The Rundown - Deep Dive: Nvidia’s AI Chips Are the New Oil (And That Could Make or Break Its Business)

Episode Date: May 31, 2025

Nvidia isn’t just a tech company anymore. It’s powering the AI revolution, shaping national security strategies, and brokering deals with world governments. From Saudi Arabia’s new supercomputers to U.S. export bans, Nvidia is now at the center of a global arms race, and it might be too important for its own good. In this episode, we unpack how Nvidia became the most powerful company on Earth… and why that power comes with a price.This video is for informational purposes only and reflects the views of the host and guest, not Public Holdings or its subsidiaries. Mentions of assets are not recommendations. Investing involves risk, including loss. Past performance does not guarantee future results. For full disclosures, visit ⁠⁠⁠⁠Public.com/disclosures⁠⁠⁠⁠.

Welcome back to the rundown for another weekend deep dive. Today, we are talking about a company that you might have heard it before. Invidia. It's probably not an exaggeration to say that Nvidia is now the most important company in the world. They've become the backbone of the global AI economy. And these days, they're doing more than just selling billions of dollars in AI chips. They're caught in a full-blown geopolitical tug of war. The U.S. government is banning exports of their best chips to China.

other countries like Saudi Arabia and the UAE are buying up Nvidia's latest chips like their Pokemon cards. And Nvidia's CEO Jensen Huang, the leather jacket wearing rock star of the tech world, is now bouncing between Beijing and Riyadh like a one-man diplomatic envoy. So in today's episode, we're going to break down how Nvidia quietly became the supplier of record for the AI revolution and why everyone from multi-trillion dollar tech giants to sovereign nations suddenly can't build anything without them. We got a good one for you today.

So let's dive in. Before Nvidia was powering the entire artificial intelligence industry, they used to be just a scrappy chipmaker trying to make video game graphics look better. The company was founded in 1993 by Jensen Huang, a Taiwanese immigrant with a Stanford engineering degree and a big bet. He thought the world was going to move towards graphics-heavy computing. And my man was not wrong. See, that might seem obvious now, but that wasn't the case back in the day.

Back in the 90s, the CPU was the star of the show. That's what most people cared about when it came to their computers. The GPUs, which was the graphics card that Nvidia made, were just a sidekick at the time. That was a chip that was in charge of the visuals, but the CPU was the brains of the computer. So not many people cared about the GPUs until the late 90s and early 2000s. That's when there was an explosion of 3D computer games. And man, that was big for Nvidia's business, because, Because the better the GPUs, the better the graphics you would have when you played the games.

Invidia graphics cards became a crucial component for PC gamers. And for the mid-2000s, Invidia was a gaming graphics card company, and they were really good at it. But then something happened in the 2010s that changed everything. In the early 2010s, researchers at Microsoft, Google, and Stanford noticed that the same GPUs that were built to make Call of Duty look good were also perfect for training neural networks. So the math behind deep learning is insanely complex. It requires doing millions of matrix multiplications all at once. CPUs could barely keep up, but GPUs were built specifically for this kind of parallel processing.

And then the breakthrough moment happened in 2012 when a team from the University of Toronto used Nvidia's GPUs to win the ImageNet competition. Their model called AlexNet blew everyone else away, and this basically became the blueprint for modern AI. In fact, Jensen Huang himself calls this the big bang for modern AI. And it was a light bulb moment for him. He saw that Nvidia GPUs could be used for more than just gaming. So while other chipmakers like Intel were still focused on PC chips and AMD was chasing console deals,

Nvidia went all in on AI. Nvidia started rebranding themselves not just as a gaming company, but as a platform for accelerated computing. And Jensen made some strategic moves inside of Nvidia, like increasing the investments and development of Kuda, which is Nvidia's secret sauce. Kuta is a software platform developed by NVIDIA, and it was released all the way back in 2007. Now, without getting too technical, Kuda essentially tells GPUs how to do calculations efficiently. And it makes it easy for developers to build AI models. And here's the thing.

Once you start building on Kuta, switching is a nightmare. You know, it's kind of like buying an iPhone and being locked into the Apple ecosystem where all your photos and text messages are on ICloud. You're probably not going to switch. Kuta is kind of like that. Once you start building on Kuda, you're unlikely to switch. And Kuda only runs on Nvidia hardware. So for the last decade, a big chunk of the AI industry was built using Kuda.

