The Six Five team discusses NVIDIA Q1 Earnings.
If you are interested in watching the full episode you can check it out here.
Disclaimer: The Six Five Webcast is for information and entertainment purposes only. Over the course of this webcast, we may talk about companies that are publicly traded and we may even reference that fact and their equity share price, but please do not take anything that we say as a recommendation about what you should do with your investment dollars. We are not investment advisors and we ask that you do not treat us as such.
Transcript:
Daniel Newman: Let’s go on to Nvidia because this is the topic du jour, topic du jour week. It’s you.
Patrick Moorhead: I know. I know. Just when you think you’ve got the topic du jour, a new one comes along. Let me hit the lead upfront. Nvidia upped their forecast for the next quarter by four billion dollars.
Daniel Newman: With a B.
Patrick Moorhead: With a B, which increased their market cap more than their competitors are worth. That got a little bit of attention. The side story is they beat revenue by 10%. They beat EPS by about 20%, but it was that upward guide that got literally everybody talking out there. Forget their gaming market is down precipitously, right? Forget about Professional Workstation being down. Auto up, good. Nobody cared. Everybody cared about what generative AI meant. Let me try to break it down for you and hopefully it’s pretty obvious if you’ve listened to the pod is that you need something to first of all train these large models, and these large models can be text, which is an LLM. It can be based on images like stable diffusion, and it can be also videos, which you can pull down a model from a hugging face and have fun.
You probably saw the Joe Rogan fake video podcast that was cranked out by essentially putting in a text script ingesting hours and hours of his video, and boy did it look real, right? So just back at the envelope, think of these types of models, these large models taking 10x the capabilities, the GPU capabilities to train. The reason I say GPU and not ASIC is that, quite frankly, Nvidia is the standard right now for training as it’s just so flexible. You want to do 27 different types of models, GPUs are your bet.
Now, Nvidia does put little tiny ASICs in there to accelerate things like transform models and things like that, but in the end, it’s just the raw horsepower of that GPU and the flexibility that keeps that going. I was on Street Signs last night, CNBC Street Signs, and a lot of debate on, how long does this go? Let’s just say for a second the training of these new generative AI models are only with the largest of companies. I can see that. The inference though, when you then take those models and you run those models and you blend those against your private data, that’s going to be an opportunity that I think is going to be there for everybody, whether it’s Intel, AMD and, of course, Nvidia.
I think Nvidia’s training probably we’re looking at probably a three quarter phenomena. I’m not going to say before we’re done, but before we see that getting back to what I would call normal growth. There’s other companies then as we’ll talk about like Marvell who are seeing the benefit of that, if nothing else, to string along these GPUs. So again, nothing else mattered other than that data center number and that data center number that was driven by generative AI training this quarter.
I think that the way that I look at this is give Nvidia time for these AAA games to come out, and also these year over year tough comparisons on the gaming side. If I look at gaming, last year they were all in the three billion range. In fact, a year ago, that gaming number was 3.6 billion. Now, it’s 2.2 billion. NFTs flamed out, which even though Nvidia did its best to segment gaming cards just for that, they neutered gaming cards for crypto. What happened is given that there were so many different types of NFT and so many different types of crypto, the GPU just became, even if it was neutered for let’s say Ethereum, all the other versions of it became the ultimate device. So what we’re seeing again is a recalibration of Nvidia’s gaming number. There we go.
Daniel Newman: All right. So there you have it. Metaverse is on pause, but AI is all the rage. Now in the future, I do see there’s going to be a coming together, by the way, of these two things. People aren’t fully appreciating just yet, but there will be, and this will, of course, fuel a bit of a web3 revolution too because I do believe we’re going to need to have some way to create tokenization of content and assets because as fake becomes easier to create, how do we separate real from what’s not real? That’s going to be increasingly difficult and I think that will be an opportunity for some of these different applications.
Look, killer quarter. Don’t mistake the two years ago when I said Nvidia would be the next trillion dollar company, when I wrote that on Market Watch. It wasn’t a mistake. It was obvious. AI had the biggest upside. It’s going to revolutionize industries and it will be industry five will be the industry revolution fueled by AI. Last summer when I was beaten up by the whole Twitter universe and for going on CNBC and saying, “Hold on.” I picked two. It was at their absolute lows. It was Nvidia and Microsoft. Nvidia had fallen like 75% and I said something along the lines of, “This is going to be the opportunity of a lifetime.”
Pat, I don’t know what I would do if I didn’t spend at least 10% of my day doing victory laps, but I do like when I get it right. So what we’ve gotten right here is that every company is buying every single piece of silicon they can get their hands on in order to be ready for the inference boom. So right now, you leaned into training, Pat. This is what it’s about. Right now, everybody understands that there’s a huge requirement to be able to train all their data, and that means tons and tons of hardware GPUs will be required. You can probably glean that the largest swath of this massive four billion dollar guide up is hyperscale, meaning these folks are going to be building out the infrastructure to be able to support the next wave of infrastructure platforms and software that are going to be built on top of things like Azure and AWS.
