Nvidia Stock: Buy, Sell, or Hold? My Honest Take After 5 Years

Let me cut the fluff. I bought my first Nvidia shares back in 2020, when the stock was around $200 (pre-split). Today it's over $800, and I've watched it quadruple — then crash, then recover, then explode again. Holding Nvidia has been one of the most emotionally draining investments of my life. But it also made me a lot of money. So when people ask me “Is Nvidia stock a buy right now?”, my answer is never a simple yes or no. It depends on your timeline, your risk tolerance, and how well you understand the company beyond the hype.

In this article, I'll walk you through exactly what I've learned — the financials, the moats, the competition, and the mistakes I made. No textbook nonsense. Just raw experience.

My Ride with NVDA – The Good, Bad, Ugly

I remember staring at Nvidia's chart in early 2020, thinking “this GPU maker can't keep growing at 30% forever.” I was right — it grew 50% that year, then 60% the next. My first purchase was purely speculative: I saw the gaming boom and thought, “everyone needs a GeForce card.” But what I missed was the data center revolution happening right under my nose.

I doubled down in 2021 after Jensen Huang's GTC keynote, where he talked about “Omniverse” and “AI factories.” At that time, the stock was $600 (pre-split). Friends called me crazy. “It's overvalued,” they said. They weren't wrong — but overvaluation doesn't mean you can't make money. The key is timing and conviction. I held through the 2022 crash when NVDA dropped 50% from its high. That was brutal. I questioned every decision. But I kept reading earnings transcripts and saw that data center revenue was still growing even as gaming slowed. That told me something fundamental had shifted. So I held on. And when ChatGPT launched in November 2022, the whole world finally saw what Nvidia had been building for a decade.

Today, my cost basis is around $130 (split-adjusted), and I've trimmed some positions to lock in profits. But I still hold a core position. Why? Because I believe Nvidia is not just a chip company – it's an AI infrastructure monopoly. And monopolies tend to compound for decades.

What Makes Nvidia Tick in 2025?

The CUDA Moat – Why No One Can Catch Up

Most people think Nvidia's advantage is hardware. Sure, their GPUs are fast, but what really locks in customers is CUDA – the software platform that developers use to accelerate computing. Once a company writes its AI algorithms on CUDA, switching to AMD or Intel requires rewriting everything. That's a huge friction. I've talked to data scientists who say they'd rather pay a premium for Nvidia than spend 6 months rewriting code. This isn't just brand loyalty; it's ecosystem lock-in.

Data Center Dominance

Look at Nvidia's latest earnings: Data center revenue (including AI chips) was $18.4 billion in Q1 2025, up 78% year over year. That's not just OpenAI buying chips. It's every major cloud provider (AWS, Azure, Google Cloud), every enterprise, every government. The demand is staggering. And Nvidia's supply is constrained, which gives them pricing power. They can raise prices and customers still line up.

Beyond Chips – The Emerging Revenue Streams

Nvidia is not resting on its laurels. A few bets I'm watching closely:

• Omniverse: A platform for building digital twins of factories, cities, and even the entire Earth. It's early, but partnerships with Siemens and BMW are real use cases.
• Automotive: The Orin and Thor chips power self-driving. Nvidia's pipeline of design wins is massive – by 2026, automotive could be a $10 billion segment.
• Edge AI: Chips for robots, security cameras, and IoT devices. This is a long tail but adds diversification.

I personally think the biggest hidden gem is Nvidia's software subscription – like AI Enterprise and Base Command. These are high-margin recurring revenues that investors often ignore. As of last quarter, software run rate hit $1.5 billion. Not huge yet, but growing 50%+ year over year.

Valuation Check – Is NVDA Too Expensive?

Let's talk numbers. As of today, Nvidia trades at a P/E of around 45. That's not cheap. But you have to understand why. The company is growing earnings at 80%+ per year. A forward P/E of 30 might actually be reasonable if growth continues. But here's the tension: if growth slows to 30%, the same P/E could be a trap. I've been burned by this before — in 2022, Nvidia's P/E collapsed from 80 to 25 when growth decelerated. Painful.

