In this episode of The Deal Room, host Speaker A interviews Stephen Barnett, head of corporate finance at AmplifyMe, to dissect Nvidia’s sky‑high valuation and contrast it with Intel’s steep decline. The conversation begins with an overview of the three main valuation approaches used for non‑bank firms—market‑based (enterprise value and market cap), absolute (discounted cash flow and sum‑of‑parts), and relative (trading comps). Barnett stresses that while the mathematics are straightforward, the outputs hinge on layers of assumptions about future growth, margins, and terminal rates, making valuation a “best‑effort crystal‑ball” exercise. He highlights Nvidia’s data‑center segment as the current growth engine, noting its 122% YoY revenue surge and the industry’s 25‑30% GPU market expansion, and explains how a sum‑of‑parts model can assign different multiples to fast‑growing versus slower units. The discussion moves to market multiples, revealing Nvidia’s EV/Revenue at ~50× and EV/EBITDA at ~88× versus peers like AMD (10× and 40‑50×). Analyst target prices range from $80 to $250, illustrating the subjectivity and pressure on equity research to justify high or low ends. Barnett shares his forward model projecting a 35% CAGR through 2029, with multiples compressing to ~14× EV/Revenue and ~23× EV/EBITDA by 2026, suggesting the company may “grow into” its valuation. The conversation then pivots to Intel, presenting a stark contrast: market cap $84 B, EV/Revenue ~2×, revenue down 14%, and gross margin 42% versus Nvidia’s 76%. Historical data shows Intel’s revenue falling from $79 B (2021) to $55 B, while Nvidia’s rose from $17 B to $96 B. Barnett attributes Intel’s fall to strategic inertia and missed shifts in the computing landscape, despite its past dominance. Throughout, the tone remains conversational yet analytical, with a measured, cautiously critical sentiment toward both companies’ valuations and strategic choices.
Chapters
Welcome back to The Deal Room, and following on from our previous episode, if you haven't checked it out, we did a breaking news one about Nvidia's earnings where Piers and I had a bit of a deep dive. In this episode, I am grabbing Stephen Barnett, who's our head of corporate finance at AmplifyMe. I want to unpack the big number, the valuation number that people always band around with Nvidia, but I want to go a little bit deeper. I want to kind of look at that in terms of getting a career insight as to what the people in investment banking look at when they're trying to come up with these valuations, but then also talk strategically. You know, what's so interesting, even with Piers and I talking about earnings, is the earnings outlook and how the company's trying to manage that. So it'd be interesting to bridge that gap of what analysts are looking at, how the company manages that, and also using a company like Intel, who was the Nvidia of many years ago. Um, young people might not recognize that fact, but it indeed was like the poster child of technology and evolution in tech. And now there's a big divergence between these two firms. So what went right? What went wrong? Stephen, I hope you're going to help me unpack this. So how's it going?
Yeah, pretty good. It feels like it's been a while. I think it's been a couple of episodes where we've had guest speakers or we've done something a little bit different. So yeah, how's your summer been?
Great. I mean, yeah, a couple of trips away, and I did manage to hear your conversation with Peter. So shout out to Peter if he does tune into some other episodes. I thought he did a— I thought that was a great conversation to get an insight from bulge bracket to boutique and things like that. So the point I'm making is, my summer's away, Stephen, you're still in my ear.
I'm glad to hear it. I hope I'm in lots of people's ears as they're lounging on the beach or next to the swimming or whatever it might be. But I think it's worth— yeah, it's worth spending time on this Nvidia story because it is, it's the first headline on so many news pieces, and it's something that we need to spend a bit of time trying to unpack. Is it overvalued? Is it undervalued? Is it the top of the bubble? Or is it actually just about now coming into its own in terms of the core financials of the company justifying that massive valuation?
So Piers and I did have a bit of a tussle over that, so I'm really interested. I'm not going to say what we tussled over because I know you haven't listened to our conversation yet, but what I would like to do is then, what's your approach and how would you go about this, this kind of idea of is it fair value and which way would they go next?
