Microsoft’s True Value of $3.8T, Nvidia’s Cloud Revenue Cut, Can AI Be Conscious?

0:13Welcome everyone to the Informations TI TV. My name is Rocket Drew. It is Thursday, July 2nd. First up, the information published exclusive reporting about NVIDIA agreeing to financially backstop GPU purchases and taking cuts of the revenue. Our NVIDIA reporter will join us to share more about what she learned. And our financial analysis columnist says Microsoft may be valued at more than it's currently trading for. We'll hear the reasoning behind that. I'll also dig into some technical topics with a couple of guests, like how AI models continue to struggle with visual reasoning, and whether or not AI could ever be conscious.

0:53And to wrap things up, I'll be joined by my colleague Laura Bratton to talk through our latest AI Agenda newsletter about model routers. It's going to be a great show, so let's get right on into it. The Information published exclusive reporting that NVIDIA is planning to financially backstop young cloud providers that rent out its AI chips in exchange for a share of their revenues. Our NVIDIA reporter Phoebe Liu broke that story, and she joins me now to share what she learned. All right, welcome on the show, Phoebe. Thanks for having me. Of course. So you wrote that this is part of a program that some people at NVIDIA have dubbed the AI Compute Partnership.

1:33Tell me more about what you learned. Yeah, totally. So there were two deals that were kind of announced in the last month or so from Firmus and ShareNAI, which are both kind of Australian-based new cloud slash data center builders somewhere along that spectrum. And it made me kind of think, okay, if there are two, there must be more. So I started kind of calling around and asking people how widespread that was. And it turns out it is part of kind of a pretty codified program that NVIDIA is starting to provide kind of younger, newer cloud customers a credit wrapper to help them buy NVIDIA's pretty expensive GPUs at lower interest rates in exchange for a cut of recurring usage-based revenue.

2:19And that's a departure from kind of previous strategy where NVIDIA has backstopped leases, provided equity investments, and done other things to stimulate the ecosystem. But it's kind of part of that trend where NVIDIA wants to get other people to buy its products. I think lots of companies do this, but I think NVIDIA has done it at an extent that's very interesting given how big of a role it's playing in the ecosystem. So the way it works is that NVIDIA guarantees that if capacity is unsold, it'll lease it back at a flat rate of X dollars per GPU per hour. I would love to know what X is. and then the revenue share would be kind of a percentage of the upside and scale down over the life of the contract which tends to be five to six years as far as I understand and basically that guarantee makes banks or other lenders more willing to lend money to these newer clouds that like don't have good credit and I don't know they would probably be offered like a 15 % interest rate or something and that may not be a viable business for them slash they wouldn't be able to scale as quickly.

3:26But with NVIDIA's backing, they can get their interest rates much closer to investment grade. So it just seems like another tool in NVIDIA's toolbox of ways that it's helping customers be able to afford its GPUs. Do you think NVIDIA expects that in practice, it will have to rent some of that capacity? Or is the bet that it'll never come to that because the NeoClouds will be so successful? Yeah, I'm sure they've talked about it. I think publicly they posture that they never expect to pay a cent and basically they're just leveraging their balance sheet um because they have such a strong one i don't know they i think they made 160 billion in net income in the last year or something like that which is crazy and yeah they're the world's largest company by market cap so yeah i think they're betting that demand will continue to be astronomical um over at least the next couple years um i think they posture like longer than that, but we'll see.

4:25At least in the case of Firmis, I talked to one of their co-CEOs, Tim Rosenfield, and he said it took them just a couple weeks to get customer commitments for the full length of the contract. So for them, it wasn't hard. I don't know what the case will be for others, but yeah, as far as I know, they haven't had to pay much, if anything yet and don't expect to. But yeah, we'll see. I think there are a lot of people on the internet that feel otherwise, but it's kind of a time will tell situation. Sure. There usually are people on the internet that feel otherwise. But if it comes to that, NVIDIA will be able to afford it because of their balance sheet.

5:07What would they do with that capacity? Would they have some use for it if they do end up having to rent some of that capacity back? Yeah, totally. So I think there is one such deal already with Lambda, where NVIDIA rented some capacity back and gave it to its own researchers, which was interesting because you wouldn't think that NVIDIA's own researchers would need more compute, given that they work for the company that makes it. But that's what the information previously reported happened with Lambda. I don't know what it's going to do with any additional capacity that it rents from these deals.

