Holding NVDA

Breaking down the Nvidia business for everyday investors

Hello.

I want to talk about the company, the business, the stock: Nvidia (NVDA).

Really, another NVDA article?

Yes, there are MANY articles and opinions out there about NVDA - the majority of them more positive and bullish, and the minority more skeptical and bearish. And yet, I’ve struggled to find one that paints complete but accessible narrative for the average person. Typically, I see discussions that are:

• Very technical about NVDA stock details (price predictions, P/E ratios, RSI, forward earnings, EBITDA margins)

• Very detailed about Nvidia’s underlying tech (tensor cores, deep learning accelerators, the intricacies of their chip architectures)

• Heavily focused on Nvidia’s AI future, assuming the reader is armed with all the necessary historical context

My guess is that even the best of these discussions - while super compelling and insightful - can quickly alienate the average person. And if that’s happening, then this average person probably lacks a solid grasp on the fundamentals that make Nvidia so exciting, how and why Nvidia “works.” I have to assume that this person would prefer a better understanding here, and with that, the ability to confidently assess the company’s long-term future.

So, below, I’m hoping to build that understanding and confidence with an accessible overview of what Nvidia (the business) does, and why I hold NVDA (the stock).

“Let’s not talk about the stock price”

As of today, early November 2024, NVDA is trading at around a $140 per share. But, I want to make this clear - I don’t think the stock price (within reasonable fluctuations) matters much right now. Bold claim? Here’s my point - I buy and hold NVDA for the long-term - 10 years or more. I am expecting and hoping for LOTS of growth from today’s price. So, whether next week or next month the price is $160 or $120 per share, my confidence in its long-term potential is the same. That adage “time in the market beats timing the market” applies here, as in, “time holding NVDA beats timing a given purchase of NVDA stock.”

A minor caveat - I do try to reassess each stock I hold (generally fewer than 20 stocks) at least once a quarter. “Are the reasons I bought it still intact?” “Has anything fundamentally changed about the business to alter my confidence?” More often than not, the answer is “no” and I continue to hold, or buy more.

Ok, time to dive in.

The AI Revolution

We’ll start here: Nvidia is no longer just a tech company that (mostly) makes fancy graphics cards for gamers. Today and going forward, Nvidia is at the center of the AI revolution.

This is maybe where I start to alienate some AI-skeptical readers, but bear with me. I’ll try to convince you that AI is much more than hype and buzzwords. I firmly believe that the potential for imminent, wildly impactful AI is very real, but also, only recently so. This newness is an interesting place to dig deeper. Historical context matters a lot here.

For decades, AI has been a topic of constant speculation, often described as being on the verge of transforming industries and economies, yet repeatedly falling short of those expectations. Historian and author Matt Ridley discusses how AI was, for much of its history, overhyped and under-delivered - promising breakthroughs that always remained just out of reach.

What’s different today - and why I believe this moment is a tipping point - is that the technology has finally caught up with the vision. AI is now making real, tangible progress. Recent breakthroughs in machine learning and neural networks, particularly the development of transformer architecture in 2017, have changed AI from a theoretical concept into a powerful, real-world tool.

These advancements, especially in unsupervised learning, allow AI to move beyond narrow applications into more generalized intelligence - able to generate text, create images, and even write code. At the same time, hardware breakthroughs, led by companies like Nvidia, have enabled the large-scale training and deployment of these models. This combination of cutting-edge software and high-performance hardware has turned long-term AI visions into rapidly accelerating reality.

So, basically ChatGPT?

About two years ago, ChatGPT debuted and forever changed the future path for AI. Is THAT the AI revolution I’m talking about? Not at all. But, it was a big movement forward. These large language models (LLMs) can suddenly process vast amounts of data and perform tasks that were unimaginable just a few years ago. And they’re quickly evolving - every few weeks there’s an update to one of them (in the current mix: Chat-GPT, Perplexity, Claude, Gemini, Llama, and a handful of others).

While the underlying tech is impressive, the primary shift for the average person has been the sudden, public accessibility of AI. Seemingly overnight, AI is interactive, user-friendly, and mainstream. This, as much as the tech itself, fuels excitement and discourse about the future.