Well, it's, it's really interesting to think about this, and I always advise when we're teaching valuation. I always advise, there's the valuation, there's the technical elements of valuation, and we'll talk a little bit about discounted cash flow and comparable company analysis and things like that. There's that element that's financial and mathematical and calculation-based, but then there's also all of the assumptions that go into these technical analyses that are based on based on little more than gut feeling, right? Is, is Nvidia going to continue its barnstorming growth in the context of an ever more exciting AI use case market where the big tech companies spend hundreds of billions of dollars on ever more sophisticated GPUs and Nvidia manages to keep up with the pace? Or is this, you know, is that wave of AI investment GPU investment, data center investment from the big tech companies. Has that just happened and they're gonna retrench over the next couple of years? And obviously if they retrench their AI spending, then Nvidia is going to be the big sufferer from that retrenchment. So again, when you're an investor, this is— when you're an investor, you need to have your necks on the line, right? There is no right answer. And winding that back a little bit, when we talk about valuation, there's no right answer. We are using a whole host of assumptions layered upon assumptions layered upon assumptions so that even if I'm slightly wrong in one of my input assumptions, so my revenue growth rate for the data centers revenue stream, I'm slightly wrong on that, uh, projecting that out 5 years in the future. My valuation could be wildly out. So we need to understand that this is a real rough— this is not a science. Valuation is our best efforts guess at getting a crystal ball out, looking into the future, and sticking our neck on the line and giving it a go.
You mentioned their data centers, and I know that was a key metric that people look at. So where do you begin? With someone like NVIDIA? I'm, I'm assuming there's a standard format of doing this analysis, but then how do you tweak it for the case of NVIDIA?
Yeah, it's, it's really interesting. The standard way of valuing a company that is not a bank— I mean, you've got 3 different classes of valuation that we tend to teach. The first is market-based, which is a function of the market's value that it places on NVIDIA. So that is simple enterprise value and market capitalization. The market capitalization of NVIDIA this morning was $2.8 trillion. The second class is absolute valuation. So these are valuation methodologies that build up the company's value from a foundational set of principles. So one might be a discounted cash flow. So how much are we valuing the future free cash flows that we estimate forecast that Nvidia is going to generate discounted back to the present day using a, using a discount rate. Again, loads and loads of assumptions. Another way that we can do absolute valuation, and this might be more relevant for Nvidia, is you can do what's called a sum of the parts, and you can take out different elements of Nvidia's business. Again, data centers being the blockbuster growth engine at the moment, you can value those different elements of the business separately. And you can therefore create this sum of the parts analysis in order to value the company. So I might value the data centres business at a higher multiple or at a lower discount rate than a different part of the business. And these sum of the parts analyses are, are super, super useful, especially for businesses that are that have got some very fast-growing areas and some very slow-growing areas. Or for conglomerates, you know, you think of something like a Google and you want to, you want to value YouTube slightly different from the way that you'd value its Google Maps and its Gmail and, and its advertising services and things like that. And then the final valuation method is relative valuation, where you look at Nvidia's comparable companies, whether that's AMD or TSMC or Arm, and you say, all right, what are investors willing to pay for these companies' profits, their EBITDA, and what are investors willing to pay for these companies' revenues? Then you use those multiples, those EV over revenue, EV over EBITDA, or price earnings multiples, and you align where you assign those multiples to NVIDIA's earnings to get another valuation method. You put this all together and you create what's called a football field, which is a stacked bar chart, or which is a— sorry, it's just a simple bar chart where you are looking at the valuation ranges for NVIDIA from all of these different methods, picking a kind of rough midpoint. Again, this is not— this is not exact. There's a bit of finger in the air, a rough midpoint and going, yeah, this is kind of what we think. But as, but as that explanation suggests, it's all, it's all very, very subjective. And if you look at analyst estimates, which I know that you do a lot, and when you look at analyst estimates for Nvidia share price expectations over the next year, you know, these are all smart people But they come up with everything from $80 a share— NVIDIA's current share price is $120— to $250 a share. That is the, that is the range of smart people's opinion on this company.
On that point, I've always wondered how much pressure there is as the, the equity research team, or whomever's responsibility it is, to engineer a result that fits at the high or low end of what you've just described.