5:45I don't know if they would consider reselling it to other people at, I don't know, lower rates or something like that, or just use it itself because it seems like NVIDIA is placing a bigger and bigger bet on training its own open weight models, which is a fascinating bet that I would love to talk more about another time. But I think because of that, that would require more capacity and they would definitely have use for it. Got it. What are the kind of risks that these arrangements could create, both for NVIDIA or for the NeoClouds? Yeah, totally. So I guess I feel like when CoreWeave first came onto the scene, everyone was like, oh my god, this is so risky, like they are not going to survive.

6:27And they kept going around saying like, hey, like all of our debt, even though it's high interest is kind of guaranteed by people with good balance sheet by customer contracts, and like taken out against the depreciating value of the GPUs. I think there's still some kind of fear that running a NeoCloud business is inherently very risky because of how highly leveraged it is. Just like getting the capital to buy GPUs at a scale big enough to run such a business is pretty hard. And yeah, there's a lot of debt. It's very capital intensive. And that inherently comes with a lot of risk, especially because nobody really knows.

7:14Like people know that GPUs continue to be useful for a long time, but useful life and kind of economic life is a little bit of a different story. So I think it's difficult to predict what a GPU will be worth in financial terms down the line. And that risk is kind of faked in. But again, I think it's pretty much mitigated by NVIDIA's backstops. And then if it gets to the point where NVIDIA is not able to afford this, I think we have bigger problems with the market and the AI boom, bubble, bust, whatever you want to call it in general. Sure. Yeah, we'll have bigger problems. Yeah. And then what about the revenue sharing part of this agreement?

7:59So in order to get NVIDIA's backstop, the NeoClouds have to promise a portion of their revenue. Yeah, so this is something I'm still asking around on, and I would love to know more about what the revenue share exactly is. My understanding is that it's kind of, so NVIDIA promises that it'll buy back the compute at some kind of fixed rate. And what I've heard, and I'm not 100 % sure on this yet, but the theory is that if a cloud provider is able to sell at kind of a far higher rate than what NVIDIA has guaranteed it would buy it back for, NVIDIA's revenue share would be kind of a difference of those two rates, if that makes sense.

8:41So it's kind of a bet that demand is far, far higher than supply right now, which would impact pricing. And that would kind of scale down to zero over the life of the contract, as far as I understand it. So it would kind of be more cash to NVIDIA at the beginning of the contract. And then by the end, it's kind of letting the NeoCloud go like fly its wings and be independent or whatever. Yeah. And then I guess in terms of the revenue share agreement more generally, I feel like although NVIDIA has been very strong on how it's not explicitly forcing anyone to exclusively use NVIDIA chips and the full stack of NVIDIA products, I think having such a close and financially dependent relationship on NVIDIA would kind of implicitly encourage that loyalty, which would help the NVIDIA ecosystem.

9:35as well. Yeah. So you set this up as sort of the latest step that NVIDIA is taking towards being the central bank, playing central bank for this whole ecosystem. What are you interested in learning next in this story? What do you think could happen next? Yeah, totally. Definitely interested in kind of more specific numbers. My understanding is that a lot of these deals are case by case and kind of depend on each customer's specific situations. So trying to understand kind of what factors are at play there would be really interesting. And just trying to see, I think NVIDIA has offered this pretty broadly, but I don't know how many people have actually taken them up on it.

10:17I know a couple people have declined it because they don't need it, for example. And it's just like, who else is doing this right now that we don't know about? That would be interesting to find out. And also kind of what we've been talking about, like how widespread will demand be? Will NVIDIA ever have to pay a cent? I think there are a lot of other kind of backstopping deals like this. um nvidia did one actually with core weave um last september i don't know or don't think there is a revenue share agreement as part of that but i think when nvidia said they would buy all of core weave's unsold capacity um investors like let out a big exhale the stock went up a ton um and it kind of just like soothes the market so i'm curious if that's going to continue to play out slash if that's going to allow more new clouds to go public.

11:12Yeah, I think the biggest question at the end of the day is how long will demand continue to be as crazy as it has been? And yeah, if I knew the answer to that one, I would maybe be in a different industry. So fair. So fair. Or I don't know, maybe would stay in journalism. I think there's - No drones, right. Yeah. Well, it's a big development. So thanks for coming on and breaking it down for us, Phoebe. Thanks, Racket. Microsoft is currently trading at a$2.9 trillion valuation. But my colleague, Anita Ramaswamy, is out with a new analysis for her True Value column that values the company at up to$3.8 trillion.