And yet. Has that much changed? Critics point out that, despite all the hype, AI hasn’t fundamentally transformed daily life for most people, even with tools like ChatGPT. Sure, AI can generate grocery lists, make photo filters, transcribe podcasts, manage calendars, and write a sophomore history paper - cool, but hardly revolutionary. Beyond fun tricks and conveniences, there’s little impact on our day-to-day routines, Others point out that, despite the prevalence of “AI” on earnings calls across nearly every company in the S&P 500 - AI isn’t actually monetizing at scale yet. Isn’t that kind of important for it to become anything close to revolutionary?

Fair points. But, I think this is missing the bigger picture. The true value creation from AI is still in its infancy. Sectors like healthcare, finance, entertainment, manufacturing, retail, transportation, and even agriculture are just beginning to see how AI can solve complex problems and drive efficiencies. From automating supply chains to optimizing energy use and reimagining customer experiences, the breadth of AI’s impact is only going to grow - and grow rapidly - across almost every industry.

Tech leaders are heavily investing in AI, but not for any immediate returns. They understand where this is heading. Industries are racing to either catch up or get ahead, recognizing that AI is going to fundamentally reshape the world. We are on the verge of another big shift, much like the Industrial Revolution in the 19th century (which transformed manufacturing and society) or the rise of the internet in the late 20th century (which redefined communication, commerce, and information-sharing). Just as those technological leaps revolutionized industries and daily life, AI is poised to become a similarly transformative force - one that will redefine how we work, solve problems, and interact with the world.

Show me the money

Recently, I came across an insightful post on Reddit that laid out some AI use cases I hadn’t heard before. That drove me to dig up many more. I’m hoping that the summarized collection below will help my overall point here: there’s HUGE growth, and HUGE financial incentive, in the current and future paths of AI.

Below are examples of what AI is either already doing or on the verge of doing across various industries:

• AI in healthcare: Larry Ellison from Oracle described how AI can dramatically reduce manpower in healthcare while enhancing the patient experience. AI handles many repetitive tasks currently being done by one or multiple doctors and nurses - from analyzing sonograms to summarizing patient histories. This frees up medical professionals to focus on more critical care.

• AI’s role in insurance: A mid-sized insurance company is using AI to automate the analysis and generation of insurance policies. A task that currently requires around 1,000 employees will soon only need a few dozen, reducing labor costs by 95% while improving speed and accuracy.

• AI embedded in consumer technology: The iPhone 16, as one example, will have increasingly more AI functions embedded over the next year - features like object recognition and restaurant details with a simple camera point. This is only the beginning of what AI could eventually do in consumer tech.

• Retail and Personalized Marketing: AI analyzes customer behavior in real time to provide tailored recommendations, personalized promotions, and inventory management. Companies like Amazon use AI to predict buying patterns, maximizing sales and minimizing waste.

• Agriculture and Crop Management: AI helps farmers analyze soil conditions, forecast weather, and optimize crop yields. Companies like John Deere use machine learning to reduce water and pesticide use, driving sustainable farming practices.

• Banking and Fraud Detection: AI identifies suspicious transaction patterns, enhancing fraud detection and security measures. JPMorgan and other major banks rely on AI algorithms to monitor and prevent fraudulent activities, saving billions.

• Transportation and Logistics: AI powers route optimization, predictive maintenance, and autonomous driving. UPS and other logistics companies utilize AI to streamline delivery times and reduce fuel costs, significantly boosting efficiency.

• Energy Sector and Predictive Analytics: AI analyzes data from sensors in energy plants to predict failures, enabling proactive maintenance. Siemens uses AI in smart grids to optimize energy distribution, reducing costs and carbon emissions.

• Global IT spending and AI’s productivity impact: Annual global IT spending is expected to reach $5 trillion in 2024, with an 8% increase over last year. If AI merely boosts productivity by 10%, that’s $500 billion in potential annual value creation, without even factoring in entirely new products and services.

