook as well as their desktop business. And then the gaming business is gaming GPUs for desktops. And also their semi-custom business, which goes into gaming consoles from PlayStation and Xbox. And then the embedded business is mostly the former FPGA business that they acquired when they bought Xilinx there.

3:44And so talk a little bit about, I see here, client revenue was a record $2.8 billion, up 46 % year over year. What was behind that growth there? That's some pretty staggering growth. Yeah, again, a common theme for AMD is they have been gaining share a lot at the expense of Intel primarily. So in the PC notebook space, they've been getting share there against Intel. Intel obviously has a lot of challenges there on the PC front. And then on desktop, desktop is probably the area where they're doing extremely well. That's kind of an area that Intel has admittedly taken their eye off the ball, and as a result, they're hemorrhaging share there in the desktop business.

4:28And let's focus in on that CPU segment here. How much of a threat is AMD right now to Intel or the other way around, Intel to AMD? Talk about that rivalry. You know, it's a segment that has been dominated by Intel. They historically have had 80 % share. I think we've seen AMD's share creep up over the years from 20 % to probably 25 % to 30 % at this point in time. I would expect that AMD would continue to gain share over the next year or so. Intel has very ambitious plans to kind of stem that share loss. They're obviously focusing on taping out their latest ship called Panther Lake on 18A Foundry.

5:11But that's going to take some time to kind of stabilize there. But I think the other thing that you want to watch on the PC front is they are both under some threat of share loss to ARM. ARM's taken a lot of share in the PC market, notably at Apple. Apple's gone to ARM architecture across the board. We've seen other merchant suppliers like Qualcomm launch their ARM CPU. It's getting kind of modest share right now, but over the longer term, we do think that ARM is going to continue to gain share on the CPU front. When you look at AMD on a fundamentals basis, what do you think of the valuation right now?

5:50we think it's fairly uh fairly valued at this point in time right you know the valuation is you know it's trading at um roughly you know 40 times our next year's earnings numbers um that doesn't kind of factor in the potential contribution from the open ai deal that they recently announced or all that's another kind of three dollars in earnings but we think it's uh we think it's fairly valued at these levels since and it's you know trading at a premium to nvidia that we prefer at this point in time, trades at a premium to Broadcom that we also prefer. And as you were listening to the call last night, what questions still remain for you about the company?

6:28What did you hope that they would give you more clarity on that still remains a little bit of an open question? I think there's two questions. They most recently over the last couple of quarters have stopped disclosing what their AI GPU revenues are. They now kind of lump it within their data center segment. They used to kind of give you kind of a breakout on a quarterly basis. They do give you some breadcrumbs so you can come up with an approximation of what their AI GPU revenues are. Obviously, that's an important metric that allows you to kind of measure them up against the incumbents like a broad comp and like an NVIDIA.

7:07I think the second kind of question, you know, we'll have to kind of wait to get more clarity on this is next year they are introducing a significant architecture change to their GPU solution. They're going to launch MI450, which is their next generation GPU. And they're also going to launch Helios, which is their rack architecture. So you're going to go from this paradigm of them basically just chips to now shipping full-stack solutions to their customers. And we saw that NVIDIA went through a lot of growing pains this year to get their rack architecture up, GB300. I think there are similar concerns that we have around what their transition plan and how smoothly they're going to be able to transition to this Helios rack architecture as well.

7:58Right. Well, John, I want to thank you for coming on. It was a very interesting quarter, certainly with the OpenAI and the Oracle News, and it was interesting to see how that played out in the results. That is John Venn from KeyBank here on TITV. Pinterest also reported earnings last night. It caused quite a bit of a stir among investors with shares sinking this morning. AI was, of course, a big focus on the call. I want to bring on Catherine Perloff, our advertising reporter, to break it all down. Catherine, good morning to you. How are you doing? Hello. Hello. Hi, Akash. Let's talk Pinterest, and then we're going to talk also about a story that you published earlier this morning.

8:33But Pinterest, gosh, shares are moving big time on the earnings. What did you make of it? Were the results really anything that bad, or what was the concern? I don't think the results were—I think they were pretty good. They've been growing by, like, around 17%. I think it was 16 % the first quarter. They've been growing pretty fast this year and faster than they did last year. And Bill Reddy, the CEO who came in in mid-2022, I think has really turned around the company since, yeah, in the pandemic when they were losing users and maybe lost focus a bit. I think the investor reaction was because their guidance was for lower growth than they've been growing in the past couple quarters.