I mean, yeah, totally. And I think that quite often valuation models are the justification to a very experience-driven gut feel of the future direction of a particular company. We wouldn't probably say it in as many words, but I know just looking at my Nvidia model right now, I know that I can, I can tweak one single assumption on my terminal growth rate in my discounted cash flow, and I can fit whatever valuation that you and the director wants to put on this. And then I can back solve the explanation for this terminal growth rate, and I can say Nvidia is going to, you know, should have a terminal growth rate of 6% because it's at the vanguard of technology and artificial intelligence, and therefore that justifies a $160 per share valuation target. And you're just as right or wrong as anyone else that has made another assumption.
So, so the bit that, um, to fill you in, so where me and Piers clashed a little bit, we did resolve that and land on a middle ground, but was this idea of, for me, my perspective was Anthony Cheung, because he was talking about the deceleration in some of the growth rates. And then we were talking about the comps being ever challenging, just given how phenomenal their earnings beats have been. But I was kind of talking about it more from a normalization, and we were kind of landing in this kind of Nvidia is now going to have to grow into its valuation. But it was almost as if you're trying to envisage what the future looks like because their earnings on the line items, I mean, top line being 122% revenue growth is insane, but their outlook was pretty, pretty weak against top-end Street estimates and some of the other numbers, issues about Blackwell and things like that. I guess my question is there is when you're valuing, for me, he's kind of glass kind of empty, I'm glass half full. I think it's healthy that the numbers are coming down and that the strategically the company, it seemed like, are being quite vague about things like upcoming technology rollouts like Blackwell, what revenues they can expect. So yeah, just wanted to get your thoughts on that, that kind of comparable and then managing a company which is growing insanely fast still, but is coming off that peak.
Yeah, it's, it's really interesting. So I'm just looking at my NVIDIA forward-looking model at the moment, and let's say, so in 2024, the company generated $61 billion of top-line revenue, and it's on course to more than double that, uh, in the full year 2025. We've just had the, the half-year earnings release, um, earlier this week. If I project that the company's compound annual growth rate from 2025 through to 2029, the 5 years, is 35%, which is an average of a bunch of brokers. So starting off at 100% growth in 2025 and then decaying down to about 10% growth by 2029. So still pretty good, but the GPU market growth is, is forecast to be 25 to 30%. So it's not an unrealistic growth expectation. Then if I go down to my valuation metrics, my EV over revenue and my EV over my EBITDA, in FY 2024, my EV revenue is 50 times, which we all, you know, our mouths are open and go, wow, how is the industry, how is the, how are, how is the investment industry paying 50 times revenue for this company? And they're paying 88.2 times EBITDA for this company. These metrics, these valuation multiples, are extremely high relative to their peers. AMD is valued at about 10 times, uh, revenue and about 40, 50 times EBITDA. But then if I look 2 to 3 years in the future and I've applied my growth rates and I've applied my gross margin assumptions and my cost of goods sold assumptions to my FY 2026, assuming that EV of $3 trillion stays about the same. So EV over FY, full year 2026 revenue, is looking like it's going to be about 14 times. And EV, current day EV over 2026 forecast EBITDA profit, is going to be about 23 times. So you can see how the speed of growth of this company, even if it slows down, you grow into your forecast. The problem is the market, the industry, as we, as you would've discussed on your podcast, always want ever more bombastic blockbuster exceeding expectations by X, Y, and Z. What's actually happening is they're fulfilling expectations and they're fulfilling the valuation that's been put on them. and they're kind of playing catch-up. But that doesn't mean that Nvidia should be worth $5 trillion. It should be worth the amount of money that it's worth.
And that's why I think the shares can, can take that 10%. Well, I think they finished down 6.5% the day after their earnings, after dropping a bit more in aftermarket. In the context of the bigger picture, I think it's a good, good airing of the laundry to be a bit more realistic. With what they were trying to achieve in the conference call with analysts, I felt when I listened to it.
Totally, totally. And I mean, from all of the different valuation methods that I've got on my screen in front of me, from some of the parts through to DCF through to, uh, trading comps, my range is $70 to $110 a share. Current is about $120 or just under. So, you know, I still think, I still think it's pretty frothy. but it's growing into that valuation.
