NVIDIA is a company well-known to 1990’s gamers. It quickly built a reputation for making top tier computer graphics cards. Much more recently, it has used the heart of its graphics technology to make some exciting headway in the field of artificial intelligence (AI).
In the past two weeks alone, NVIDIA’s stock price is up over 30%. What’s going on?
Exciting Q1 results
The immediate spark has been the company’s Q1 results, which NVIDIA announced on May 9. The chart below shows revenue, gross income, and net income for 2018 Q1 vs 2017 Q4:
Let’s cover a few quick notes:
- NVIDIA has a 2018 fiscal year that runs from February 2017 through January 2018. That’s why the reported year numbers stray so much from the calendar year.
- Gross profit is revenue minus cost of goods sold. Gross margin, a more commonly seen metric, would be gross profit divided by revenue.
- I thought this quarter’s results were supposed to be exciting. What’s so exciting about all the revenue and profit numbers going down?
As often happens with financial analysis, a sequential quarterly comparison can be misleading. Let’s shift our baseline, and compare the first quarter of this year to the first quarter of last year:
Okay, now we’re getting somewhere. We see huge jumps across the board: a 48% increase in revenue, a 54% increase in gross profit, and a 144% increase in net income.
It’s one thing to know the all-in numbers. But it’s more helpful if we can dive a little deeper. The chart below shows NVIDIA’s quarterly revenue, broken into its market platforms. You can think of the market platforms as applications for NVIDIA’s processing technologies.
Now we’re starting to see what’s going on. Gaming is still NVIDIA’s largest market platform. That’s not surprising. That’s how NVIDIA made its name.
Counter-intuitively, it’s NVIDIA’s smaller market platforms that are more consequential for investors. Look at the trend in Datacenter revenue. Year over year, Datacenter revenue has nearly tripled, from $143 million in 2017 Q1 to $409 million in 2018 Q1.
What’s the Datacenter platform? It’s the area where NVIDIA is targeting AI. Long story short, NVIDIA is positioning its graphics processing units (GPUs), and the software that drives them, as an ideal way to perform the heavy lifting of artificial intelligence. And this heavy lifting typically takes place across large numbers of computers packed into datacenters.
If you think AI is the way of the future, you are very, very excited about NVIDIA’s results. Its Datacenter market platform is showing growth far and away above the rest of the company’s offerings. And the sky is the limit.
Here’s what the company said in its press release:
“The AI revolution is moving fast and continuing to accelerate,” said Jensen Huang, founder and chief executive officer of NVIDIA. “NVIDIA’s GPU deep learning platform is the instrument of choice for researchers, internet giants and startups as they invent the future.
“Our Datacenter GPU computing business nearly tripled from last year, as more of the world’s computer scientists engage deep learning. One industry after another is awakening to the power of GPU deep learning and AI, the most important technology force of our time,” he said.
Datacenters are just getting started, in terms of adopting GPUs for AI-intensive tasks. We’re still learning what AI even is, and what problems it’s best able to solve. NVIDIA is poised to capitalize, as researchers better understand the value-generating potential of AI.
What about the dip from Q4?
We still haven’t addressed why the Q1 results were so exciting, given the big walk back in revenue, gross profit, and net income from the prior quarter. Let’s look at the revenue by market platform chart again:
The culprit is gaming. It’s the largest segment. Its revenue declined by $321 million sequentially, whereas NVIDIA’s revenue as a whole only declined by $236 million.
What explains the big dropoff in gaming revenue, from 2017 Q4 to 2018 Q1? The holidays. Its graphics cards are big gifts for gaming aficionados. Its SHIELD TV streaming device is having some success for cord cutters. We need to keep this in mind, when we compare sequential quarterly performance. A big revenue drop off from Q4 to Q1 isn’t so scary, when you keep in mind the once a year gift rush we see in Q4.
Is NVIDIA’s stock price too high?
Let’s do a little arithmetic, to try to put NVIDIA’s valuation in perspective. Keep in mind, as of the time I wrote this post, NVIDIA has a market capitalization of $84 billion.