9:14I think there were some, you know, missing investments type of thing where, but basically the guidance was revenue growth in the fourth quarter for 14 to 16%, which is a little bit slower than the previous. It's a little bit. It's not that much. No, it's not that much. I do think obviously the fourth quarter is the shopping season. It's supposed to be the big quarter for advertising. And, you know, so I think that was maybe why there was an outside reaction. But yeah, I do think it was maybe a little bit too concerned from investors, though I do think this broader space of social media companies is so dominated by Meta and Google.

9:48People always really scrutinize like some of the smaller, smaller firms like Pinterest or Snap. Well, and certainly when you see a company like Meta, their revenue comes mostly from advertising. They're growing faster than Pinterest. They're obviously a much bigger company. And so I think it does pose a little bit of a structural question around you have a smaller company that's growing slower. What does that say about the overall positioning in the sector? How much did they talk about AI on the call last night? Yeah, you know, I think they obviously they have, they've released a bunch of products, especially last week.

10:20They, like, released a new shopping assistant, which is sort of a more chatbot-like experience where you can kind of ask more conversational questions about, like, I want to, like, you know, have an outfit for a wedding or, you know, in a more conversational way. And they'll produce recommendations from the kind of litany of photos that's on the platform. There was questions about sort of the threat of ChatGPT and also these commerce integrations that, you know, Walmart has integrated with ChatGPT. And what did they think about that being very much in the sort of advertising and commerce space?

10:54I mean, they don't do transactions on Pinterest, but they're advertisers. They're commerce companies. And they were kind of like, well, we already are an AI-powered shopping assistant. And, you know, when they say AI, a lot of that is kind of the older machine learning. But it's true. And they have always, you know, used AI. They also said that, yeah, they have 600 million monthly active users. I think ChatGPT says they have 800 million weekly active users, so that's more. But they're kind of like, we're still in the conversation. We're a place people go. And, you know, don't count us out, basically.

11:31Well, I think that's a good segue into a story that you published this morning along with our colleague Anne Guillen, who reports on all things e-commerce. And I actually want to bring Anne in to help us chat about this. So Anne and Catherine, you wrote the story today about how OpenAI has sort of positioned itself as a could-be super app, right? They're integrating commerce. They're integrating a lot of these consumer websites into their platform. And the great thing about your story is you went out to the potential partners for OpenAI and said, how are you thinking about these integrations? Are you in favor of it?

12:07Are you stampeding towards it like companies like Etsy and Walmart? Or is it giving you pause about how the business model could change? And what are these consumer websites telling you? And also maybe just sort of define the scope of which websites and which companies we're talking about here. That would sort of help us anchor things. Sure. So Catherine and I spoke to several retailers, several e-commerce brands, several big consumer internet companies. that have historically, you know, like Pinterest, attracted a lot of commerce companies as advertisers. And so we spoke with a wide range of them.

12:50And really, the thing that kind of kept coming up over and over again for some of these companies that are maybe a little more reluctant to partner with OpenAI or want to kind of wait and see how this is all going to play out is these companies really just want to, they want reassurance that they're still going to have a lot of control over the user experience and the shopper experience. That's been a big thing that's come up for me just more generally in my e-commerce coverage over the past, you know, several months and even, you know, a couple of years is brands really want to control shopper relationship as closely as they can.

13:31And so introducing something like ChatGPT and their commerce features that potentially kind of moves shoppers a little bit further away from the brand and from their own website, that's something that is definitely making some brands, not all of them, but some brands are definitely thinking twice about whether this is something they want to really pursue. And what are the long-term risks that they are thinking about? But this idea of adding a middle layer to the relationship or disintermediating the relationship, what is their concern long term? Sure. So, I mean, all of these brands and retailers, they want to own the customer relationship, which is really the name of the game.

14:13They want to have your contact information so they can send you marketing emails and text messages. And they want to know your purchase history. They want to know what you're browsing for on their website. All really valuable data points. that they can use to market to you and ultimately get you to come back and purchase from them again and again. So I think what some of these brands are worried about is if some of your transactions are happening on ChatGPT, are they going to have as complete of a picture of you as a shopper and have that kind of really important understanding that they get from you when you come to their website?

14:53So that, from the retailer perspective, that is kind of their worry about introducing a middleman into this process, if you will. And Catherine, what is the advertising angle here? Because I imagine this could ultimately, if OpenAI owns this relationship, it means really you can't advertise to these customers because they're not really on your platform anymore. Yeah, I mean, you know, I think that theoretically the ads could still show up within the kind of chat GPT interface. But, you know, I don't think it's been worked out what the revenue model would be or if OpenAI would get a cut of that.

