Nvidia - AI Market Growth 2016-2025

https://www.statista.com/statistics/607716/worldwide-artific…

I do not know if all of these markets are 100% opportunities for Nvidia, but Nvidia is the enabling technology for at least every single product being projected here.

The market for AI is expected to grow from $7.345 billion 2018 to $89.847 in 2025. 2025 is really just around the corner. Amazing how time moves so much faster the older you get. I hope it does not geometrically accelerate as each year goes by. From this projection, and I doubt many would quibble that the growth will be very large as AI is the largest technological change since the silicon chips that it requires to power AI itself, is 12x growth from this year to the end of 2025. Most estimates like this underestimate some trends, but sometime some trends take longer to take place than anticipated.

In this projection I believe the most speculative is the autonomous automobile market. We are still a bit away from full level 5 automated passenger cars. However the AI market does not need that to take off as autonomous machines are lower hanging fruit. Things like drones, tractors, trucks, construction equipment, and the like.

At present nearly all of this requires Nvidia GPUs and software. AMD does not have anything but an asterisk, and Intel has CPUs that are ancillary to the GPUs in learning, but can stand alone or with GPUs in inference, along with FPGAs, but those are all growing much slower than the GPU aspect of AI. Nvidia presently has ROE > 50%, meaning this company returns more than 50% for each dollar it invests in its business. An absurdly large number. There are around 47 start ups trying to become the “Nvidia killer” but as far as I can tell they are mostly like Google’s TPU chips that Nvidia has managed to largely outperform and the TPU has not made a dent in Nvidia’s business.

For a company with an effective world wide monopoly in multiple enormous and high growth markets that may exceed the TAM for any market in the world, their current PE of $24 TTM (after Q4 ends with this quarter), and probably 16x PE for next year…

I will leave it to others to comment. Low valuation is not necessarily a positive criteria, but at some point something has to give. Currently I think the market has equated Nvidia with AMD and general semi’s. Reality is most likely something else. Anyone else agree or disagree?

Tinker

9 Likes

I will leave it to others to comment. Low valuation is not necessarily a positive criteria, but at some point something has to give. Currently I think the market has equated Nvidia with AMD and general semi’s. Reality is most likely something else. Anyone else agree or disagree?

Yes, Tinker, our thinking on NVDA, growth in its markets, and its REALLY low valuation are aligned. This week, I doubled my position in NVDA and I may well increase it further. I sold most of my PSTG and even some NTNX to put into NVDA. The reason is that I see NVDA to very a much low risk investment. Why? Because it has financial stability, has essentially no competition, is disrupting many huge markets while it disrupts itself (disrupting itself at a fast pace prevents some other technology from getting a foothold and disrupting NVDA). Yes, I NVDA is getting taken down…is it due to AMD or the semis or the general market selloff? I don’t know why but that doesn’t matter; my decision was clear. Technology is accelerating faster and faster every year, and to a large extent NVDA is a big driver for this and NVDA will collect the dollars as it changes the world. With the drop in price and all of the new products, improvements, and new huge target markets revealed in the last year, I believe that the probability that NVDA delivers 10x investment return in the next 5-7 years (an important investing criterion for me) has gone up a lot. Therefore, I put and may continue to put my investment dollars there.

Chris

9 Likes

I believe that the probability that NVDA delivers 10x investment return in the next 5-7 years (an important investing criterion for me) has gone up a lot. Therefore, I put and may continue to put my investment dollars there.

Chris:

I love your posts but that is impossible. To be a 10 multiple from here would put its market cap at $1.2 TRILLION (more valuable than any stock today including AMZN or AAPL)…no possible way that could happen in 5-7 years.

We have been tracking the revenue growth for its main sources (gaming, data, autonomous) and what we see is here:

https://discussion.fool.com/nvda-npi-review-33064502.aspx?sort=w…

That data (and I know you keep the best data so this is not a competition :wink: suggests that “data” is likely to reach parity with gaming in the next 2-3 years. Autonomous is way behind that adoption has many more hurdles that people realize not the least of which is governmental regs.

So IMO, either you have to lower expectations to say a 3-5 multiple on 5-7 years…or double/triple the time frame.

IMO, a better way to consider NVDA would seem to be a lower risk adjusted return. But let’s also consider that while its forward PE may be 22.7 or so…between 2011 and 2014, it brandished similar PE’s…so this is not an anomaly.

