Chip Market Dynamics: Growth, Valuation, and AI Influence

The compute market is rapidly expanding, favoring chip companies with scale and expertise.

### The Evolving Landscape of Chip Companies and AI Applications
* 💻 The compute total addressable market (TAM) is growing significantly, driven by the **increasing demand for foundational models and custom compute chips** from companies like Amazon.
* 📈 Marvell, along with India and Broadcom, represents a selective focus for investment in the compute sector, showcasing the **importance of scale and expertise** in capturing market opportunities.
* 🤖 Murphy's leadership at Marvell has created **substantial value**, particularly through effective **M&A** and the development of **AI custom compute chips**, making it unlikely for him to leave the company.
* 📊 Nvidia's valuation, currently at a **$1 trillion market cap** with a forward earnings multiple of **30 to 35 times**, is considered reasonable given its profitability and growth trajectory, despite concerns about sustainability.
* 🚀 The tech industry is on the brink of a **productivity revolution**, with digital agents transforming labor dynamics and software companies vying to be the **primary data management platforms**.
* 📉 While Palantir has achieved high valuations, not all companies can replicate its success in AI; its **unique approach to customizing AI solutions for clients** sets it apart in the competitive landscape.
* 🌐 The shift from chip dominance to software innovation reflects a broader trend in the tech industry, with firms recognizing the need for **application-layer advancements** to sustain growth.
* 💡 As AI continues to disrupt various sectors, only a select few companies will **excel in harnessing its potential**, underscoring the **complexity of integrating AI into business operations**.

Basically, you look like you were in every chip name apart from Intel, in many ways. Tony, is that right? And how are you looking at the ecosystem and its strength?

You know, when you look at our holdings, we currently are focused on the compute aspect of things. And so, look, India is a big holding, as well as Marvell and Broadcom. My view is that the compute TAM is

growing significantly, and the foundational models are kind of growing exponentially every year. And so the chip companies that have the scale, the expertise to take out the leading edge are, I think, going to be advantageous in terms of capturing this compute.

And so I think, you know, I like that space long term, but we are selective in our holdings here. You know, specifically those three, especially. We'll talk about Marvell, for example.

Would it be a worry, as the leader did, to jump ship to Intel? Well, I think that Murphy has created a ton of value over the years. You look at when I picked it up; it was in a sub-$10 billion market cap.

And he's, over the years, done a lot of great M&A in capital allocation, spun businesses, and acquired a world-class custom team that is right now delivering a lot of AI custom compute chips for the likes of Amazon, for example.

And so, you know, I think that it would be a concern if he jumps ship. I do think it's, you know, he's put a lot of work into Omaha Valley. He's got a great team there. And they're kind of firing on all

cylinders coming out of the last few years. So, you know, for me, from my perspective, it seems like he does have a lot of good reasons to stay at Marvell. And so I'm, you know, I'm taking his word for what he said last night, kind of affixing

value here. Looking at your top ten holdings, Marvell is one of them. At the end of October, it's about 2% of the fund. The top holding is in video, as you mentioned. Interestingly, we had CEO Matt

Gorman on the show, on the program across our network yesterday, and we're going to be playing a bit of that sound later. But Amazon is the best points contributor to the Nasdaq today. They had a whole raft of chip announcements.

What does that mean for NVIDIA? What does it mean in terms of their own customers building custom silicon? Yeah. I mean, I don't think that story is anything new. It's just generally the evolution of compute. Essentially, you have new workloads

that need more general-purpose Nvidia chips, and then over time, as the workloads get bigger and they mature and stabilize, you kind of bring in the custom compute. And so, you know, I think the two kind of coexist, and they have coexisted over

the last decade.
And so going forward, I think of them as kind of a symbiotic
relationship in some ways. And that is just evidence that the
chip's total addressable market (TAM) is increasing. The computing demand continues to
increase because the workloads are getting

to be so big that it makes sense to have
a custom chip. But you know, what's interesting is that
previously, five or ten years ago, we thought, hey, AI would
just be used for things like

identifying what's a cat and what's
not a cat, I imagine. So now it's like transformed into
something that we no one could have thought about or predicted.
So, I think that it's just we've had this debate before about why.
But I think that what you've seen is

that both Nvidia, AMD, and custom
chip providers have done well as the tandem has increased.
'I am not a cat' will forever be part of internet history.
What then about Nvidia's valuation at these levels?
You're talking about how finally we're

getting a return on investment from an
application layer perspective of AI. We're actually using it in our everyday
life and work life. Can Nvidia's growth in terms
of its valuation continue? Have we priced it all in?
Yeah, I think that's a great question.