11:56Anita joins me now to break down her assessment. Anita, welcome to the show. It's great to have you. It's great to be back, Rocket. So we got to get it out of the way first. Is it just because Microsoft owns so much open AI? Is that why they're undervalued? Yeah, it's a good question, especially as all of these discussions swirl around open AI and when they're going public. My thinking is that it probably doesn't help right now, considering a lot of the questions around the business and its long-term future, but also it's a pretty small stake in the grand scheme of things. What you have to remember is Microsoft is a$2.9 trillion company.

12:29It's massive. That's what their enterprise value is today. And their stake in OpenAI is probably worth a couple hundred billion at most. And so at the end of the day, that's not really what seems to be driving the undervaluation in Microsoft shares. Yeah, that's pretty compelling. I guess writing about AI all day long, I forget there are like other companies besides OpenAI and Anthropic. So what made you decide to take a closer look at Microsoft's valuation in the first place? Well, it's a couple of things. I mean, first of all, I think this 20 % under, you know, sort of correction that's happened, whatever you want to call it, the stock side that's happened this year for Microsoft is notable.

13:04That's more than all of its other sort of hyperscaler peers. So if you look at, you know, the Googles of the world, the Amazons of the world, they've had some ups and downs, some bumps, and certainly a tough month. But they are not trading down as much as Microsoft is on a year-to-date basis. And so I wanted to examine, you know, this is a company that has tons of resources, very deep pockets, and has been investing in AI from the early days, really, when it comes to Gen AI. And so how is it that their stock is not doing well and have they lost their lead? The other reason I thought it was interesting to look at Microsoft's valuation right now is just the idea that, you know, a lot of times when you see these conglomerate sort of businesses that have different business units that do different things, they tend to trade at a discount relative to what each of those business units would trade if they were just standalone businesses.

13:53So I thought it was an interesting thought exercise. You know, maybe years from now, we could even see a scenario where companies like Microsoft or Amazon or Alphabet choose to break up because their businesses are valued so differently and sometimes even attract different investor bases. What is the intuition for discounting that basket of different businesses? Is it that maybe one of those business lines might get cut on a whim and we don't know how it's going to go? You know, I've always wondered this because the conglomerate discount doesn't make a whole lot of sense to me, but I think it's something along the lines of just different types of investors looking at different businesses in different ways.

14:29And so some investors who are, let's say, focused on semiconductors might value that a certain way versus they might not place as much emphasis on a software business. I also think that part of it has to do with the focus of the company and where they're investing their resources. I see. So you say it's okay, it's reasonable to have some kind of discount, but Microsoft's is too steep. So let's go sort of bucket by bucket through each portion of Microsoft's business that you consider here. So you look at cloud, software, and personal computing. Let's take them each in turn. Let's start with the cloud side.

15:02Tell us what you found there. Yeah. So what I found on the cloud side, Rocket, is that this is the hardest part of Microsoft's business to value. As you mentioned, they have three different business units and cloud is the fastest growing, and it seems to be the most promising when it comes to AI. They are growing at a rate of about 40 % year over year. And it's been really sort of breakneck growth with the demand for cloud computing. Their Azure unit is housed within this cloud division, and Azure is the one that's growing around 40%. If you look at the overall cloud segment, there's a couple of other things in there, like software that helps support some of their cloud infrastructure operations.

15:39So that growth rate is a little bit lower when you look at the cloud unit overall. all, but the engine and the growth driver really is Azure, which is their core cloud business. The problem with this business is that it's on the lower margin side compared to Microsoft's other core business of selling software. So they're doing around 42 % operating margins, and this is growing fast. So the question investors are asking themselves is, will this dilute the margin over time? Got it. Got it. Okay. So that takes us into the software side. The margins are better there, but do you think there's something else that's undervalued about the software side of the business?

16:17Yeah. So I guess before just getting into software, I did want to mention that the conclusion I came to in terms of actually valuing the cloud business, despite how challenging it is, is that it's probably worth somewhere in the ballpark of$1.7 trillion. That's based on some similar companies like Google and Amazon, but we didn't really have a good comp in terms of a standalone cloud business that we could compare Microsoft's cloud business to. So it's possible that's a little too high, but my estimate is that that's worth 1.4 to 1.7 trillion. Now let's talk about software because this is their quarter business.