I know this is a different animal now with data and AI and autonomous…and not just a gaming stock…but 10 multiple seems a bit optimistic.

I know Saul’s board hates these discussions so if you want to discuss more, we can chat at the NPI. I wont respond further here.

Thanks for all you do Chris…is a pleasure to read your posts.

7 Likes

I know Saul’s board hates these discussions so if you want to discuss more, we can chat at the NPI.

Hi Duma, your discussion seems like a sensible discussion of a high growth stock and I don’t see any reason that it doesn’t fit on our board.

Saul

5 Likes

Thanks for the nice post. And thanks for all of the analysis that you do (Tinker too). The posts are very helpful to me.

To be a 10 multiple from here would put its market cap at $1.2 TRILLION (more valuable than any stock today including AMZN or AAPL)…no possible way that could happen in 5-7 years.

We have been tracking the revenue growth for its main sources (gaming, data, autonomous) and what we see is here:

First, let me say that I didn’t say that NVDA stock will return 10x in 5-7 years. I said it can. I still think it can. Yes, in 7 years it will be near the end of 2025. I evaluate my holdings continuously so if it has increased 5x in 5 years then I would be looking for 10x in 5-7 years from then and might well decrease or sell for faster growing opportunities.

During 2018, NVDA announced some pretty compelling products and services that are not listed in your table. For instance, CLARA is the medical imaging AI enhancements. There was a good demonstration of it during the GTC Japan keynote presentation and it was first announced at the main GTC event in the Spring. There are 3 million legacy imaging devices installed worldwide. CLARA will enable these more primitive systems to produce images (when enhanced by CLARA) that are 3D and far superior. How will NVDA charge for this? Will sell per instrument subscriptions that enable them to connect to NVDA’s server over the internet to enhance the images produced by the instrument or will they charge a per image enhancement fee? I don’t know but I do know that NVDA is good at monetizing their technologies and there is very clear value to medical centers that have legacy instruments (cost to enhance compared to cost to replace the legacy instrument with more advanced instrument…why replace when the can get the benefit of better imaging at a lower cost).

Automotive is another market in your table but in aggregate non-passenger vehicle market opportunities for automotive (as defined as anything that moves and can benefit from independent decision making (AKA thinking) is much bigger than the passenger vehicle market opportunity.

What other market opportunities will be revealed in the next 7 years? Surely there will be at least several more $20B+ market opportunities that NVDA will target. As NVDA lowers the cost, miniaturizes, lowers power requirements, increases the reasoning power, etc., the solutions offered by NVDA will proliferate into the markets of the world and create new markets that do not exist today. NVDA will collect much of the financial rewards of this. This is already unfolding before our eyes. Will we autonomously mine asteroids (how many trillions would that be worth)? Will we harvest and tend to all crops autonomously? As Kurzweil has written, the accelerating pace of technology is difficult to observe and image when looking backward. Technology is accelerating and it is deflationary. As a side note: let’s see what happens with inflation and the Fed.

Perhaps a 10x return in 7 years is optimistic, but it is not impossible as you wrote. When one thinks about NVDA perhaps it is better to view it through the lens of disruption through efficiency improvement (which forces adoption by the necessity of economic survival) and enablement of creating new markets. By looking at only the markets that NVDA is in and has already announced that it will target, one misses a huge piece of the opportunity that NVDA presents. There will be new markets (in the next 7 years) that NVDA enters that are not currently considered in calculating NVDA’s value. However, I would agree that NVDA’s path to a 2-3x return in 2 years seems much more clear than a 10x return in 7 years. But I will make another prediction: NVDA will very likely surpass INTC in market cap during 2019. Currently, INTC is at $210B while NVDA is at $120B.

Chris

20 Likes

However, I would agree that NVDA’s path to a 2-3x return in 2 years seems much more clear than a 10x return in 7 years. But I will make another prediction: NVDA will very likely surpass INTC in market cap during 2019. Currently, INTC is at $210B while NVDA is at $120B.

I would agree with that as well. IMO, NVDA is a long term hold but as I said I view that as a risk adjusted return.

IMO, we should hold stocks of companies that the world could not imagine being without or would suffer greatly. Without NVDA, progress in data, AI, etc…these industry initiatives would suffer and in some cases come to a halt.

Think of all the companies whose stock you hold…how often can you say that about each of them…the world would be worse without them…no clear readily accessible substitute with equal or better promise/value.

