DA Davidson Head of Technology Research Gil Luria joined Bloomberg Open Interest to explain why Nvidia’s ecosystem and chip demand could support growth through 2030, while warning that debt-heavy data-center builders may face serious pressure.
--------
Like this video? Subscribe to Bloomberg Tech on YouTube:
/ @bloombergtech
Watch the latest full episodes of "Bloomberg Tech" with Ed Ludlow here:
• 'Bloomberg Tech' With Ed Ludlow Full Episodes
Get the latest in tech from Silicon Valley and around the world here:
https://www.bloomberg.com/technology
Connect with us on...
X: / te...
What Gil Luria said
Written from the verified transcript and checked against it. Every figure links to the moment it was said.
Gil Luria, head of technology research at D.A. Davidson, said NVIDIA's investments in the AI ecosystem, such as the $8 billion in European startups, are aimed at making the ecosystem more robust, but the real driver is AI model consumption, not NVIDIA. He noted NVIDIA's valuation is 'very pedestrian,' trading at a discount to the market despite expected 70% growth next year, and that the cycle could last through 2030 based on ASML's comments. Luria highlighted a split market: NVIDIA, Broadcom, Micron, and SK Hynix trade at low multiples, while other semi and optical stocks trade as if the cycle continues. He warned that neoclouds like CoreWeave, borrowing at 9% to build commoditized data centers, may destroy value, and that chasing capital at high rates is a concerning governing factor. He called for Anthropic to disclose real financial numbers.
Key takeaways
NVIDIA is trading at a discount to the market despite guiding 70% growth next year, and the AI cycle could continue through 2030.
CoreWeave borrowing at 9% to build data centers with only 5% returns is destroying value, and chasing insurer capital is a concerning sign.
ASML's cautious capacity expansion, with orders into 2029 and 2030, is a positive signal consistent with the ecosystem's outlook.
The market is split: NVIDIA, Broadcom, Micron, and SK Hynix trade at low multiples, while other semi and optical stocks trade as if the cycle continues.
Anthropic and OpenAI need to disclose gross revenue, margins, and training costs as they are the best indicators of market willingness to pay.
Numbers and commitments
Figure
What it refers to
Type
At
$8 billion
NVIDIA's investment in European startups over the last year
1:24Gil, based on those comments and everything else that we've seen investment-wise from Jensen over the last number of years, is he single-handedly ensuring that the AI party keeps going?
This is the beginning of AI in Europe. In the last year alone, we have invested $8 billion in European startups. And in the next week, I'm going to announce another $2 billion more. And so we've invested tremendously in just about every single layer of the AI infrastructure here in Europe. And so the next layer above that is investments. These are going to be the superstar companies of the next generation.
Today, demand is far greater than supply. And the reason for that is because we're in the beginning of this new industrial build out. And so we need to build AI infrastructure around the world. And so demand is much, much greater than supply. Someday, of course, the two will come into balance. Now, whether it's five years from now or seven, ten years from now, which is kind of the duration of a new technology build out, except this one requires a lot of factories and requires energy and construction. And so there are natural limiters that keep the industry from building too fast. But eventually, it surely will happen that demand and supply will come into balance, but you're not going to see that in the next several years.
That was NVIDIA CEO Jensen Huang speaking in Paris, and joining us now is D.A. Davidson head of technology research, Gil Luria. Gil, based on those comments and everything else that we've seen investment-wise from Jensen over the last number of years, is he single-handedly ensuring that the AI party keeps going?
He has a very big role in it. NVIDIA has a very big role because they have so much resources, and they're really in the eye of the storm. So what they're doing is they're finding the places where the ecosystem needs to become more robust, and they're investing there. Whether it's fiber or new clouds or optical capacity or, in this case, more geographic diversity for the model companies, NVIDIA is finding ways to invest in those places to make the ecosystem more robust. And they're really in a position to do that because everybody needs their chips and because they have the capital to do so. But it's not that they're driving the wave. What's driving the wave is AI model consumption. It's companies and consumers buying more, paying more to use models for work and personal lives. And he's just making sure that that capacity is available because ultimately, if the capacity is available, more GPUs will be sold. And obviously, the companies he's investing in are more likely to buy NVIDIA GPUs than they are competitor GPUs.
Sure. And also, the bonus would be if he actually made a return on those investments. I mean, made a business model of it, right, and it worked out in some cases extraordinarily well for him. How do you value NVIDIA with that included? So, you know, returns on potential investments.
So the investments they've made have mostly gone up, but that's still, in NVIDIA's case, a relatively small part of their valuation. So we're not sure that we need to give them separate credit for that. What's far more important is that they have been successful in creating this ecosystem. Let's not forget that started in 2017 when he delivered the first GPU to Sam Altman at OpenAI. So he's been doing it for a while, and the return on investment is more the GPUs sold than it is markups on those positions. NVIDIA's valuation right now is very pedestrian. They are trading at a discount to the market in spite of the fact that they guided us that even next year... Right now, they're growing almost double. Even next year, they're going to grow 70%. So the market is skeptical of NVIDIA's growth. That's why you see Jensen with a little bit of a chip on his shoulder. But if they continue to deliver, the market eventually will give them credit that, as he is saying, the cycle looks like it'll be a three- to five-year cycle at least. He's saying a five- to ten-year cycle. But based on what everybody else is saying, including ASML this morning, the cycle is likely to continue through 2030 based on what we know now.
Yeah. I mean, it's funny. It almost feels like NVIDIA, you know, right in the middle of the maelstrom, is the one company that markets are actually valuing maybe appropriately. Right? This sell-off that we've been seeing in memory stocks, is it a healthy thing, for example?
9 more exchanges in this transcript
Sign in free to read the rest of this interview. No card required.