There’s no bulletproof way to answer this question. Here’s one thing we can do, as a back of the envelope calculation. Let’s assume NVIDIA’s net income grows at 10% annually. How many years of that kind of growth would it take to deliver $84 billion in profit, assuming a 3% discount rate?
The answer is 22. NVIDIA’s net income would have to grow at 10% every year through 2039, to deliver $84 billion in cumulative (2017) profit dollars. Yikes.
Let’s try to add a little context. Apple has a market capitalization of $806 billion. How many years of 10% net income growth would it take Apple to deliver $806 billion in (2017) profit dollars?
The answer is 11. Apple is about half as expensive as NVIDIA. But remember what the market is telling us. It’s saying that NVIDIA is much more likely than Apple to sustain its profitability growth. That’s why you pay a premium for NVIDIA stock. You’re presumably paying for the added profit dollars you’ll likely get in the future.
Of course there’s an even quicker way to arrive at this conclusion. Look at the price to earnings (P/E) ratio for these two companies. Apple’s P/E is 18. NVIDIA’s is 47. This is a direct measure of the premium you pay for NVIDIA’s current earnings dollars. Even though every earnings dollar is the same, regardless of the company it comes from, NVIDIA’s dollars today come with an increased likelihood for even more such dollars in the future.
No one can predict the future of AI
When it comes to investing, all the financial models in the world can only get you so far. No one can predict the future. And because no one can predict the future, no one knows whether NVIDIA is priced appropriately or not.
The AI movement is particularly sensitive to our predictive failures. No one can say (a) whether an AI boom will ever materialize, (b) if a boom does arrive, when it will get here, and (c) if there is a boom, how long it will last. No one knows.
That’s how stock investing turns into a form of gambling. If you think AI will change the world, and you think it’ll happen in the next decade, and you think NVIDIA will capture a good fraction of the value, then NVIDIA’s stock is cheap. You’re buying General Electric before electricity became mainstream. You’re buying Ford as the Model T is rolling out of the manufacturing facility. You’re buying Apple in 1997, as Steve Jobs is setting his turnaround effort into motion.
But what if this much hyped AI boom never materializes? Or what if it appears in a different form? What if GPU-filled datacenters aren’t the backbone of our eventual AI-focused future? What if NVIDIA’s competitive advantage washes away, before the market matures?
You can never know the answers to these questions. That’s why there’s no right answer in investing, unless you’re the jerk asking questions in hindsight. You have to make some judgment calls. You have to place a bet.
In the past 52 weeks, NVIDIA’s stock price has more than tripled. Clearly, some investors are betting heavily on NVIDIA. These optimists see the emerging trend in datacenter performance, and are wagering heavily that the trend will continue going forward.
Two takeaways from NVIDIA’s stock price explosion
One idea we need to keep in mind is our inability to predict the future. Different people will have different beliefs, and these beliefs will manifest in wildly different valuations, for NVIDIA and other companies.
A second idea we need to keep in mind is the importance of today’s small product categories. For NVIDIA, gaming is king. It contributes over half of total revenue. The Datacenter market platform is relatively small. It contributes only a fifth of total revenue.
But Datacenter is where NVIDIA and its investors expect the future to be. It’s a good example of how focusing on a company’s largest product or service can distract you from what’s most important for the future. Gaming is a nice business, but it doesn’t have the potential of AI. It will keep growing, but it’s only one market. AI can appear all over the place.
When you value a company, you look to the future. Sure, it matters what assets and liabilities the company has today. But what’s more important is how the company will use those assets, now and in the future. Do the assets offer a sustainable competitive advantage in a booming market? Or will they simply allow the company to grind out a bit of cash each year, with little hope of meaningful growth?
These are the kinds of questions we have to consider when analyzing the financial trajectory of a company. All the discounted cash flow models in the world can’t really tell us, for instance, if NVIDIA’s stock price is reasonable or not. Our answer to that question tells us much more about ourselves, and our worldviews, than it does about our modeling capabilities.