저자: ToshiyaHari · 외국인
요약 보기
골드만삭스 반도체 애널리스트 Toshiya Hari는 엔비디아 매수 의견을 유지하며, 최근 주가 급락은 과도하다고 봄. 가속 컴퓨팅 수요가 하이퍼스케일러를 넘어 기업·국가로 확대 중이고, 경쟁 우위(혁신 속도)가 유지되며, 내년 실적 기준 PER 20배 초반으로 밸류에이션 매력이 있다고 설명. 다만 2026년 지속성과 AI 투자 ROI 논란은 리스크로 언급. 인텔은 공정 기술이 TSMC에 뒤처져 회복이 어렵다며 매도 의견 유지.
다듬은 글
We are here at the Goldman Sachs 2024 Communacopia and Technology Conference. I'm here on set with our executive editor Brian Sozzi. We are thrilled to be joined by Toshiya Hari. Toshiya is of course uh the lead analyst covering semiconductors at Goldman. Thank you so much for making time with us. We got to start on the Nvidia of it all.
You've got to buy on Nvidia despite the record sell-off we've seen in the stock over the last week. Make the case. Why are we still a buy? Yeah, thank you for having me. First of all, it's great to be here. Um yeah, so we've had a buy rating on Nvidia for quite a long time. And to your point, the recent performance hasn't been great.
Um but we do remain positive on the stock. A couple of things. I would say first of all, demand for accelerated computing continues to be really strong. Uh we tend to spend quite a bit of time on the hyperscalers, the Amazons, the Googles, the Microsofts of the world. But you are seeing a broadening uh in the demand profile into enterprise, even at the sovereign states.
Uh number two, I think their competitive position continues to be really strong. I think at this conference you'll hear from Nvidia Jensen Huang. Uh we have Lisa Su. We have a couple of the hyperscalers at the show as well. So there is a bit of a, you know, merchant versus custom silicon debate going on.
But we we do think within merchant silicon, Nvidia is the go-to. And even versus custom silicon, they've got the edge in terms of the pace of innovation. And last but not least, from a valuation perspective, um perhaps for the wrong wrong uh reasons, but the stock has come in quite a bit. Uh on our numbers on next year's earnings, the stock is trading in the low 20s, which I think is very inexpensive for a stock with this kind of profile.
Do you think Do you think the sell-off is just overdone? Yes, we do. Um I think clearly the stock had great great momentum uh until sort of the June-July time frame. And stocks never go up in a straight line. So we kind of appreciate uh why the stock has given back some. Uh I think the debate around is it sustainable into 2026?
It is a fair debate. are looking beyond 2025. They're valuing Nvidia on 2026 already? Valuing on '25, but depending on how the outlook is into '26, what kind of multiple you apply to '25 numbers is kind of the key debate. And and interestingly, we were having a very similar debate 12 months ago in 2023 when the stock was sort of range-bound the back half of '23.
And that was due to again multiples going from call it 50 times down to 30-25. And back then it was, "Okay, '24 is kind of in the bag, but what about '25?" So we're having a very similar conversation as of today into '26. When you look back on this earning cycle so far, too, there was this big question about ROI on AI.
Investors getting a little bit of fatigue, perhaps, on the amount of money being spent on AI. How much of a headwind could that be for Nvidia if companies are getting dinged for their spend on the expensive Nvidia chips? Mhm. Yes, I think um it is something to look out for. Um again, uh lack of customer ROI, whether it be the hyperscalers, whether whether it be the enterprises that are trying to leverage AI, that becomes a problem at some point, right?
We need monetization for this to be sustainable. I think the hyperscalers do make a pretty good case uh in terms of their ROI. I think if you look at ad revenue, you look at things like Instagram, the the the platforms, they are leveraging AI. And they are getting the return. So so we're confident as a collective unit that this is sustainable.
But to the extent there's a lack of monetization, that could be an issue for not just Nvidia, but for the all for the overall ecosystem. You mentioned the hyperscalers. That's the Amazons, Metas, Microsofts of the world. They're building their own chips now. Uh and they're certainly cheaper than Nvidia's.
Can they ever compete directly with Nvidia and take up more market share than they are currently? So they they do. They have been competing with Nvidia in some cheaper form. Uh I would say most of the custom silicon is there to address internal workloads. In the case of Google, it would be Google Search.
But GCP, which is external-looking, uh they're still buying uh a boatload of Nvidia GPUs. And and you can see make the same case for for Amazon. So I think you'll see a bifurcated market where again the internal workloads are addressed by their custom internally designed silicon. But what's outward-looking, they'll continue to buy Nvidia GPUs, AMD GPUs, and so on and so forth.
Important to note, you're not just the Nvidia analyst for Goldman Sachs. On the other end of the spectrum, you're covering Intel. And on the other end of the spectrum, you have a sell rating on Intel. Now this stock is near what, historical lows? They're announcing thousands of layoffs, earnings depressed.
What does Intel need to do? I won't say save itself, cuz that's just ridiculous. But how do they get their stock working? And when do their earnings start to improve? And when does the investment thesis start to improve? Mhm. Yeah, it's a great question. Um to your point, to your question, we've been pretty cautious on the stock for some time.
I think for starters, it would help for the company to start hitting what they guide to, right? Uh they they've disappointed the market uh quite frequently. And and that's never good for for a stock price. That's number one. Number two, medium to long term, they definitely need to get um on par if not uh exceed or surpass TSMC from a process node technology standpoint.
Because without that, they are behind the curve uh from a foundry standpoint. If they don't have the best tech or good tech, customers are not coming to to to Intel. They're they're going to continue to work with TSMC. And then on the product side, you can't really compete effectively with the likes of AMD, Nvidia, you know, Qualcomms, so on and so forth.
Do you think they will ever be Intel again where they are in that same discussion with an Nvidia and they are their time is past? Never say never. Um but I think it's fair to say that it's going to be an uphill battle. Because again, they in hindsight, they were perhaps complacent, uh didn't make the right investments.
They sort of allowed TSMC to catch up and to to beat them. Uh if you're a customer of TSMC and you've got this massive opportunity set in front of you that is AI, for you to essentially leave TSMC and go to Intel or any other foundry for that matter, that's a lot of risk. So I think for Intel for sure, but even companies or foundries outside of Intel to compete with TSMC in the broad growing eco-ecosystem, again, it's going to be an uphill battle.
Toshiya Hari, thank you so much for joining us and making time. We really appreciate it. And thank you to our audience for listening. Stick with us for more of your markets coverage right here on Yahoo! Finance.
Comments 0