Nvidia's AI Dominance: A Closer Look at the Chip Giant's Future

300 Billion Reasons to Buy Nvidia Stock Now | The Motley Fool

Nvidia (NVDA 1.40%) stock has pulled back from its peak in June as the stock market seems to be debating the future of the AI boom. Some investors think the rally overheated as the new technology has yet to produce a "killer app" or make any major disruptions in the economy. There was evidence of this when Alphabet and Microsoft stocks both pulled back following their recent earnings reports as investors questioned the level of spending on AI infrastructure and wondered if those tech giants would get a return on those investments. However, there are plenty of AI bulls remaining, and among them is former Google CEO Eric Schmidt. In a recent conversation at Stanford University, Schmidt commented on the evolution of artificial intelligence, saying that companies were planning to spend tens of billions of dollars or even hundreds of billions of dollars on AI infrastructure. Microsoft and OpenAI, for example, are planning for a massive data center and supercomputer project, known as Stargate AI, that could cost as much as $100 billion. Schmidt went on to say, "(OpenAI CEO) Sam Altman is a close friend. He believes it's going to take about $300 billion, maybe more..." The "it" here refers to meeting their needs for AI infrastructure. Continuing, he said, "If $300 billion is all going to Nvidia, you know what to do in the stock market," though he added, "That's not a stock market recommendation." Still, it's a reminder to investors not to take their eye off the long-term goal when it comes to AI as the buildout is only just beginning. Keep in mind that Schmidt's quote just refers to open OpenAI's infrastructure needs, so if that's correct, the demand in the industry is much greater than that. In fact, Sam Altman is trying to raise as much as $7 trillion to expand the world's semiconductor industry in order to achieve artificial general intelligence (AGI), or AI that is as capable or more capable than a human. AGI is OpenAI's stated goal, and other tech visionaries like Tesla's Elon Musk are working toward it as well. Investors should expect them to continue to build the needed computing power until AGI is achieved. Nvidia CEO Jensen Huang has predicted that artificial general intelligence is five years away, so the market is likely to see AI investment increase during that time. If you look past the market volatility and doubts among investors, nothing has changed about the tech industry's expectations for AI. The race among the "Magnificent Seven" and other companies will determine the winners in the industry for the next generation and building the necessary infrastructure is crucial. Nvidia remains the far-and-away leader in data center GPUs and other components needed to run AI models like ChatGPT, and it's likely to continue to dominate that market even as competition from AMD and Intel comes online. Nvidia had an estimated 98% share of the data center GPU market in 2023, and the company's data center revenue jumped 427% in the first quarter to $22.6 billion. That growth rate will slow as it laps the initial generative AI investment boom that followed the launch of ChatGPT, but Nvidia will almost certainly grow at a high rate as the race to AGI picks up. Investors can debate the proper valuation for Nvidia stocks, but the strength of its business and its future potential are hard to dispute. The demand ramp for its products still has a long way to go, according to the needs of AGI and statements from the likes of Schmidt and Altman. In fact, Schmidt's biggest concern about the surge in AI infrastructure was finding enough electricity to power these massive data centers. While that could be a challenge down the road, taking his advice and buying Nvidia's stock now makes a lot of sense.

Is Nvidia the Future of AI or a Bubble Waiting to Burst? | The Motley Fool

Nvidia (NVDA 1.40%) has been on an incredible run since the start of 2023. The stock has risen around 700% and has been powered by impressive revenue and earnings growth along the way. A direct line can be drawn from this performance to artificial intelligence (AI) demand. But when massive hype surrounds a technology, companies involved in it can be caught up in an investing bubble. When this bubble bursts, it could take years (or decades) to recover. So, is Nvidia the provider of the future? Or is it a bubble waiting to burst? Perhaps the best comparison to today's AI gold rush is the internet boom in the late 1990s and early 2000s. Companies like Cisco and Oracle were spearheading the stock market with the networking equipment needed to proliferate on the internet. However, that bubble burst, and it took Oracle nearly two decades to set a new high, while Cisco is still below its all-time high. One of the major differences between what happened in the early 2000s and what is happening now with AI is the speed of the buildout. The internet didn't go up overnight, and the demand was less than expected. For Nvidia, which builds the graphics processing units (GPUs) that do the heavy lifting of AI model training, the demand does exceed expectations. Nvidia is selling billions of dollars' worth of GPUs every quarter, a number that has steadily risen over the past few quarters. Furthermore, with the big tech companies purchasing most of these GPUs and telling investors that 2025 will be another year of heavy capital expenditures due to building out the computing power needed for a leading AI model, Nvidia is likely in good shape from a demand perspective. While the demand is there, the stock could still be considered overvalued and stuck in a bubble if the business can't deliver enough growth. But is that the case? Right now, Nvidia recently traded for about 68 times trailing earnings, which is very expensive. But that's the wrong metric to use here because its business is rapidly evolving. Instead, I'll use the forward earnings ratio (P/E), which uses analyst projections for the next 12 months. It isn't perfect, but it provides a better picture because the market is a forward-looking machine. From that perspective, Nvidia trades at about 43 times forward earnings. Which still isn't cheap but isn't as bad. By using those two metrics, investors can calculate that analysts project about 59% earnings growth over the next year. Investors might look at the first quarter's results and see 629% growth in earnings per share (EPS) and conclude that this will be a breeze, but that's flawed thinking. Starting in the second quarter, Nvidia will be up against tougher year-over-year comparisons, which might not provide the same impressive growth figures investors have seen over the past year. These are levels significantly above their historical ranges, which indicates that Nvidia can charge a premium for its products because it's the industry leader. As big tech companies start to design their own hardware that can exceed Nvidia's GPU performance on workloads specifically set up to be optimized on a custom chip, those margins could fall. Nvidia is still delivering great products and strong growth, so it's not a bubble. However, I think it will face increasingly harder challenges over the next year. I consider the stock expensive, with challenges on the horizon. It might overcome them and continue its historic run. Or it could get crushed. I'm not sure which of these outcomes will occur, so I will stay on the sideline for now.