16:52It's their biggest business. It's the thing that Microsoft is really known for, and they have a huge advantage in the fact that so many enterprises use Microsoft's Office Suite. But, you know, the hard part about software, while it is very high margin, around 58 % operating margins, they are not growing nearly as fast as the cloud business. They're growing it around, you know, the high teens growth rate. Well, I imagine they're also getting swept up in the SaaSpocalypse concerns to some extent here. Is that a factor? Yeah, definitely. So when you look at Microsoft versus Alphabet, I said that, you know, Microsoft has traded down more than Alphabet a year to date.

17:27But if you look at like Salesforce, ServiceNow, SAP, Microsoft is actually, their stock performance this year has more closely mirrored some of the pure play software names. So that suggests that investors are looking at Microsoft and, you know, thinking that they are going to be vulnerable to disruption from things like Cloud Cowork out of Anthropic and other AI tools. Sure. Yeah. Someone can sit down and use Cloud Code to vibe code their own version of Microsoft Word. then what do they need the actual Microsoft Word for? On the other hand, that's a good sign for their cloud business, presumably, that people would use AI so much.

17:59Yeah, it totally is. And there are some bright spots in the software business, I would say. I mean, we have Copilot, which is Microsoft's own sort of AI assistant agent that they've been promoting. But the challenge is that so far, at least the investors that I spoke to and the customers have not really said that, you know, they prefer using Copilot over other services. It's more that they're using it just because it already comes in their software stack. So, you know, we have to see how the technology evolves. But the bottom line here, Rocket, is that this business, if you look at, you know, comps that are similar in the software business, is probably worth around$1.7 trillion.

18:35It's possible that that is way too high of an estimate. You know, that's just my two cents on it. But if you look, there are some investment banks that have floated estimates like Goldman Sachs. They value software business closer to$500 billion. So it just depends on what you think the future of the SaaSpocalypse is going to look like. Got it. Okay, so that's the cloud side, the software business, and the last bucket is personal computing. What did you find there? So this is Microsoft's smallest line of business. It's the least sort of scrutinized by investors. And that's because it's smallest in terms of revenue.

19:10It makes up around 20 % of their overall top line. The margins are really low. They're closer to the 2020s range. versus the other two in the 40s and 50s, this is the unit that houses the gaming division. And they're going through a big shakeup in gaming right now. They have new leadership. We've recorded on that quite in depth. But at the end of the day, this is just a small unit. And so the Surface PCs, and if you think about the hardware investment involved in this personal computing unit, it is very different from the other two lines of business Microsoft has. So what I did to try and value this unit is that I took some comparable companies such as Apple, HP, Sony, and Electronic Arts.

19:51Sony and EA are video game makers, so that kind of reflects some of the value in this business. This business is actually shrinking in terms of, or not shrinking, sorry, it's decelerating in terms of its revenue growth. And so that's an area of concern for the future. And so the value I ultimately ascribed to it was$350 billion based on some of its industry peers. Got it. Okay, so that's how you get to this total of up to$3.8 trillion for the whole business. And yet, like you said, the stock is down 20 % over the past year. So why is that? What are the big concerns that Microsoft is facing? One of the investors I talked to for this story put this really nicely, and he sort of framed this as a bit of a double whammy that Microsoft is going through right now.

20:35On one hand, you have the SaaSpocalypse fears, and people see Microsoft as a core software business. There is this perception that Copilot has not kept up technologically to some of the faster growing alternatives like Cloud Cowork. So there's SaaSpocalypse fears. And then in the second bucket of concerns, there is the question that investors are asking of all the hyperscalers about, are the billions of dollars that you guys plan to spend and have been spending on capital expenditures to build out your cloud businesses, is that going to be justified by the amount of revenue that your AI products actually bring in?

21:09And so that remains to be seen. And that's something that all of these bigger tech companies that have been investing heavily in AI have gotten punished for lately. And Microsoft is at the perfect nexus and intersection of those two sort of negative outlooks right now. Got it. Well, that certainly puts Microsoft in an interesting position in the industry right now. Thanks so much for coming on and walking us through your thinking about it. Definitely a company that will be very interesting to watch in the coming months. Thanks for having me. The most advanced AI models still cannot solve some visual puzzles that even kids can solve.

21:43Andrew Dye worked at Google for 14 years, but now he's focusing on the visual reasoning capabilities of AI models at his new startup, Elorian. Andrew joins us now to tell us all about it. Welcome on the show, Andrew. Thanks, Tremont. Thanks for having me. Yeah, absolutely. I'm excited for this conversation because visual reasoning is a really fascinating area of AI right now. I want to start by understanding the limitations of current models. I think some people will hear the idea that there are limitations. They'll say, well, what do you mean there are limitations? ChatGPT makes me great pictures all the time.