All right, well, look, talk to me about Intel. Intel, I, I never hear people talk about other than I see its sticker on people's laptops, so it's evidently still a major player. But, but why aren't we talking about them in the same light of what we're talking Nvidia?
Yeah, I wanted to do a tale of two companies because the rise of Nvidia has been so singular, and especially for people that haven't seen a bunch of different shooting stars from valuation perspective and from, uh, excitement and investor hype perspective, I just wanted to take a step back and compare NVIDIA with the rise and staggering, staggering fall of Intel. So the headlines and the two companies have recently brought out their most recent earnings reports, and I just want to do a bit of a side-by-side comparison. So the market capitalization of Intel is $84 billion. NVIDIA's, as we discussed, as of today, $2.9 trillion. Um, Intel's EV over revenue is 2 times. NVIDIA's, well, we've said it, 40 times. So there's a massive, uh, NVIDIA's revenue growth year on year last year, 122%. Intel's revenue decline, 14%. NVIDIA's gross margin, 76%. Intel's gross margin, 42%. And now this is the staggering thing. In 2021, Intel's revenue was $79 billion and NVIDIA's revenue was $17 billion. Fast forward to trailing 12 months revenue, and Intel's revenue has gone down to $55 billion from $79 billion, and Nvidia's has gone up to $96 billion from $17 billion. How unbelievable is that in terms of a reversal of dominance and fortune? And I mean, so there's so many things that we can talk about. The 1st of August, and the August earnings report for Intel. It was a shocker. Share price down 26%, plans to cut 15,000 workers. Interestingly enough, we discussed data centers earlier on. Their data center revenue shrunk. If you think— so Nvidia's data center revenue, uh, for the 3 months, uh, that we just had was $26 billion. This time last year it was $10 billion. Intel's data center revenue in the last 3 months, $3 billion. The year before, $3.1 billion. So what the heck has happened for this company that, if, if we step back a little bit, was basically a founding member of the computer age? Intel was the company, not only from a technology perspective but from a management perspective. It effectively brought in through its chips, uh, it brought in the computer age, and it was one of the most valuable companies in the world. And it had this wonderful duopoly— so two companies dominating a market— it had this wonderful duopoly with Windows, where everyone bought a Windows, of course, and every Windows was powered by Intel chips. I don't know if you remember back in the day, the Intel Platinum, uh, Pentium, sorry, the Pentium processor. I remember that very early on in my computer buying history, the Pentium.
Cutting edge back in 1993. And so they were at the kind of— they were at the absolute cutting edge of all things that were good and innovative in the world of computing in the 1990s and into the early noughties. And then it all started to unwind.
Right, so things, things are pretty bad, is I guess what you're saying. But how did it get that bad, and how quickly? You would think that they could see, I don't know, signals happening requiring some sort of strategic shift. I'm guessing that they either didn't react or didn't see that. So where did it all start to go wrong for Intel?
Yeah, I've got a, I've got 5 strategic missteps and maybe 1 reason for those strategic missteps. So, and again, these might be quite prescient for Nvidia, which is at the top of the pile at the moment, but there's always new technologies coming down the pipe and, and they may not be quickest to react to the next wave. So I would say its most famous, Intel's most famous strategic error came in 2006. you might have heard about this one— where it declined Apple's request to produce chips for the iPhone because they were resting on their laurels that, you know, I— we've got 85, 90% of the chip market for, for desktop computers and laptop computers. I'm not sure the cost is going to be worth it for the iPhone. It's a little bit small fry, it's a little bit too expensive. So the— I— so Apple went to ARM and Samsung. So ARM as the, uh, the designer and Samsung as the manufacturer. And then now it uses TSMC. And TSMC, Taiwan Semiconductor Company, is now a $770 billion market cap company because they cottoned on to the fact that the infrastructure, the semiconductor industry is going to have to reposition for the mobile age. Intel could have done it, but they didn't because they were the big incumbent and they had these blink— these strategic blinkers on to say, yeah, well, you know, iPhone could be good, but you know, we've got a great business.