With Its Blackwell Chips Delayed, Should Investors Delay Buying Nvidia Stock? | The Motley Fool

While there has been a lot of excitement surrounding the launch of Nvidia's (NVDA 1.40%) latest chips based on its new Blackwell architecture, both customers and investors apparently will need to wait a bit longer. According to reports, the shipment of the chips will be delayed, although by how long remains to be seen. Given the strong demand that was expected for the chips, let's take a closer look at the issue and the impact it may have on the stock. As first reported by the tech publication website The Information, a design flaw in its Blackwell B200 chip was found "unusually late" in the production process. The issue is believed to stem from the company being one of the first to use Taiwan Semiconductor Manufacturer's new CoWoS-L packaging technology ("Chip on Wafer on Substrate with a Local silicon interconnect," if you're curious) and the placement of the bridge dies connecting two graphics processing units (GPUs) being less than perfect. As a result, Nvidia has apparently decided to revamp the design of its Blackwell GPUs, which is expected to delay the start of shipments by three months or more, according to some reports. Customers and partners have indirectly confirmed the delays. Meta Platforms says it doesn't expect to receive Blackwell GPUs this year, while Super Micro Computer executives say they don't expect any real volumes from Blackwell until the March quarter. Alphabet, Microsoft, and Meta all have huge Blackwell orders worth "tens of billions of dollars" that they are looking to get filled, according to The Information. Meanwhile, it was reported earlier this year that Amazon was moving its Nvidia AI accelerator orders from Hopper to Blackwell. While Hopper orders will likely help fill some of the void resulting from the Blackwell delay, the risk is that there is an air pocket if these customers just wait on their large orders. UBS analysts, however, have come out and said that, after speaking to Nvidia customers, the firm expects the chip delay to only be between four to six weeks and that the delay will be "invisible" to most customers. This is much shorter than the initial delay that was reported, which has helped lift the stock off its recent lows. The length of the delay for Blackwell will likely have a big impact on Nvidia in the short term. A short delay will likely be good for the stock with hardly any impact on its 2025 results, while a delay of three months or more would be looked upon unfavorably, especially after the idea of a shorter delay has already been floated by a Wall Street analyst. There also is the question of whether the design flaw could cause a chip failure or whether it was just leading to less-than-expected production yields. Either way, it appears that the company's decision to delay production of the chip and fix any issues is a smart move. Longer term, the bigger question surrounding the chip issues is whether Nvidia has sped up its development timeline too quickly. The company has cut its planned development cycle for new chip architectures from two years to one. This should keep demand and prices high, but it's also an aggressive schedule with limited room for error or delays. Being at the forefront of new technology and delivering mass production of your products are two separate things that may not always coincide in the future as well. So these are some risks to consider. That said, with a number of customers having absolutely huge orders for its chips, demand is not an issue for Nvidia. Customers are currently more concerned about falling behind in the artificial intelligence (AI) race than overbuilding capacity. As large language models (LLMs) become more advanced, they will need more and more computing power, which means more GPUs will be needed. For example, Meta said its Llama 4 LLM will likely need 10 times the computing power of its previous version to train. And that right there is the biggest reason to own Nvidia. With a dominant market position in the GPU space, Nvidia will continue to be the biggest beneficiary of the continued push for more computing power. At the same time, the stock trades at a forward price-to-earning (P/E) ratio of about 30 times based on 2025 analyst estimates. For a company with the growth and long-term prospects of Nvidia, that valuation is quite attractive. So while there are risks associated with the Blackwell delay, long-term investors can still look to buy Nvidia at current levels.