That is a high bar indeed and that is why I agree it is a great risk adjusted return stock.

But there are the “race horses” that will no doubt kick NVDA’s returns…but they are far more volatile and unpredictable.

Sometimes, IMO, our portfolio’s need a little less “hair on fire”, and a little more stable but true. That is how I see NVDA. I own it and will for some time.

11 Likes

One thing overlooked as it is said, sensibly, that NVDA was once routinely valued at this multiple (yes, I know Duma knows Nvidia is no longer the same company, not even close, but I do not think people realize just how far removed NVidia is from 2014) is that to date we have seen AI used for some basic functions that run through the core data centers. Starting in 2019 and really ramping in the next 2-3 years medical will take off (I linked to an article (at least in my own brain) citing the new unanimous opinion of medical centers beyond just imaging adopting AI, and medical is an enormous market), and autonomous machines are going to start rolling out enough to be noticed in at least some places, with things geometrically growing in 2020 and 2021…

At the current valuation NVDA is valued as if this were 2014, as if it were AMD, as if it was just a slightly faster growing Intel, say of 2005 and none of it is even close to the truth. Nvidia at $120 billion marketcap is about equivalent to (I don’t have any real numbers to estimate this but I will anyways) around a $5 billion Nutanix (but with no effective competition), or a $4 billion Arista, or the like. Nvidia is like a very powerful, extremely well funded, revolutionary start-up, with an ROE (Return on Equity) of 54.8%

https://www.macrotrends.net/stocks/charts/NVDA/nvidia/roe

Can any other company in the world claim something even close? Every dollar invested is returning more than 50% for shareholders! And yet Nvidia has just put a toe into the AI waters as of yet. The real harvest just moves beyond the asterisk stage starting by late 2019, if not sooner. I believe Google is rolling out its autonomous car service in limited release within the next 3 to 12 months. They are already in beta service and playing with different pricing structures. Google is the only autonomous vehicle competitor that Nvidia has, and like with the TPU, Google plans to launch as a service, not as a solutions and systems provider.

I think saying that Nvidia was valued like this in 2014 is like saying that Lincoln may have some impact in history, but he is just a president after all however.

Numbers will speak for themselves, and yet we are so focused on Q3, the dangerous Q3. This could of course, due to multiple issues, such as mid quarter release of their new generation of chips, Fab down for a few weeks in August, who knows what turn into a “miss” that Cramer predicts. That is quite myopic indeed.

As most know Nvidia was part of my triad. But, for similar reasons that SAul sold it off, so did I. But at this point, back down to $120 billion marketcap, mid-teen one year forward P/E even if earnings growth slowed to just 20%, and the stock totally crashed, more than the market at large, based on nothingthat is specific at all to Nvidia…there is no question it is worth another look. Just the multiple compression being removed again is worth a double in the next year.

I believe there is much fear about Q3, and if Texas Instruments is in anyway a bell weather for all chip stocks. Q3, I can see reason for anxiety (albeit Intel’s strength in their earnings were expressed as strength in gaming and strength in big data in the data center (ie, exactly where Nvidia presently dominates - AMD has little presence in the data center and their gaming chips just got antiquated with Nvidia’s new generation release), so one can go either way. Which creates an unknown and a risk.

Against this is when/if all the rest starts to roll out starting from this very tiny base valuation (relatively speaking). The problem is, there is no “if” but just when. And Nvidia is now valued almost as if there is neither an “if” or a “when” at all, but rather Nvidia is again a dominant growth engine for gaming as it always has been. Nice and good steady business to be in, but highly cyclical.

Tinker

12 Likes

NVDA will very likely surpass INTC in market cap during 2019. Currently, INTC is at $210B while NVDA is at $120B.

Or by 2020 imo, at least.
I believe NVDA has an fairly attainable path to 50% gain in next 12 months from here.

In 2020, AV becomes more and more material and DC has probably become biggest BU. The market realizes nvidia isnt just a momentum favorite ala 2016, but is akin to Intel in the 1990s and early 2000s = everywhere.

Dreamer

2 Likes

IMO, we should hold stocks of companies that the world could not imagine being without or would suffer greatly.

I like that a lot Duma. Thanks for putting into words!

Saul

4 Likes

I’m going to start re-reading Philip K. Dick in order to figure out what the future AI world will be like… :slight_smile: