NVIDIA Just Revealed Its Next $200 Billion AI Opportunity!!

Hey guys, welcome back to another episode on What the Chip Happened. What I believe is the greatest semiconductor and AI YouTube channel out right now. Now, in today's episode, what I want to do is take a closer look at Nvidia because Nvidia just announced something pretty interesting that I believe is completely changing the AI market once again and that I believe is extremely bullish for the growth opportunity of not just Nvidia, but the rest of the ecosystem as well. But before we take a closer look at that, I also want to share some recent updates impacting OpenAI. There has been a slowdown in their growth that has some investors a little bit worried and I want to share my overall thoughts there as well. So, let's take a closer look in today's episode. But before we begin, if you're serious about semiconductor and AI investing, I break down major earnings and conferences inside my community. Institutional quality research built for retail investors, get 33% off at whatthechipappen.com. Now the later part of this topic we're going to talk about Nvidia and how it's trying to secure land power and shell because that's the next critical resource for AI factories. Now luckily if you are part of my community at what the chip happened this is something we talked about this morning during the exclusive live stream. So make sure to check it out at whatthechipappen.com. So OpenAI this morning announced the following that as models become more capable the risk associated with developing and testing them internally also grow. They mentioned that they temporarily paused reinforcement learning training on our latest models intended for deployment for two weeks while we harden and redteamed our research environments and expand monitoring coverage. So they do have kind of a a full list here of like hey our AI models for the past few weeks we've seen a bit of developments in the AI industry. First that open AI hugging face incident. Now they're saying that preliminary evidence of their upcoming model Astra is showing that this model is insane for cyber security capabilities and that's not something that they could theoretically release if they don't feel comfortable and obviously could create a lot of issues with things like national defense. So they have to play this safely. They kind of talk a little bit of how they're trying to slow down and what they're going to be working on in the next few weeks to try to improve their security and try to make sure that they're able to contain uh this model. Uh so a lot of it it's interesting as we're going to see the company did see a bit of a slowdown in growth in revenue. So, some skeptics might be saying this is just open AI trying to fight that negative sentiment coming from some of that negative news on the growth numbers that we're about to discuss and try to give investors a little bit calmness in their mind. If I was an investor of OpenAI and then I heard this, it's like, okay, now I have the ability to not be so fearful of my investors. So, there are some skeptics that are saying this is more of a PR stunt uh to be able to showcase to investors that they shouldn't be worried about their investment. I don't know necessarily how much I believe that, but what I would say is in this AI market, it changes very often who is the leader. The past few months, the past few quarters, I believe Anthropic has been the leader in the AI industry. Most recently, I personally just from personal experience have started using codecs a lot more and OpenAI models a lot more than Anthropic. So, it's not impossible to believe that while Enthropic might have won the first half of this year, it's not impossible for OpenAI to do better in the second half of the year. Now the other thing that I found pretty interesting here is they were kind of mentioning that when you are trying to safeguard compute and these models you also need a nice amount of compute for to be able to do that safeguard for that monitoring overhead. And to me this is actually kind of like a case of hey if AI needs to get stronger and better you also need to have AI rails and AI monitoring and AI overhead be stronger as well. So it's interesting to think that that alone can also create a bottleneck of compute because if you are using and here's 20% but it's not 20% of their whole compute it's 20 of the inference compute being monitored it it still indicates that hey there has to be some form of compute usage for this and again that reduces the total compute overall. So I thought that was interesting and would love to hear some of your thoughts there about that. Now when we look at the real numbers, so Wall Street Journal reported that Wall Street, not a Wall Street Journal, I don't know why I always think of world star, you know, world star, Wall Street Journal reported that OpenAI told investors its revenue grew by 18% from the first quarter to the second quarter. While its losses deepened, results that disappointed some shareholders who hope the startup will show more catching up to rival Anthropic. The company said its revenue grew to 6.7 billion in the 3 months that ended in June, up from 5.7 billion in the first quarter. Now, they mentioned margins sank further. And maybe this is also one of the main reasons OpenAI has been pushing their IPO a little bit longer than expected. Obviously, they had the legal issues. Uh, but now that kind of that is a little bit over, I think they're going to want to show that they're winning before they go IPO so they can get as much money as possible. Now, while OpenAI did about 7 billion, Anthropic more than doubled its revenue to 11.6 billion in the same period, marking the first time its sales surpassed its older rival. The company also supposedly swung small on small operating profits. Now, this is why I'm excited to hear more about the Ashwans from these companies and going public because we will be able to really see the numbers, right? These are all reports based on people familiar with the matter so on and so forth. So one of the main reasons of success for Anthropic is Enthropic did really good with cloud code, right? I think that's kind of one of the main AI coding solution um that really grabbed attention. Now OpenAI is definitely doing a a lot better. But this fear, right, this fear of Open AI slowing down, I'm going to believe is going to create fear for the semiconductor industry because the first thing that a bear can say, see, open AAI is not going to have the money to pay all the funds and all these RPOs that it's making with all these cloud providers. This was supposed to be a company that's growing, supposed to be growing massively, even on a sequential basis. What is going on? It is right to make that point because it is true. OpenAI is supposed to be growing at crazy levels. Now, to me, one thing that I have always stated in this channel and I still believe to be true is I don't care what AI model wins. The demand for intelligence will always be there no matter who it is. If open AI fails, Anthropic is going to win, right? Let's say you have let's say you need 100 intelligence. Let's just put a number in intelligence 100. It could I don't care if it's 50/50. I don't care if it's 6040. I don't care if it's 991 because one of those players is going to need all the compute. The great thing with a lot of these AI solutions is a lot of these AI chips, especially from the market leader like Nvidia, if OpenAI fails, Nvidia and those chips can easily go to Anthropic, right? And and and we're going to see kind of with what Nvidia is doing in a bit to really showcase this, right? It doesn't really matter because anthrop intelligence will be needed and it doesn't matter who serves it. It's just whoever has the most amount of intelligence and whoever has the most intelligent product will be able to have the demand of all that compute out there. But again, we love to see more of the S1 and we love to see more updates. I'm pretty sure OpenAI is going to read this Wall Street Journal piece and is going to kind of share some of their overall thoughts about it. I know sometimes anthropic and open AAI are set to the to for revenue to be recognized differently. So maybe open AAI can share as oh look if we actually recognize this revenue similar to anthropic this would be our revenue run rate but would love to hear from the team in the upcoming months. Now again right now it is an interesting pace. I do believe it's going to create negative sentiments toward the overall market but at the same time I don't believe it matters who wins in this AI space. It could actually even be Grock or Meta, Mark Zuckerberg with Meta. One of those is going to win and whoever wins is going to need all the compute cuz intelligence intelligence demand is going to continue to grow. Now, the next part that I want to focus on is this, right? Nvidia on August 17th, so yesterday, announced that securing the infrastructure of intelligence, land, power, and shell, the next critical resource for AI factories. And again, if you have been following this channel, this should not be new news to you, right? Nvidia is one of those players that tries to reduce the hurdles of anything that is in the way of AI development. One of the first hurdles was what? Actual AI labs. So Nvidia makes sure to help AI labs continue to grow, continue to innovate cuz then you're able to create a product. All right. So what's after creating a product? Nvidia decided to help out creating AI infrastructure dramatically and increase the supply chain. It did this through various ways. It did this through building up neoclouds. It did this through showing how to design servers with its MGX platform. It did this by showing how to run AI cloud with its DGX cloud. It showed how to design data centers, which is what it's doing right now. And and then we saw just a few a few weeks ago, it's trying to show the market how to properly finance the space. So Nvidia is trying to hit all these hurdles from either supply chain or just overall AI development. Now, one of the other issues that they see in the supply chain is the land power and shout. If you are an AI company and you're trying to sell gigawatts of AI chips, it's very hard for your customer to say, "Hey, let me find not hard. They can do it. A selected amount of you can do it, but not everybody has the ability to do it or has the full teamwork to do it. It's like, hey, I want to buy a gigawatt of chips." Great. But if you're going to buy a gigawatt of chips, you need to be able to have land power shell to do it. And Nvidia wants to make sure that you don't just buy a gigawatt of chips and you don't know what to do with them and you have a lot of chips just turned off and create this fear of kind of an AI bubble. So Nvidia is doing now where they're reducing the supply chain risk. Hey, we're going to rent out we're going to lease out some of these land powers and shells because then when you come in and order 1 gawatt of compute, I can tell you look, you don't have to wait for them to turn on. This is what I have. I have this whole supply chain ready for you from chip building, from massive server racks, from building the data center to now we also are available to help you with the land power and the shell. So I think this is a massive massive win for Nvidia and Nvidia can do this because it has that massive market share. Now what I thought was pretty interesting is Nvidia announced a massive deal with OpenAI in the AI economy. Compute is revenue and they mentioned AI factories require a full stack of critical resources advanced chips packaging memory and networking and Nvidia in the past few years has been making sure that this supply chain portion is done extremely well right it's done investments in optics it's made great partners and great investments in kind of the manufacturing for both its chips for also the servers and so