All these examples paint a picture of AI not just as a fun tool, but as an embedded element in countless sectors. And this is mostly stuff happening already. If AI keeps advancing and evolving at a rapid clip (which, thus far, it is), consider how much power it has over the next few years to reshape entire industries, drive productivity, and create value on a massive scale.

What was once an ambitious, unfulfilled vision is now quickly become reality. And, at the heart of it all, is Nvidia.

Nvidia: “The Electrical Grid” of AI

Nvidia’s technology is the infrastructure enabling the AI revolution, generating the power needed to run everything from machine learning algorithms (examples: ChatGPT, Perplexity) to the data centers that run AI models (examples: Google Cloud, AWS). Nvidia is laying the groundwork for AI and is uniquely positioned at the very center of what I believe will be a transformative shift in how industries operate. Their combination of high-performance hardware and cutting-edge software has turned long-term AI visions into a rapidly accelerating reality. This is what fuels my confidence in their long-term growth potential as a company and a stock.

To illustrate Nvidia’s place and impact in the world today, especially as it relates to the AI revolution, I’m going to make what I hope is a useful analogy: the electrical grid, each part within it representing a critical layer of the AI ecosystem:

• Electricity (AI itself): Electricity powers industries, transportation, and homes, just as AI powers a growing number of applications in healthcare, finance, entertainment, and beyond. AI, like electricity, is becoming foundational to nearly every sector.

• The Electrical Grid (Nvidia): Nvidia serves as the grid - the foundational infrastructure that enables AI to reach every corner of the economy. Just as the grid makes electricity available to homes, businesses, and industries, Nvidia’s technology distributes the power of AI by providing the essential hardware and ecosystem needed to build, scale, and run AI models.

• Power Plants (Nvidia’s GPUs): Nvidia’s high-performance GPUs (graphics processing units) are like power plants, generating the raw processing power needed to fuel AI models. They are the engines behind AI’s rapid growth, powering models for Nvidia’s major customers.

• Power Lines (Nvidia’s Ecosystem: CUDA, Software, and Data Centers): Power lines carry electricity from power plants to end-users, and Nvidia’s ecosystem - including CUDA, its software libraries, and AI-powered data centers - transmits the power of AI to applications and businesses.

• End Users (Companies and Industries Adopting AI): Just as homes, businesses, and factories depend on electricity, AI’s end users - companies across nearly every industry - rely on Nvidia’s infrastructure to drive their AI initiatives. Nvidia’s customers are not just tech giants but also industries racing to adopt AI to remain competitive.

Just as the advent of the electrical grid transformed society - making electricity widely accessible and foundational to modern life - Nvidia’s infrastructure is now enabling AI to achieve real, transformative impact across industries. The vast network Nvidia has built, with its powerful GPUs and deeply integrated ecosystem, opens up so many possibilities, allowing AI to accelerate at a pace that would have been unimaginable just a few years ago.

In this era of AI-driven transformation, Nvidia is providing the “electricity” for everyone, laying the foundation that powers the future of AI and, ultimately, the next stage of technological revolution.

The NVDA Moat

The best businesses have a competitive moat that sets them apart from the competition. Nvidia’s moat is quite, quite wide in this rapidly evolving, advancing AI landscape. In a space where innovation is constant, Nvidia’s combination of advanced technology, strategic positioning, and market leadership has created barriers akin to owning the power grid - its competitors will likely struggle to overcome them for years and years. At the heart of this moat is Nvidia’s dominance in two critical areas that drive AI: its cutting-edge hardware and its deeply integrated ecosystem.

1. The Hardware: GPUs (aka the power plants)

Let’s get some technical jargon out of the way. For our purposes, there are essentially two kinds of chips: CPUs and GPUs:

• CPUs - often considered the “brain” of your computer // are like a single-lane road where only one car (task) can go at a time

• GPUs - Nvidia’s specialty // excel at rendering graphics, Bitcoin mining, and training/modeling AI // designed for parallel processing // are like a multi-lane highway where thousands of cars can move simultaneously

AI relies on the ability to process vast amounts of data all at once - something best done by GPUs, and Nvidia builds the best in the world right now. These GPUs, like Nvidia’s A100, H100, and the upcoming Blackwell chips, power the AI systems of tech giants like Google, Microsoft, Tesla, Meta, and Amazon (among many others), making Nvidia the go-to provider for scalable AI solutions.

