00:00And you're right. I think the tech crisis points to the fact that this disruption is real
00:06when it comes to the AI impact. And NVIDIA is the poster child when it comes to the demand
00:13that they have seen for their chips. And look, all the hyperscalers, we started the year thinking
00:21they're going to grow their CapEx 30%. Now that number is 60%. Some of that increases because
00:27of memory prices. But then you are also seeing an increased plans to add capacity. And that's
00:34where NVIDIA will likely have upside to their numbers tonight. The question still remains
00:40around the sustainability of growth. Because when you are growing 70% at the scale that
00:45NVIDIA is at, you know, $300 billion run rate for 2026, the numbers, the growth rates will
00:52taper. It's a given. So that's the part that I think the visibility is what you want to
00:57figure out in terms of how much visibility the company has and what that means in terms
01:02of growth rates.
01:03I mean, again, I'm looking at the FA function, which is a consensus of all the analyst forecasts
01:09here. And again, in calendar 26 or calendar really 26, 50% revenue growth. I mean, it's
01:17just extraordinary here. What does the street need to hear from NVIDIA tonight? What are
01:23they focusing on these days?
01:25I mean, Jensen has called out demand being stronger at least 10 times, you know, since
01:32the start of the year, CES, Davos. And yesterday, you know, the fact that AMD had this deal with
01:38Meta, where they ended up giving away 10% of the stock as warrants. To me, that's a sign
01:45Nvidia still is the leader when it comes to, you know, everyone clamoring to get their
01:51chips. And everyone else is trying to make sure they stay in the race, which is what
01:55the move I thought AMD had yesterday.
01:58Yesterday, I mean, you basically put the surveillance cork in my mouth and Mandeep lectured me on the
02:02use of AI. I look at it as a $20 subscription. How in God's name did they pay for the
02:09billions
02:10of data centers with $20 from Alexis, $20 from Paul and $20 from Tom?
02:17So right now they are subsidizing it and it is being priced below the costs that, you know,
02:23these hyperscalers have to pay to Nvidia to get the chips because Nvidia clearly is not taking a hit
02:29on their margins. They still have that mid 70% margins. So prices will go up. That $20 subscription
02:36will go up in price. This is familiar, Paul. It sounds like streaming. I know, I know.
02:40And also the enterprise API usage. That's a big part of adoption. Don't just focus on the consumer
02:46part. That's corporations. That's corporations. They don't pay $20 per person. Yeah, they're
02:50signing multimillion dollar deals right now. I mean, Nvidia, you just mentioned $300 million
02:54in revenue. I just went on to look at AMD, $45 billion of revenue. It's not even close, dude.
03:00So is Nvidia that much ahead of everybody else in terms of this story?
03:06They are. And that Blackwell chip ramp that they showed last quarter, I mean, you will see it in
03:12full gear in the results tonight. And now they have announced the next architecture, the Rubin
03:17architecture. So that's how far ahead. They are at least, you know, 18 months ahead of everyone in
03:22terms of the efficiency and the compute capacity and power that they are able to show with their chips.
03:28Across the nation today, Mandeep Singh with us with NROG Ron. I just can't say enough about
03:33Bloomberg intelligence ability on technology. We threaten to have both of them on together,
03:38but they can't be in the same room at the same time. Okay, Mandeep, here's the real world on
03:43depreciation. I had a 2019 Intel cheese grater. Fancy, fancy Mac computer. Fancy people like me can buy.
03:51It has a depreciated value of under $400. The fancy Nvidia chip in the bottom of my
03:58cheese grater is now holding up a martini at 5 p.m. I mean, tell us about the depreciation
04:05and the technological pressure that Mr. Wang is under to get better, better, better. And everything
04:11else becomes a coffee table for a martini. Yes. And that's a great point about the long-term,
04:17you know, multiple that this company can get despite its growth is because of that aspects.
04:23But right now, the real constraint is power. So everyone has a finite amount of power and you
04:29want the best compute. And that's why people are willing to upgrade to Nvidia's latest architecture
04:35because they can't add more power. They want to use the chip that is best optimized for the power
04:41they have available, which is why Nvidia is releasing their new generations so quickly within a year,
04:48in some cases. It's Blackwell. It is? I don't know. Somebody whispered it in a year.
04:53Look at the rental pricing of the older versions of their chips.
04:57Rental pricing of older chips. Yes. I mean, they're still off the charts. Everyone wants to get a hold
05:03of whatever the market has to offer in terms of the rental price. All right. So in the last three,
05:09four, five, six weeks, we've had this sell-off in software stocks. What are you and Anurag telling
05:14institutional investors these days about that? That it's going to take a while to clear this
05:19out in terms of these companies adapting, you know, the LLMs situation now. They have to integrate
05:27LLMs. They have to offer, you know, functionality that's built on top of LLMs. And it's going to take
05:34a while to play. But don't I just run out and buy Salesforce.com hand over fist here? I mean,
05:39is this, is AI going to put a Salesforce.com? Not out of business. The terminal value is not
05:46in question, but it's going to take a while to figure out the margin impact, when the growth
05:50will rebound. And it's not going to happen within a quarter or two. See, finally, Tom,
05:54these tech analysts, investors actually have to do some work. Before they could just sit back and say,
05:59my business is growing 20, 30%. He's built his career on just my business is growing 20, 30% a
06:04year.
06:04Ian Anurag. Now I think they got it like, oh boy.
06:08Okay, one final question. You know why it's waiting for us here. We got to get them on.
06:13Mandeep Singh, Microsoft AI, transform with Microsoft AI where trust meets innovation.
06:20Great MBA statement. Okay, wonderful. What's the level of panic at Microsoft right now?
06:26Good question.
06:27In the AI wars?
06:28I mean, look, the whole open AI situation is what's really causing the panic. Microsoft is still
06:35deeply embedded to Paul's point. You know, they're a system of record in a lot of the cases.
06:40And most likely you are going to add AI on top of Microsoft software. You're not going to rewrite,
06:48you know, what Microsoft is doing with Office or Windows or all the other software they have.
06:53The question is, they don't have their own LLM. And that's where the reliance on open AI is hurting
07:00them because open AI wants to get into applications themselves. And that whole thing is not panning
07:05out very well.
07:05I thought LLM was a law degree out of the United Kingdom.
07:08Yeah, I know.
Comments