15:28I think, you know, it's interesting. A big trend in the past couple of years in the world of advertising is more and more retailers are creating their own ad businesses. You know, we see like Walmart has a big ad business. Obviously, Amazon has a big ad business. You know, Instacart, though, Uber, all of these sort of, you know, not traditional media companies are having ad businesses because they can sell the fact that they have a bunch of attention from users. And if you don't have, you know, if you don't know who those users are, as Am was saying, and they're not directly on your site, it's hard to make that pitch.

16:02So I think that's the concern. And that was a concern folks talked about with us. But I think, you know, pushing back on that, you know, this is still an experiment and we don't know if it's going to work. So, you know, we talked to this company, Thumbtack, which is one of the original partners on this super app announcement, and they're a place where you can, like, book a plumber or someone else to help with home repairs. And, you know, right now, the CEO told me that it might not be a better user experience to do that through ChatGPT than through their app. So, you know, maybe ChatGPT is just kind of a symbiotic partner for them and getting them more users.

16:38But the relationship is not necessarily, you know, parasitic or, you know, bad at this point if ChatGPT doesn't sort of disintermediate them, which it might not. But this does sort of strike me as part of what the Amazon and perplexity dispute was just yesterday, right? I mean, we saw that news playing out. This is, again, Amazon saying, well, if we don't own the relationship, then it is a threat to our business, and we ought to target that. And what is OpenAI saying about all this? Well, so OpenAI is kind of counter to a lot of this is that they have designed this so that merchants still own the customer relationship.

17:19They were thoughtful about designing the checkout feature in a lot of different ways. You know, merchants control the payment processing. They do get customer information that they need to fulfill orders like shipping addresses and things like that. When you make a purchase through ChatGPT, it shows up on your credit card statement as the merchant that you purchased from, not ChatGPT or OpenAI, which, you know, isn't always the case with some of these other kind of new AI shopping tools. So they have tried to design it in a way where shoppers are still interfacing with the merchant in some way.

17:58I think just the concern from merchants and retailers is more just kind of that behavioral shift if people get used to going to ChatGPT as a shopping destination versus going to some of these individual sites. But OpenAI is definitely saying that they've designed this in a way where merchants still have control and where some of these other app companies and other internet companies have control. So it'll just be interesting to see how the consumer behavior shifts and what kind of some of the downstream impacts of that are. Right. Well, Anne and Catherine, I want to thank you for coming on. That is Anne Guillen, our e-commerce reporter here at The Information, and Catherine Perloff, who covers all things advertising here at our newsroom.

18:47Okay. It is quite the moment for the chip sector, with the big cloud companies touting their own chips and OpenAI diversifying its bets, so it's not only relying on NVIDIA. Meanwhile, companies like Intel and Broadcom find themselves in interesting situations as well. And so I want to bring on David Yaffe, a professor at Harvard Business School and a former longtime Intel board member for a bit of analysis on what's going on right now in the chip sector. David, welcome to the show. It's great to have you here. Thank you for having me. I want to talk about chips, but I just want to tee up a little bit of the discussion that we just had in the last segment, which is OpenAI finds itself now doing commerce.

19:28We've seen they've got a social media network with their Sora 2 model. They're is everything, everywhere, all at once. And I want to get your take on this strategy as a whole, because you are a strategy professor. You've been teaching this for many decades at HBS. Look, the strategy seems to be working pretty well for them. The numbers, they speak for themselves. My question for you is, did you anticipate that this strategy would work so well for them? Has it exceeded even your own expectations? Is it a good strategy? Well, you have to look at it from two perspectives. Obviously, the revenue growth has been spectacular, but their losses have been equally spectacular.

20:10So it goes in both directions. As a strategy professor, you try to look at both sides of the equation. But nonetheless, there's no question they have been growing faster. They have been innovating faster. Their ability to attract great people and even bring them back after they're being offered spectacular bonuses from places like Meta certainly exceeds almost everybody's expectations. Right. Well, let's shift to the chips side of this story. I want to get into the different angles that the chip companies are taking, but broadly, we've seen some circularity and some vendor financing, I guess, in the chips cloud and AI ecosystem.

20:51One of the questions we like asking people who have such a great personal view on things is what they think of all this. What do you make of all of it? Well, when I was on the Board of Intel, we actually didn't do this. It was an explicit decision not to invest in our customers' businesses. And there were a lot of reasons for it. I think that NVIDIA has got a different competitive position now. It's much stronger than Intel was even at its peak. It has a little bit more flexibility to do things. It's clearly driving demand for GPUs. It's a very, very high margin business for NVIDIA and that makes a lot of sense.