This is especially true for hyperscalers, the massive cloud providers like Amazon Web Services (AWS), Microsoft Azure, and Google Cloud. These companies are the "power plants" of the digital age, running enormous data centers that process AI workloads at unprecedented scales. Hyperscalers represent some of Nvidia’s largest and most important customers, relying on Nvidia GPUs to enable their AI services for industries ranging from healthcare to retail to finance. The ongoing growth of these cloud giants and their increasing investments in AI further reinforce Nvidia’s dominance in this space.

And fortunately for Nvidia, these aren’t one-time purchases. AI is advancing so fast that companies continuously need to upgrade their hardware to keep up with the increasing demand for more power and speed. Nvidia's regular release of newer, more advanced chips ensures that their customers are tied to its hardware pipeline - connected to Nvidia’s power grid - creating a valuable recurring revenue stream.

I won’t go deep on them today, but Nvidia’s competitors here are companies like AMD and Intel, who also produce GPUs, but at least so far, they are currently inferior and also lack Nvidia’s comprehensive ecosystem.

2. The Ecosystem: CUDA and Beyond (aka the power lines)

Nvidia’s ecosystem is the second, equally critical, pillar of its moat. At the center is CUDA, Nvidia’s proprietary platform that allows AI developers to communicate with and harness the power of Nvidia’s GPU hardware. Once companies build their AI systems around CUDA, switching to a different provider becomes costly, time-consuming, and complex. CUDA essentially locks Nvidia customers into its ecosystem, creating a highly sticky relationship that deepens Nvidia’s hold on the market as more of its GPUs become integrated across the companies investing in AI. It's extremely difficult for competitors to disrupt this advantage.

Imagine building an entire city with a specific power grid. Switching to another grid would require replacing all the infrastructure, causing massive disruption. This is the same challenge companies face when they rely on Nvidia’s CUDA ecosystem. Once integrated, moving away threatens their entire AI operation, requiring extensive code rewrites and workflow adjustments.

And Nvidia’s ecosystem extends well beyond CUDA. Its software libraries, development tools, and AI frameworks help companies develop, train, and deploy AI models at scale, while Nvidia’s AI-powered data centers provide processing power for computationally heavy tasks. These data centers act as critical hubs for training massive AI models, further embedding Nvidia’s technology in AI infrastructure.

3. Supporting the Moat

Wild as it sounds, Nvidia’s dominance actually extends beyond its hardware and ecosystem. Several key factors further strengthen its moat and reinforce these two primary pillars:

• Developer Community and Network Effects: Millions of developers worldwide are now trained in CUDA and Nvidia’s tools, creating a powerful network effect. As more developers contribute to the ecosystem, the more valuable and entrenched it becomes, raising switching costs for companies.

• Partnerships and Industry Standards: Nvidia’s partnerships with industry leaders like Google, Microsoft, Amazon, Tesla, Meta, Oracle, IBM, and more make its hardware and software industry standards. This preferred adoption among major players strengthens Nvidia’s foothold in AI infrastructure.

• AI Research and Vertical Integration: Nvidia is also deeply involved in AI research through initiatives like Nvidia Research. By pushing the boundaries of what AI can do, Nvidia stays ahead of the curve and influences the future of AI. This vertical integration gives Nvidia a unique advantage, as they’re not just providing tools for AI development - they’re shaping its evolution.

• Cloud and AI Software Platforms: Nvidia has expanded into cloud and AI software platforms, such as Nvidia AI Enterprise, enabling companies to run AI models more efficiently. This presence across both on-premise and cloud-based infrastructure solidifies Nvidia’s role in the AI-as-a-service (AIaaS) market. Here, Nvidia competes with cloud providers like AWS and Google Cloud, who offer some AI processing power but without Nvidia’s specialized hardware integration.