I mean, a $1 trillion market cap is
pretty astounding if you look at the
profitability and free cash flow.  Even so, a 30 to 35 times forward earnings
multiple is pretty reasonable.

And so while the stock price has gone up a lot, I think that when you look at valuation, it's not extreme, not at 100 times earnings, which I think says the market is saying there is a little bit of a

concern about how sustainable the numbers are, whether profitability can continue at these levels. And so, I think the valuation at face value is actually relatively moderate, given its history and its growth. It is extraordinary to take a step

back and think that on this day we are at a record high for the Nasdaq 100. Apple's at a record high. We've got Nvidia continuing to rise in terms of market cap. And then you have the Salesforce numbers and its own record high in terms of share price today.

What do you make of their application of agent force and whether it's actually going to make them more profitable as well as getting revenue off the ground? Yeah, I think it's an exciting time in tech and that, you know, it could be this

next productivity revolution. You're kind of upending labor as we know it with digital agents and digital labor that essentially doesn't sleep, doesn't even take breaks, operates 24/7 and can operate alongside humans. And so I think that's what's really

exciting. And all these software companies are now kind of seeing the opportunity and going after it. Everyone wants to be the platform where you store your data and operate. And so whether that's Microsoft, ServiceNow, Salesforce, like I

I think it's a big opportunity.
And so we've kind of seen, you know, the first year of chatter pretty much dominated by the chip companies, and really dominate the infrastructure
layer. And so now I think that we're seeing it
on the app layer, which is really

exciting because you need to see it in
the app layer for the chip companies to continue doing well.
And, you know, I think we are kind of seeing that in the last earnings
season. We've seen that shift, as you say,
movement from all-in on chips, then

shifting into software.
But boy, is it playing out in some of the valuations.
I'm looking at Palantir trading at 180 times future earnings; they've been the
darling of the AI story. Will all companies aspire to reach
those valuation levels, or is

Palantir overvalued?
You talk about how AI is going to disrupt the
workforce. But I feel that the market's been trying
to price that in for quite a while. Yeah, absolutely.
And I think that not every company can

be like what Palantir is doing.
I think doing AI is really, really hard. And they've been elite for the
last decade. And I think of them bringing
AI to the government and the enterprise.

And so I think what's special about them
is that they've got this very centralized ontology for AI, and then
they can customize that for clients, and clients can build on that.
So I think of them as a platform.

I mean, you think about what the United States government is trying to do now is to be more efficient. And part of that is bringing AI and technology to automate things and to modernize people's processes and data. And so I don't think every company can

be like Palantir. I think there's like some special ones in the AI software, kind of similar types of area, like, you know, kind of like Applovin, for example, or Axon. I like, you know, where there's like specific vertical applications where it

can be applied and really drive a lot of automation and intelligence.

In Search of AI Opportunities

Welcome to the Bloomberg Technology YouTube channel. Subscribe for original series and mini-docs on game-changing technologies, profiles on innovators, live conversations with experts, and breaking technology news & analysis from Bloomberg TV.

Bloomberg Technology

Current holdings focus on compute, with significant investments in India, Marvell, and Broadcom.

Compute TAM is growing significantly, favoring large-scale, expert chip companies.

Selective holdings, particularly in Marvell.

Discussion of Marvell's CEO and potential impact of his departure.

Assessment of Marvell's strong performance and reasons for CEO to stay.

Impact of Amazon's Chip Announcements on NVIDIA

Amazon's chip announcements and their implications for NVIDIA and custom silicon.

Coexistence of general-purpose Nvidia chips and custom compute.

Symbiotic relationship between general-purpose and custom chips, reflecting increasing computing demand.

Evolution of AI workloads leading to increased demand for custom chips.

Unexpected growth of AI applications beyond initial predictions.

Nvidia's Valuation and Market Performance

Question regarding the sustainability of Nvidia's growth and valuation.

Assessment of Nvidia's valuation as reasonable despite its high market cap.

Market concern about the sustainability of Nvidia's profitability.

Record highs for Nasdaq 100, Apple, and Salesforce.

Impact of AI on Software Companies and Productivity

Potential of AI to revolutionize productivity through digital agents.

Software companies competing to become the leading platform for AI data and operations.

Shift from chip-focused infrastructure to application-layer AI.

Importance of application-layer AI for continued chip company success.

Discussion of Palantir's high valuation and whether it's overvalued.

Challenges in developing AI applications and the uniqueness of Palantir's approach.

Palantir's centralized AI ontology and its application to government and enterprise clients.

Palantir as a platform for AI applications in government modernization.

Not all companies can replicate Palantir's success; examples of other successful AI companies in specific verticals are given.