21:32It's locking some customers in because once people get totally committed to the CUDA and the NVIDIA environment, then they have more switching costs to try and go to alternatives, which I actually believe will be emerging, let's say in the next year or two. On the other hand, there's a lot of risk associated with it. They could be stimulating a lot of artificial demand, which is not sustainable. They have certainly vendor risk, just like we saw in 2000, when people were investing heavily in the telecom sector and saw huge losses as a result. So that's the problem with circularity, which is on the one hand, you are stimulating demand.

22:12On the other hand, if it isn't actually sustainable demand, And you end up with potentially a lot of very bad debt and a lot of poor consequences down the road. You say alternatives coming up in the next year. Who are you watching? Are we talking AMD here? Who are you paying attention to? Well, actually, I'm paying attention to a lot of different players because it's not just AMD. I mean, AMD is the only company that really has a GPU competitor that can compete directly with NVIDIA on the core technology. But there are two other things we have to be thinking about. One is that there are a variety of other types of accelerators, whether we're talking about TPUs at Google or some of the other capabilities that Amazon and Microsoft are trying to develop that would allow them to do some of the same things in a more narrowly focused way, but without the expense of a Blackwell GPU.

23:08In addition, the big question down the road is how much of this capacity that's being put into place is going to be used for training, which requires a GPU, or is going to be used for inference. And for inference, you don't necessarily need a high-powered GPU. You can use a CPU and an accelerator, which is a fraction of the cost. So down the road, the question is, will that same demand for NVIDIA GPUs that we see today be sustained or will we start to see a tail off? So do you actually see these hyperscalers chips, the TPU and Tranium, which were very much at the center of the earnings calls last week, do you see them taking significant market share from NVIDIA?

23:54It depends on the time frame. If you said to me in the next 18 months, the answer is absolutely not. If you said to me the next three to four years, then I think it starts to eat into the share. And what about the CPU race here? You've got Intel, you've got AMD. This is a space you know very well. How much of a threat does a company like Intel pose to AMD, or how much traction is AMD getting eating into Intel's market share? Well, AMD has very successfully eaten into Intel's market share in the last five or six years. I suspect that that is going to continue for at least the short run. The real question is, can Intel get its act together?

24:37And getting its act together means two things. One is it has to figure out what its real strategy is with its foundry in manufacturing. Is it going to actually compete with TSMC or is it going to be more heavily focused on its own internal demand? And second, it's got to actually get refocused on products. I think Pat Gelsinger had a very ambitious strategy, but it was all focused on the foundry. They lost a lot of their focus and attention on making great new products that would potentially be able to compete with NVIDIA or Broadcom or AMD. And I think that the current management is at least moving in that direction.

25:16And so on those decision vectors here, what ultimately is your recommendation? I know you're not on the board anymore, but look, I did these cases. I went to an undergraduate program, the Ivy Business School in Canada. We did the cases too. At the end of it all, you got to give a recommendation. What is your ultimate recommendation to Intel then as to how they should play along both those decision vectors? Well, I've written about this extensively, a number of pieces in Fortune over the last year. And my view is that they actually need to spin off their manufacturing, create a separate manufacturing operation that would then be able to compete with TSMC, and that their product business should then really focus on making products in this space of CPUs and accelerators that would enable them to compete much more effectively with AMD and NVIDIA.

26:09There are lots of reasons why I think they should spin off the manufacturing. The most important one is that Intel never had a service culture. And being a foundry, servicing other companies, really is a real service business, not a product business. And Intel failed every time in my almost 30 years on their board of trying to be a high-quality service provider. So I think they need a different culture. And they also have the problem that if I'm NVIDIA or AMD, I'm going to hesitate before I put my chips into an Intel fab because it's still going to be a potential competitor down the road. And so what makes it a much more effective foundry is if it's independent and there's no perception of conflicts of interest.

26:54Right. Well, it's a fascinating race to watch. And Intel is certainly in a moment here that I anticipate there's going to be more changes to come. David, thank you so much for coming on the show. We really do appreciate it. Thank you for having me. Okay. As we talked about yesterday, this week we've released our list of the information's 50 most promising startups that our sources cannot stop telling us about. Every day this week, we are previewing a company on that list. Today, we are talking about Pickle Robot. The company is building robots, not for pickles, but for supply chain automation.