So while most people still saw Nvidia as a gaming company, they were quietly building the infrastructure that would later power everything in AI from self-driving cars to chat GPT. And this is what I love most about the Nvidia story, because they didn't just luck into the AI dominance they have today. Jensen Huang saw this coming. He had Nvidia doubled down on Kuda after AlexNet,

way before anyone had heard of ChatGPT. So once ChatGPT was released and kicked off the AI arms race today, Nvidia was ready to capture all the incoming demand for their chips. And it made Nvidia into possibly the most powerful company in the world. And now that's starting to catch the attention of politicians and governments all over the world. All right, so let's fast forward to today because Nvidia just dropped another monster earnings report this past week. NVIDIA's revenues were up 69% year over year, and their data center business, which is where all their AI chips live, was up 73%. And the star of their show was their Blackwell chip, which is Nvidia's new next-gen AI chip platform.

It already makes up nearly 70% of their data center revenue. On the earnings call, Nvidia said that Microsoft alone had already deployed tens of thousands of Blackwild GPUs, and collectively cloud giants are buying 72,000 of these chips, every week. And just for some context here, each chip costs like $30,000. And Nvidia is selling 72,000 of these chips every week. So basically these Blackwell chips are like Taylor Swift concert tickets of enterprise computing. They're expensive, but everyone still wants them. Tech companies these days are spending like it's the dot-com bubble all over again. Oracle plans to spend $40 billion on Nvidia chips to power Open AI's new data center in the U.S.

and then Elon Musk said that his new Colossus factory in Tennessee will need 1 million Nvidia chips. So yeah, Nvidia probably not going to have any demand problems anytime soon. But here's the thing. Invitya still had to take an $8 billion hit to their revenue this year because of the U.S. government's export restrictions to China. Earlier, this year the Trump administration blocked Nvidia from selling their H20 chips to Chinese customers. And that's a big deal because these H20 chips were custom designed for the Chinese market. There are watered down version of their high-powered chips, specifically designed to comply with rules set by the Biden administration. But now, Nvidia isn't allowed to sell AI chips at all to China right now, which

resulted in a $2.5 billion revenue hit in Q1 from unsold H20 chips. So this is having a meaningful impact on Nvidia's revenues. See, before all these restrictions, China used to account for over 20% of Nvidia's revenues. Now that's down to 13%. Meanwhile, US customers make up about 47% of Nvidia's sales, which is more than double what it was five years ago. So, Nvidia is becoming more dependent on the U.S. market. Now, some analysts on Wall Street didn't seem to be too concerned about the revenue hit that Nvidia took from the China export restrictions, because NVIDIA actually forecasted that the damage would be about $1 billion less than what they had predicted back in mid-April. Some analysts still see a lot of upside if the U.S. and China

can finally hammer out some sort of trade deal. Wedbush analyst Dan Ives said in a note responding to the earnings that the $8 billion run rate could easily come back to the NVIDIA story if NVIDIA can reenter the Chinese market with a new restricted chip. CEO Jensen Huang also made some interesting points on the earnings call. He warned that banning NVIDIA from China isn't going to stop Chinese AI companies from progressing. If anything, it might help local Chinese chip makers like Huawei by shielding them from U.S. competitors.

In fact, he went as far as say that whoever wins China wins AI guys. globally. Nvidia said that they expect China to become a $50 billion market with or without the U.S. So I got a feeling that Jensen might be sending a subtle message to the Trump administration. But I think he's trying to play both sides here because in that same evening, Jensen Huang went on Bloomberg and praised the Trump administration's economic policies. Take a quick listen to what Jensen had to say. The first one is utterly visionary. The idea of tariffs being a pillar of of a bold vision to re-industrialize, to onshore

manufacturing, and motivate the world to invest in the United States. And it's just an incredible vision. I think this is going to be a transformative idea for the next century for us. So yeah, it seems like Jensen is playing geopolitical gymnastics here, but Jensen really

doesn't want to give up on China just yet. In fact, Nvidia still has plans for China, which is making some senators here in the U.S. kind of upset. We'll talk more about that. in a bit. At this point, you can make the case that NVIDIA isn't just a tech company anymore. They're like a geopolitical force. Because in today's world, AI chips have become more valuable than oil, and NVIDIA controls the main pipeline. And that's why NVIDIA is stuck in a delicate balancing act. U.S. lawmakers are pressuring NVIDIA to stop fueling AI development of rival

nations, while the rest of the world is begging to buy more chips. And that puts Jensen Huang in more of a diplomatic and lobbying role than just a normal big tech CEO. Jensen used to delegate all this politics and diplomacy stuff, but now there's just too much at stake. So he's taking matters into his own hands. And in the past month, he's gone on a full-on world tour. A few weeks ago, he was in Riyadh talking AI sovereignty with the Saudi Arabia's crown prince. In the next week, he was in Taiwan giving a keynote address at a conference.