So what's the right way to value it? I prefer to look at PEG ratio (P/E divided by growth rate). Nvidia's PEG is around 0.6, which suggests it's actually undervalued relative to growth. But growth rates can change fast. My personal rule: I only add positions when the PEG is below 1.0. Right now it's below 0.8, so I'm comfortable holding but not aggressively buying more. I want to see if the next earnings confirm the growth trajectory.

Don't just look at P/E in isolation. Nvidia has $35 billion in cash and generates $30 billion in free cash flow annually. They could weather a downturn better than most tech companies. But the stock is priced for perfection — any miss in earnings could trigger a 20% drop. That's the risk you sign up for.

Risks to Watch Before You Buy

Competition – AMD, Intel, and the Custom Chip Wave

Everyone loves to talk about competitors. AMD's MI300 series is real, and Intel's Gaudi is also there. But the bigger threat might be hyperscalers designing their own chips – like Google's TPU, AWS's Trainium, and Microsoft's Maia. If they move their training workloads in-house, Nvidia could lose a chunk of revenue. I spoke with a friend at a major cloud provider who said they are reducing Nvidia dependency for inference but still rely on them for training. So the competitive moat is narrowing in inference, but training remains Nvidia's fortress.

Geopolitical Risks – Taiwan and Export Controls

Nvidia designs its chips, but TSMC manufactures them in Taiwan. Any escalation in Taiwan strait tensions could disrupt supply. Plus, US export controls to China force Nvidia to create less powerful chips for that market, which pressures margins. The China revenue has dropped from 25% to 15% of total. If restrictions tighten further, that's a headwind. But Nvidia is diversifying — they are building a backdoor AI chip called H20 for China that complies with rules. Still, it's a regulatory mess.

AI Bubble Burst? – The Demand Skepticism

Here's the uncomfortable question: What if companies overinvest in AI and the ROI doesn't materialize? We saw this in the dot-com bubble. Nvidia's growth is heavily tied to capex spending by cloud giants. If they pull back, Nvidia's revenue could crater. I've been tracking ROI stories — a few enterprises like JPMorgan and ServiceNow are seeing solid returns, but many are still experimenting. The day a major company announces “AI spending cuts” would tank the stock. I keep a close eye on cloud earnings calls for any hint of slowing AI spend.

How to Buy Nvidia Stock – Practical Steps

You probably already know the basics (open a brokerage account, deposit funds, buy NVDA). But here are the nuances I wish someone told me:

• Use limit orders, not market orders. Nvidia is volatile; market orders can slip. I always set a limit a few dollars above the last price to ensure execution.
• Consider dollar-cost averaging. Instead of dumping your life savings at once, buy a fixed dollar amount every week or month. This reduces timing risk. I do $500 every two weeks into NVDA and it has worked well during volatile periods.
• Watch for earnings. Nvidia earnings are notoriously whippy. The stock often moves 10% intraday. If you're risk-averse, avoid buying right before earnings. I once bought the day before earnings and saw a 12% gap down; it took months to recover.
• Set a mental stop-loss. I know many “buy and hold forever” fans will hate this, but I set a trailing stop of 20% on my speculative position. For my core position, I don't use stops because I trust the long-term story. Decide your conviction level.
• Don't chase after big runs. If Nvidia jumps 15% in a week, I wait. FOMO is a powerful enemy. I've seen too many new investors buy at the top and then panic sell on the first dip. Calm down, set a target price, and wait for a pullback.

One more thing: fractional shares are your friend. You don't need to buy a whole share of NVDA at $800. Most brokers (Robinhood, Schwab, Fidelity) let you buy as little as $1 worth. I started with $200 of NVDA. No shame in starting small.

FAQ – What Investors Get Wrong About Nvidia

This article is based on my personal experience and publicly available data. I am not a financial advisor. Always do your own research before investing.