The Moat, summing it up

Nvidia’s moat stands firmly on two critical pillars: its world-class GPUs and its deeply entrenched ecosystem centered around CUDA. These pillars are fortified by a vast developer community, strong partnerships, ongoing AI research, and expanding cloud-based infrastructure. Together, these layers make Nvidia’s moat exceptionally difficult to breach, with some of the biggest companies in the world relying on Nvidia’s hardware and ecosystem for their AI needs. This positions Nvidia to remain at the forefront of the AI revolution for years to come.

Risks to Consider

I think it’s relatively clear that Nvidia has stronghold on the AI landscape. That said, it’s also worth recognizing some potential risks at play: ,

1. Supply Chain Dependency on TSMC: Nvidia relies heavily on TSMC (Taiwan Semiconductor Manufacturing Company) to produce its GPUs. They are the world’s largest semiconductor foundry, but still, any disruption in TSMC’s manufacturing capabilities- whether due to supply chain issues, geopolitical tensions, or production constraints - could affect Nvidia’s ability to meet demand.

2. Growing Competition: Nvidia is the absolute leader GPUs - I’ve heard suggestions that it’s 3-5 years ahead of the competition. But, AMD (and, to a lesser extent, Intel) are racing to catch up, along with many of the big tech companies potentially looking to make their own chips. Additionally, specialized AI hardware startups are emerging, bringing innovation and niche solutions that could challenge Nvidia’s dominance over time.

3. Regulatory and Geopolitical Challenges: As Nvidia expands its footprint in data-intensive and AI-driven applications, it could face regulatory scrutiny, particularly around data privacy and security. Export restrictions, such as recent U.S. restrictions on selling advanced chips to China, could limit Nvidia’s reach in certain markets and impact revenue from those regions.

4. High Valuation and Market Expectations: Nvidia’s stock price reflects high expectations for future growth. If AI adoption or demand for GPUs fails to meet these projections, or if Nvidia misses earnings expectations, the stock could face volatility (which it typically does, anyway, despite the long-term growth).

5. Potential Overreliance on AI Growth: While AI offers huge opportunities, Nvidia is deeply invested in this sector. If AI adoption slows or its growth doesn’t reach anticipated levels, Nvidia could be more impacted than companies with a more diversified business model.

The risks exist. I don’t think they offset the strengths and opportunities, but they’re at least worth consideration as part of a well-rounded view of Nvidia’s investment potential.

Looking Ahead: Nvidia's Long-Term Potential

What matters most for Nvidia isn’t just staying relevant today; it’s their ability to shape the future of AI and stay at the center of it. I’m confident they can continue to do so for the following reasons:

• They’ve investing heavily in next-gen GPUs, like the upcoming Blackwell chips, that promise greater, and greater processing power to fuel the next wave of AI advancements.

• They’re working on their own AI models that rival the latest capabilities of ChatGPT and other LLMs. This is a space they hadn’t even touched previously. Even if they don’t end up a winner here, simply playing in this category is a competitive statement.

• Their full-stack ecosystem and the unmatched stickiness of CUDA only become more valuable with time. Competitors face an uphill battle trying to recreate this network of high-performance hardware, software, and developer support that Nvidia has been building for years.

• Their Omniverse platform is setting them up to be a major player in the metaverse, offering tools for companies to create virtual worlds, simulations, and immersive digital environments. As industries like architecture, entertainment, and manufacturing explore virtual collaboration and simulation, Nvidia’s early moves into this space could open up entirely new revenue streams, further diversifying their tech portfolio.

• I believe in their leadership, especially CEO Jensen Huang. He’s actively involved, forward-thinking, prioritizes engineers, and keeps the business constantly positioned to lead and innovate.

Final Thoughts

AI is still in its early stages, with a long runway for growth ahead. The potential unlock of productivity and value is only just being recognized now. Nvidia stands at the center of this transformation. Their dominance in GPU technology, the durability of their CUDA software ecosystem, and their constant drive for AI innovation place Nvidia in a unique position to lead the AI revolution for years to come. So, I am holding.

As always, I’m open to feedback, input, and discussion. Cheers.

— Alf London

NOTE: This is for educational purposes only and not financial advice. Please consult with a financial professional before making any investment decisions.