And then he was in Shanghai, opening a new inviative. research center over there, even while U.S. lawmakers aren't so happy about it. Jensen is now getting called out by politicians. Senators Jim Banks and Elizabeth Warren wrote a letter to Jensen Huang accusing NVIDIA of undercutting national security interests. They warned that opening up this research center over there risk giving the Chinese Communist Party access to NVIDIA's cutting-edge technology. So as you can see, NVIDIA's role is starting to look less like a chipmaker and more like a diplomatic broker, a broker with $3 trillion in market cap

and more pricing power than OPEC at this point. Now, we're seeing this play out with Saudi Arabia, the first confirmed destination of Nvidia's Blackwell supercomputers. As part of Saudi Arabia's Vision 2030 to become an AI superpower, Saudi Arabia launched a new state-backed company called Humane AI, which is fully owned by the country's public investment fund. Humane bought a ton of Nvidia chips

with the promise of buying hundreds of thousands, thousands of more over the next five years. And look, Saudi Arabia is just the beginning. All over the world, countries, especially ones with deep energy reserves and strong central governments, are racing to build their own large language models, their own chat GPs, their own surveillance tools. But they all need the same thing. They need Nvidia's hardware. So this is likely the next phase of Nvidia's growth. The last five years we're supplying big tech companies like Google, Microsoft, Amazon, and meta with a ton of Nvidia GPUs, the next five years might be about supplying sovereign governments. Jenton even has a term for this trend. He's calling it sovereign AI. And it looks

to be Nvidia's next big opportunity. And even with all the controversy and the export bans and the congressional pressures and rising competitors, Nvidia is still really the only game in town. Sure, AMD is trying to make some moves to catch up. But at this point, they just have a fraction of Nvidia's data center revenues. In fact, some days, Nvidia's market cap moves more in one day than the entire valuation of AMD. Intel is also out there, but at this point, they're just trying to survive. So nobody else seems to be even on the map when it comes to threatening Nvidia's dominance. At this point, it doesn't even matter who's building the AI models. It could be an American company like OpenAI or Meta, or it could be a Chinese company like Deepseek or ByteDance,

or it might even be a Saudi company, Humane AI. Ultimately, all these companies will need chips to build state-of-the-art models, which means that Jensen Huang might be one of the most powerful men in the world, because he's not just the CEO of a big tech company, he's now the kingmaker. So that's the story of Envideo, a company that started off trying to make video games look better, and now they're powering the future of everything, including warfare, surveillance, national AI models, and trillion-dollar economies. But as the company becomes more powerful, they're attracting the attention of governments all over the world,

including the U.S. The U.S. is already restricting the company from selling their chips to China. So the question isn't whether Nvidia can stay on top with the best AI chips. It's whether they're even be allowed to. Or will Chinese chip makers one day catch up to Nvidia with Nvidia cut out of their market? These aren't just business questions here. These are global power questions. But it's just crazy to think that in a very short amount of time,

Nvidia went from powering the graphics of war games like Call of Duty to potentially powering the AI of real life warfare. Well, all right, guys, that's it for today's weekend, deep dive. Hope you guys enjoyed today's episode. If you did and you have like 12 extra seconds, consider giving us a five-star rating on Apple or Spotify. And if you are listening on Spotify, consider leaving us a comment on Spotify, vote in our Spotify poll. All that engagement really does help us out and it helps other people find the show. Now, I made a joke earlier in the week that if enough people listen to this episode,

we might be able to get Jensen on for part two. I know it's a bit ambitious, but I had to shoot my shot, right? So your engagement and feedback on this episode could one day make that a reality. Thank you guys again for listening. Hope you guys enjoyed that one. Have a great weekend. And of course, shout out to Mike and Connor for all the help behind the scenes. And we'll see you guys back here on Monday.