much more right they've been working with Foxcon they've been working with TSMC they've been working with AMOR one way or another and the list goes goes on and on. So for the whole semiconductor manufacturing, they have helped out. But they say we are no longer a chip company. We are an AI factory company. And for us to continue to work our way down to the AI factory supply chain, we need to also make sure that we have the land power and shell or some way to improve the supply chain there. So Nvidia with its long-term visibility and supply chain partnerships has now worked to secure critical resources and is now applying that same discipline to that LPS. And the main reason is this also allows them to kind of I would say not hoard up but if Nvidia has that LPS then that compute that's going to be hosted in that LPS is going to be Nvidia's. One of them is going to be this ports Pike technology campus in Ohio and OpenAI will be the tenant. Now interesting OpenAI will be the tenant even though OpenAI is going to be game over. There has to be a little bit more to that growth story. Is there some form of compute issue? Is there just was the model just really that weaker? And if we are seeing improvements in July and August with the new models from OpenAI, what is the growth rate looking like right now? And Nvidia says that look a good portion of our big players, our large cloud providers, they have investment grade enterprises and have balance sheets for a long time. So they can secure LPS independently. But there are other companies that are just starting off or that are not as nearly as strong as the cloud providers or for example Frontier AI labs. These are the open AIs. They don't have that type of balance sheet like all the other players but they have that huge revenue growth except open right and anthropic has that huge revenue growth right now but because of that they are not able to do this LPS grab Nvidia in my opinion is trying to do a few things first it wants to continue to build up this ecosystem extremely well because more demand more AI chips the other thing is it doesn't want there to be limited amount of players if there is limited amount of players Nvidia gets squeezed in potentially margins and competition vision and whatnot. So, Nvidia wants there to be a competitive landscape. Also, a competitive landscape in the AI industry focuses innovation and innovation drives new product development and new product development drives new customer and new demand for solutions which continue to create that cycle of AI compute needs. So, they mentioned the initial deployment is expected to be 4.2 gawatt of AI factories. That's a lot. They mentioned it's roughly about 150 to 200 billion in Nvidia revenue per generation and over the 20 years the site can support upgrade cycles. So Nvidia is focusing on these big AI facilities where they can re-upgrade to newer generations every few years maybe every 5 years and that would be an extra $200 billion per segment. They also have the ability to increase this capacity by 3.7 to roughly 8 gawatt. So OpenAI is taking this and has comm committed to this. Again, this is interesting because OpenAI has had a lot of fear today. And I don't think Nvidia is going to make a partnership with OpenAI, especially something this big if OpenAI is really as scared or as bad as kind of these press release are making seem. The other thing that Nvidia has mentioned is Nvidia is selectively securing exceptional sites. So this is pretty good. I think they're only going to be focusing on big sites. The other thing is the productive life of the site extends through multiple generation each capable of producing more intelligence and more revenue than generation before. So just showcasing that Nvidia believes this is going to be a multi-revenue not just for its customers but for itself as well. The final thing the Nvidia's compute is what happens to ports if OpenAI does not use the site in the future for example if OpenAI fails right that's this question right here and I think this question explains it all and Nvidia's compute is versatile fungeible and broadly adopted the capacity can be resold to another qualified tenant across Nvidia's global ecosystem of cloud providers enterprises AI labs and startups so to me this is Nvidia's way of kind of saying what I was talking about earlier today where while it doesn't we don't know who the AI leader is going to be the thing that we know is intelligence is going to be needed. So whoever is the winner in the smartest model will need that compute. If open AI fails here that's okay cuz that compute can just be sold to the actual winner. Obviously, the market if there is that fear of open AI failing and other stuff, I do believe it will create a lot of volatility in the market, but I think the long-term goal is still pretty much the same and the long-term thesis will still play out as I see it. I I'm not a guarantee here, but that's my overall thoughts. Similar with all these crazy backlogs, right? The market is going crazy with the amount of backlogs in cloud server providers and OpenAI doing all this. But like I previously stated, even if OpenAI fails, I think that backlog will be used by the other player instead or players, right? It doesn't mean that there's only going to be one winner. But I this is one of the things that I do believe the market is going to get wrong and if there ever is a fear of one of the AI labs failing, they are going to sell off everything when that's not necessarily what's actually happening. It could be happening, but it's not a guaranteed. So, I hope you guys enjoyed today's episode. Like always, if you want to hear my thoughts, I share them at what the chipappen.com, my overall community. We've posted so many earnings report already. It's insane. We have a power terminal for the AI market, and the list goes on and on. So, make sure to check it out at whatthechip.com. Take care. Have a good day. And see you all next