22:16Isn't that visual reasoning? So where's the gap that you see right now? Yeah, ChatGPT is great at making pictures and I use Google Lens all the time to identify plants and flowers. These current models we have and these tools are really good at object recognition, what in the computer vision industry we call segmentation, localization. These are used every day by all sorts of tools. But the fact is that even though they're very good at pattern recognition for complex visual reasoning tasks, so the more common type of tasks you see in kind of like white-collar visual work, so these might be things like design or say you're a realtor and then you need to produce a floor plan that looks exactly like the house that you're selling or you're an architect that needs to design an office for exactly 50 people no more no less These models, their reasoning and individual understanding isn't at the level of humans yet.

23:27So they will be great if there are, say, a dining table and you're asking how many people sit around the dining table. But if you're in a crowded room and you ask how many people are there, then they really struggle. And this is because for simple tasks, you can do pattern recognition. you kind of find an impression of something whether that's a certain kind of client or three or four people around the table. But as soon as you have a question that requires more than one second of looking at an image to figure out the answer then these models aren't able to reason like that. Got it. Okay, those are some really helpful concrete examples.

24:10I guess it's a little surprising that visual reasoning has gotten left behind. I mean, there's this trend right now of reasoning in models that has uplifted code and math reasoning. It's a little surprising that visual reasoning has gotten left behind. Is that because visual reasoning tasks are hard to verify? It's hard to automatically grade a model's performance on them? Or what are the reasons that visual reasoning is left behind? Yeah, there's several reasons. I'd say the main one is that the way people are used to using AI is in these restricted settings, right? As an average user, most of your cases for using images and video, like I said, it's just to identify this planet you saw in your park, right?

25:00But for more complex use cases, like I mentioned, like understanding like CAD drawing. People don't think of that as an everyday use case. And the thing is that people don't use AI for these other types of work. And what that has led to is that that's led to the community focusing a lot on these recognition problems. And then you do see some of the benchmarks like MMMU, which is also called massively multi-modal benchmark, those benchmarks actually can be solved without looking at the image in a lot of cases. And then the other benchmark that is quite well known is also called ACK-AGI. This is commonly brought up as a multi-modal reasoning benchmark.

25:50But even in this benchmark, the images that are being shown to the models are only 32 by 32 or 64 by 64 pixels. And it's very hard to take a real-world problem like a floor plan and shrink it down into that many pixels. It's just not possible. But there are some visual reasoning benchmarks that are coming up. The one that has really struck for me was this one called Baby Vision, released earlier this year. And in this benchmark, they do show that the frontier models, the best of them still reason around the age of a preschooler. And any elementary school kid in their benchmarks were able to beat all the frontier models on these visual tasks.

26:36What are the kinds of tasks in that benchmark, baby vision? Yeah, so these are tasks that are not just object recognition. It's about spatial reasoning, more like geometry, understanding differences. So they have tasks like navigating through a maze, following one string along to see what's at the other end of a string, counting things that are spread across a table or a desk, finding differences between two images, finding things that are the odd ones out. So a lot of this is kind of like more multi-step, what sometimes people call visual thinking, where you have to think mentally and mentally imagine different kinds of scenarios to solve the problems, like navigating a maze.

27:28I see. Okay, so the industry is dropping the ball on visual reasoning because of the limitations of benchmarks. The benchmarks don't reward visual reasoning. They're not sort of measuring visual reasoning in a naturalistic way. That's your point with Arc AGI. The settings are kind of too toy. But your point that people aren't using the models for visual reasoning. Is the concern there that we're not able to get data from production about how people are actually using the models and then use that for training? Or that the labs don't think that the demand is there? so they're not investing in these capabilities?

28:01Yeah, so I think it's a bit of both. So the labs, definitely, including Gemini, which is where I was a data area lead, there is investment in multimodal understanding, and you see that through Project Vio, Nano Banana, but it's much more on the generative side. That's quite simply where they see the demand. But it's kind of like a chicken-and-leg situation. Until you have demand, there's no reason to invest a huge amount of resources in building better capabilities there. Unlike coding, where the demand is already there, so then the strategy is obvious. But yeah, it's a chicken-and-leg because the demand is not there, but the demand is not there because the models are just not good enough for these use cases.