00:00Get on the nostalgia bandwagon here, Mandeep, and walk us through what's happening here.
00:04It does feel like there was a bit of a lag, obviously a lot of excitement and enthusiasm
00:07for these new upstarts now on target to be worth many hundreds of billions of dollars
00:13in terms of valuation. But what accounts for these legacy companies having a moment right now?
00:19Well, all these companies were growing, you know, mid to high single digit at best,
00:24top line growth. And suddenly, you know, because everyone is looking to upgrade their
00:30infrastructure to run more AI workloads, a company like Dell is guiding to almost, you know,
00:3748% top line growth. And yes, they are, you know, benefiting from a new line of AI servers,
00:45which didn't exist two, three years back. But, you know, when you look at the size of these companies
00:52and the way they were growing in the past decade and now suddenly, you know, everyone is talking
00:57about sustaining that double digit growth, that's what's driving a reset in expectations.
01:04Dell stock popping 33%.
01:06It was crazy. I mean, Mandeep, we did that live on Tim and Carol's show. It was insane.
01:11The other stock that we talked about a lot this week is Snowflake. And I want to come back to
01:16that
01:16because the first thing I want to hit with you is talk to us a little bit about the difference
01:19between these kind of legacy physical AIs versus AI agents and why you're seeing more growth in one
01:26part than you are. As you see almost a little bit of a slump in some of the AI softwares
01:31as users
01:31decide what they're actually getting an ROI on.
01:35I mean, all these companies that are tied at the database layer, they're having their moment right now
01:43because when you talk about agenting deployments, you have to connect your enterprise database with
01:50these LLMs. And, you know, it's driving more queries for a company like Snowflake. And on top of that,
01:58Snowflake has launched their own coding agent. And we know coding agent is the real sort of big category
02:06of software that has emerged with the help of LLM. So really, they are in that right spot. Obviously,
02:13initially, the market treated everyone equally in terms of, you know, all the SaaS names getting
02:18clobbered. But now you're finding those spots within the SaaS market that are actually going to benefit
02:25with, you know, more agenting AI, more use of coding agents. And I think Snowflake probably is one of them.
02:32Let me pivot to the flashier, newer companies here for a moment, because I think the parlor game of
02:37the moment is the dueling IPOs, rather, between OpenAI and Anthropic. And we saw Anthropic taking
02:45the lead this week, becoming the frontrunner. And I'm curious how you see that binary unfolding as we
02:51push ahead to those IPOs, the way that Anthropic is positioning itself vis-a-vis OpenAI and vice versa.
02:57Yeah, look, Anthropic started off the year at around a $10 billion revenue run rate. They're already
03:06close to $47 billion in run rate in five months. And the projections are they will probably end the year
03:14at $100 billion revenue run rate. So it's, you know, a straight 10x. And they had a 10x last year
03:21as well.
03:22And these kind of hyper growth rates are really something that we haven't seen probably in the
03:29last 20, 30 years. And, you know, when I look at OpenAI, they had the lead. But it seems like
03:38in
03:38terms of growth rates, there's no doubt that Anthropic has the advantage. And even with SpaceX,
03:43you know, they're trying to pivot to NeoCloud. The acquisition of Cursor is a huge kind of inflection
03:50point in terms of, you know, shifting XAI towards growth. But Anthropic is by far the leader when it
03:58comes to, you know, these kind of growth rates. And they seem to be sustainable for the time being.
04:03It's interesting to me that you're seeing a harder investment on kind of like the back end of this,
04:09like the physical servers and the memory space. And while you're seeing some hesitation on the
04:15software side, depending on what the software does. And Mandeep taught me a new word this week,
04:20David Gurra. Are you ready for it? I'm ready. I think I'm going to get it right. Disintermediation.
04:24You don't want to get disintermediated. Mandeep, what does that mean? Six syllables,
04:29if I'm counting. And why does it matter? And why is it more important for it's easier for some of
04:35these integrated companies to keep their market share, whereas you said ones that are like a node
04:41and an add-on can get removed more quickly? Look at you. Look, I learned things. Mandeep taught me
04:46things. But explain it better, please, because that wasn't... I mean, look, the computing stack is
04:52changing. And, you know, when we think about LLMs, this sort of layer didn't exist. The model
04:59intelligence layer is a new sort of addition to the computing stack that we have, you know,
05:06accustomed to over the past 20, 30 years. And from that perspective, you have to see what
05:13is this model layer going to do going forward versus, you know, how the computing stack was
05:19organized before. And that's where, you know, you want to make sure that you are part of that
05:24modern computing stack with the LLM layer. And that's where someone like Snowflake has pivoted
05:32very nicely, you know, in terms of making sure that they're not trying to compete with the LLM layer,
05:38but really serving something that the LLM needs to deploy that agent in functionality. And that's not
05:45going to be the case for every software company. So really, the more you are adjusting your business
05:52model to make sure you are aligned with the LLM layer, I think the better prospects you have going
05:59forward. Mandeep, you've got the OG tech companies. You've got the kind of boring or historically
06:04boring tech companies as well. And Lisa and I sat down with Jitu Patel, who's the president at Cisco,
06:09just to tease an interview we're going to play a little bit later in the show. And I mean, he's
06:12wildly excited about the new Cisco and all Cisco is going to do. Yes, we think of it as this
06:15networking
06:16company, but it sees great opportunity when it comes to AI, a chance to kind of reinvent itself or
06:21have a kind of renaissance as a result of AI. Are there other companies like Cisco that you're
06:25watching that are going through a similar sort of reinvention or thinking about ways in which
06:30they can kind of rebrand or reanimate themselves because of AI? I mean, right now, everything on
06:36the hardware semiconductor side is going through a complete transformation. These were supposed to
06:43be, you know, commoditized companies that had no pricing power. Suddenly, you know, you look at all
06:50of the chip categories. I mean, we know hardware is in great demand, you know, across the board
06:57because the entire computing stack is being reimagined. The traditional chips are just not good
07:05enough to, you know, run these modern AI workloads. And that's where the compute density is what's
07:09driving these major shifts across the board.
Comments