--- title: "NVIDIA's market cap has already reached 5.5 trillion, how much more upside is there?" type: "Topics" locale: "en" url: "https://longbridge.com/en/topics/40865829.md" description: "People often say that NVIDIA has already reached a market cap of 5.5 trillion, leaving little room for further upside. Instead of buying NVIDIA, it's better to explore companies with greater potential for growth. This argument sounds very reasonable. After all, a company reaching a market capitalization of 5.5 trillion dollars is an extremely exaggerated phenomenon in the history of human commerce. Many people's first reaction is: With such a large company, where else can it go? But this question reminds me of the development history of human data storage technology. In the 1950s, the world's first computer hard disk had only 3.5M. Over the next few decades, it slowly grew, reaching the early 1990s..." datetime: "2026-05-18T06:37:30.000Z" locales: - [en](https://longbridge.com/en/topics/40865829.md) - [zh-CN](https://longbridge.com/zh-CN/topics/40865829.md) - [zh-HK](https://longbridge.com/zh-HK/topics/40865829.md) author: "[爱妻的交易员](https://longbridge.com/en/profiles/17289850.md)" --- # NVIDIA's market cap has already reached 5.5 trillion, how much more upside is there? People often say that NVIDIA has already reached a market cap of $5.5 trillion, leaving little room for further growth. Instead of buying NVIDIA, it's better to explore companies with greater potential for volatility. This argument sounds very reasonable. After all, a company reaching a $5.5 trillion market capitalization is an extremely exaggerated phenomenon in human business history. Many people's first reaction is: With such a large company, where else can it go? But this question reminds me of the history of human data storage technology. In the 1950s, the world's first computer hard drive had only 3.5M. Over the next few decades, it grew slowly. By the early 1990s, ordinary personal computer hard drives were only a few hundred megabytes. The growth didn't seem particularly exaggerated. But then it suddenly exploded: 6G, 120G, 250G, 4T, 60T, 100T… In just a few decades, capacity grew from a few hundred megabytes to the level of hundreds of millions of megabytes. If you lived in the 1950s, even if you were particularly optimistic, it would be hard to imagine that hard drive capacity would increase tens of millions of times over a few decades. Because the human brain is more accustomed to linear growth, such as a little more each year, or tenfold more in a decade. But truly great technological transformations are often not linear growth, but exponential growth. This is also why, when standing in the present to judge the future, we often underestimate the truly major trends. Bill Gates once said a very classic line: People always overestimate the change that will occur in the next two years and underestimate the change that will occur in the next ten years. This statement is especially fitting in the AI era. In the short term, the market will fluctuate violently due to orders, valuations, earnings guidance, export controls, and macro interest rates. But in the long term, if AI truly continues to change the structure of productivity, then the growth in computing power demand may far exceed our intuition today. Today, many people think NVIDIA is already very large, with limited future space. This judgment is not completely unreasonable, because from a short-term valuation perspective, NVIDIA is indeed not cheap, and the market has already given it high expectations. But the problem is: If AI is just a hardware procurement cycle, then NVIDIA might indeed be very expensive now. But if AI is a productivity revolution lasting ten or twenty years, then the current NVIDIA may not be the end point. We cannot simply use "the market cap is already large" to judge whether a company has a future. What we really need to look at is whether the industry ceiling behind it is still moving upward. In the past, we thought mobile phones, cloud computing, and the internet were already big enough, but each subsequent round of technological revolution created an even larger market. The same goes for AI. If AI in the future remains only at the level of chatbots, then NVIDIA's imagination space is indeed limited. But if AI gradually enters enterprise office work, software development, medical research, industrial simulation, autonomous driving, robotics, data center inference, sovereign AI infrastructure, then the computing power it needs is not short-term procurement, but long-term consumption. At this point, what NVIDIA sells is not just chips, but the underlying infrastructure of the AI era. Just as the electricity era needed power plants, the internet era needed servers, and the mobile internet era needed smartphones and cloud computing. The most scarce resource in the AI era is likely to be high-performance computing power. And NVIDIA is currently positioned exactly there. So I think that judging NVIDIA cannot be based solely on the phrase "$5.5 trillion market cap is already large." It's more important to ask a few questions: First, is AI computing power demand still growing? Second, have AI capital expenditures by cloud providers and large enterprises slowed down? Third, has NVIDIA's product iteration capability been weakened? Fourth, have competitors truly breached its moat? Fifth, has the AI application end consistently validated the return on computing power investment? If the answers to these questions remain positive, then NVIDIA's long-term ceiling cannot be simply limited by today's market cap. Of course, this doesn't mean NVIDIA is worth buying at any price. Even the best companies will pull back after rising too much in the short term; even the strongest industry trends will have moments of valuation excess. So for ordinary investors, the truly reasonable approach is not to blindly chase highs, nor to completely reject it because of its large market cap. Instead, it's to distinguish between core holdings and trading positions. Core holdings look at the long-term AI trend; try not to sell too early due to short-term valuation anxiety. Trading positions look at market sentiment and price fluctuations, scaling back appropriately when overheated and gradually buying back during pullbacks. Cash is a safety cushion used to deal with volatility, not an emotional tool for chasing rallies and selling in panic. I've always felt that the hardest part about NVIDIA is not judging whether it's a good company. That question is already quite clear. The real difficulty is: When it's already very expensive, do you still dare to admit that it might continue to grow larger; When it has a short-term sharp decline, do you still dare to believe that the long-term logic hasn't changed; When its big rally is exciting, can you still control your position size and desires. The most common mistake in investing is using linear thinking to understand exponential change. Seeing a $5.5 trillion market cap, one thinks there's not much room left. Just like people in the 1950s seeing a 3.5M hard drive could hardly imagine a 100T hard drive in the future. Humans often underestimate long-term compound interest and also often underestimate the endgame of technological revolutions. So my view is: NVIDIA is already very large, but that doesn't mean it has no room for growth. If the AI revolution continues to advance, its future market cap may still exceed most people's intuition. But as investors, we also cannot use "it might be very large in the future" as a reason to buy blindly. The best strategy is still: Look at the industry in the long term, look at valuation in the short term, and manage position size in operations. NVIDIA's true growth potential does not depend on whether its market cap is $5.5 trillion today. It depends on how high this AI technological revolution can ultimately push global productivity, business models, and computing power demand. ### Related Stocks - [NVDA.US](https://longbridge.com/en/quote/NVDA.US.md) - [NVDL.US](https://longbridge.com/en/quote/NVDL.US.md) - [07788.HK](https://longbridge.com/en/quote/07788.HK.md) - [07388.HK](https://longbridge.com/en/quote/07388.HK.md) - [NVDY.US](https://longbridge.com/en/quote/NVDY.US.md) - [NVDD.US](https://longbridge.com/en/quote/NVDD.US.md) - [NVDX.US](https://longbridge.com/en/quote/NVDX.US.md) - [NVDQ.US](https://longbridge.com/en/quote/NVDQ.US.md)