Of course, we will see different flavors and varieties, but there is no end-to-end system right there that’s in the market that’s ready right now the way that Nvidia is. So Nvidia has a huge leap and a huge gain, and this has been on the backs of … AMD has not been able to get the whole stack in the software right. Intel’s lagging behind. There are some ASICs. Google’s built their TPUs, and Meta’s building some proprietary AI capabilities through hardware. We know companies, Pat, you and I have tracked Groq and SambaNova and the companies that are building some specialized ASICs and accelerators. There’s some efficiencies to be created there, but Nvidia has the color on this particular market right now.
The interesting thing has been the second wave and we’ll talk about Marvell and Broadcom and some of the others in a future segment of all the other companies that are getting a similar boost to what Nvidia got. Look, Nvidia added … Pat, I don’t know if you said this. I was trying to find the stat while you were talking. So something along the lines of 260 billion dollars in market cap in 24 hours. I just want to pause, just take a breath. Okay. Let’s add up what that is. That’s Dell, Intel, and AMD. I think that’s the three of them together. That is the entire market cap of-
Patrick Moorhead: Intel is 120. Actually, AMD is 204.
Daniel Newman: Okay. AMD went up big though after this all happened though. The day of, AMD was 160 at that time because I looked it up. IBM’s 120 or 110, but my point is it’s bigger than Qualcomm and Intel together right now. So the fact of the matter is that’s not the … and it breached a trillion dollars in a day. So the fact of the matter is is that we’re just seeing the beginning of what this is going to be. I’m pretty sure that Nvidia will have a hard time keeping up with its demand for many, many quarters to come on the enterprise side to try to support the build out both on the enterprise and cloud scale.
Interestingly enough, it secretly covered for the fact that gaming is soft, Proviz is soft, Omniverse is soft, crypto is soft, automotive is soft. So it’s interesting because this one particular category right now is taking all the weight. Try, nobody else cared. The only other thing I’m going to say here is I’m just going to seed this idea. Nvidia, I have to imagine at some point is going to be under some interesting pressure with its end-to-end closed ecosystem when everything gets built on it. It seems like it’s rife for some regulatory oversight. I’m not sure. I just mean we’re spending months trying to pass something like a Broadcom VMware deal that has almost no definitive antitrust or anti-competitive attributes to it and certainly nothing that couldn’t be handled with a couple of concessions.
You now have a closed software hardware and, by the way, maybe some bundling things going on with the fact that I’m pretty sure, from what I understand, Pat, you can’t buy an A100 just the silicon anymore. You’re buying systems. Everybody’s saying you have to buy the whole systems top to bottom now. So I’m just saying it’ll be interesting to see if that gets explored.
Patrick Moorhead: Making 70 points on that. Something has to give, Daniel, and that could be regulatory, it could be a big competitive move. As we discussed on the show too, and DGX Cloud, the way that that’s operating where it’s more the gaming model where Nvidia creates all the value and then it’s distributed to others. Something’s going to change and it could be a combination of all three, but Nvidia does need to watch themselves in how they move forward, particularly given the regulatory environment.
Daniel Newman: Well, Pat, I think it’ll be an interesting thing to watch and explore as they’re trying to regulate the large language model. The only way these things are being trained and built is in Nvidia’s sandbox. I mean that’s where this is all happening right now. So like I said, you cannot easily move any of these workloads off the Nvidia platform. At this point, it’s still very difficult. Will it get looked at? I don’t know. I think our regulatory bodies don’t seem to pay attention to the most obvious antitrust issues. So I wouldn’t be surprised if this one floats under, but I guess I’m saying it was an amazing quarter for Nvidia, but if you’re looking for maybe something to be cautious about, it’s the fact that they are the only game in town right now.
Patrick Moorhead: They’re the only game in town for training. That is for production level highest performance training.
Author Information
Daniel is the CEO of The Futurum Group. Living his life at the intersection of people and technology, Daniel works with the world’s largest technology brands exploring Digital Transformation and how it is influencing the enterprise.
From the leading edge of AI to global technology policy, Daniel makes the connections between business, people and tech that are required for companies to benefit most from their technology investments. Daniel is a top 5 globally ranked industry analyst and his ideas are regularly cited or shared in television appearances by CNBC, Bloomberg, Wall Street Journal and hundreds of other sites around the world.
A 7x Best-Selling Author including his most recent book “Human/Machine.” Daniel is also a Forbes and MarketWatch (Dow Jones) contributor.
An MBA and Former Graduate Adjunct Faculty, Daniel is an Austin Texas transplant after 40 years in Chicago. His speaking takes him around the world each year as he shares his vision of the role technology will play in our future.
