AI Here and Now: Seven Critical Insights from Silicon Valley’s Tech Leaders

2026 Tech Tour: AI here and now

Jennifer Martin:
Hello, everyone, and thank you for joining us for today's T. Rowe Price webinar, Technology Tour 2026: AI is Here and Now. I'm Jennifer Martin, a portfolio specialist in T.Rowe Price's Global Equity Division, and I'll be your host for today's discussion. Technology remains the largest sector in global equities, and artificial intelligence continues to drive innovation and reshape industries as well as broader society and culture. Today's discussion will range across the current state of AI adoption, how real-world applications are scaling across industries, how technology companies are balancing growth with financial discipline, and what investors should focus on as clear paths to monetization begin to emerge.

But before we dive into today's discussion, let's highlight a few tips for getting the most out of your viewing experience. If you happen to experience any problems viewing the webinar, please try refreshing your browser or submit a question to our production team via the Ask Question widget on the left side of your screen. We also encourage you to submit questions to the panel. If we don't get to your questions today, we'll follow up as soon as we can. We've also added some informative content for you to download. If you're looking for more, please visit troweprice.com, where you'll find articles and information on the technology sector, artificial intelligence, and a host of other topics covering the economy, markets, and investments. And finally, please be sure to take the short survey after today's event. It's a great way for us to hear your feedback and improve on our future webinars.

Now, let's set the stage. As many of you know at T. Rowe Price, we explore distinct perspectives around the globe, which includes our annual technology tour to Silicon Valley, a trip that we've been making each year for many decades to gather the latest intelligence for this fast-moving sector. During our most recent trip in December of 2025, we had more than 40 portfolio managers and analysts travel to meet with dozens of public and private companies in the technology sector, where we engaged with top executives of the world's leading and emerging technology firms. It's always a great trip, and today I'm joined by our two technology sector heads, Dom Rizzo and Tony Wang, who led many of the meetings. Dom is a portfolio manager responsible for our global technology equities. And Tony is a portfolio manager focused on science and technology equities. Gentlemen, thank you both for joining us today. I'm really excited to be here. And Dom, we said we were going to start with you.

Dom Rizzo:
 Sounds good.

Jennifer Martin:
You're on record with sharing with clients that AI represents the most significant boost to productivity since the advent of electricity.
Having just returned from Silicon Valley, has that view changed?

Dom Rizzo:
Well, first off, it's great to be back at the seat and with you two coming off the trip. I mean, the trip is my favorite event of the year. And I mean, it's particularly fun for me to think back. I took over the strategy December 1st, 2022. So, the day after ChatGPT was launched, and then we had the annual tech trip that week. So, we're now three years on. And I think we started saying AI has the potential to be the biggest productivity enhancer for the global economy since electricity right after that first trip, right? And what do I mean when I say that? There's only three ways to grow an economy, right? There's capital, there's labor, and there's productivity. And if you look at the great productivity cycles historically, like electricity, you're lucky to get 1% a year to GDP growth if you have an amazing technology. So, electricity 1% a year for 30 years, basically. I think AI is going to beat that. We're already beating that. And I think that's why we're seeing such strong economic growth. I mean, it was incredible. The Q3 numbers were revised to 8% nominal growth, right, for GDP in the US. And I mean, 4% of that was true growth. And, you know, there was some inflation in there. But I really think we're at this exciting period of accelerating economic growth, driven by this new technology, AI, which is digital intelligence. And that digital intelligence is coming with immense capital intensity increases, some levels of speculation throughout the market, but overall, I think is going to be this major driver and economic force for both traditional companies and new age companies for many decades to come.

Jennifer Martin:
 Yes, and I know I heard you say in a meeting once, we've moved from "this is a cool toy" to "what are the use cases?" And you saw a lot of them on the trip, which we'll be sure to talk about.

Dom Rizzo:
 You know, Hock Tan, the CEO of Broadcom, left his earnings call. And, you know, I think he probably did a couple of sell-side calls right after, and then came into the T Rowe meeting, right? Their earnings call was at 2 pm—he's the CEO of Broadcom. And in our meeting, he had such a great quote. He said, you know, a few years ago, you could describe AI potentially as a bubble. But now it feels much more like a wave. And I think I'm coming around to that way of thinking. Yes, there will be productivity cycles. Productivity cycles always come with speculative bubbles. Our job is to navigate those speculative bubbles responsibly for our clients. But where we are at right now in this stage, this feels far more wave-like to me.

Jennifer Martin:
 Okay, that's super helpful. So, Tony, can you expand a little bit more on what Dom said about AI? Really, across all your conversations, were there any signs of slowing of the AI infrastructure build-out?

Tony Wang:
 Yeah, so AI continues to be the topic of discussion across all meetings. So, I felt like it's been great to see the evolution of how AI has been discussed. First, I think a few years ago, it was about ChatGPT launching. Then it was building the infrastructure. And now it's more about the application. How do we actually drive ROI? And so, you know, I think there's a few vectors that I'm thinking about AI. One, it continues to be unabated in terms of the spend, the interest. And I think that there's probably a few things to look for here is that where does the economic profit accrue to, essentially, and where are we on the S-curve adoption of various technologies that go into AI? And so, I think that in the first phase of AI, it was kind of Nvidia that took a lot of the stock price appreciation. And now more recently, we've seen energy be the bottleneck, networking, memory. And I think that it's interesting because that creates more investment opportunities, asymmetrical upside. So, I think that to me, what's fascinating is that, just as much as AI is becoming bigger, but like who are going to be the stocks that do the best in that kind of environment over the multi-year as well?

Jennifer Martin:
 Yes, I love that phrase, the profit pool is evolving, because I think you both saw that firsthand. And I think maybe, Tony, staying with you, and I think you kind of highlighted, hyperscalers continue to accelerate AI CapEx, and there really doesn't seem any sign of slowing. And so, I think a question many in the market are asking, and hopefully both of you can comment on this, is what we need to change for the spending trajectory to moderate, and how are you really positioning ahead of that inflection?

Tony Wang:
 Yeah, I think it's a great question and something that I think about quite frequently. And I would just frame it as, to me, it's all about the token profitability and whether it's a virtuous reinvestment cycle or not. So, when they buy a GPU from Nvidia or custom chip from Broadcom, like, you know, when they're selling that, you know, on their own cloud or deploying their own workload, are we to a point where the ROI is increasingly positive? So, there's probably two things with that is that one is that the cost curve continues to go down in terms of how much it costs to bring on new compute on a per compute basis, I would say. And so, you know, with each generation of AI systems, that's coming down. And then I think the other thing to think about is like, are these AI apps driving incremental profit? For companies, are the use cases expanding? I would say that right now, what we're seeing is that, yes, like absolutely. And the technology is getting better. It's hard to implement in some ways, but the applicability is like really vast. And I think that's what's really exciting about what's going on in the sector. And so as long as like, you know, the technology is continuing improving, agentic AI, for example, is probably the next big inflection, whether we can have autonomous agents or semi-autonomous agents, at least, execute decision-making processes that drive a ton of efficiency and then also probably drive new revenue opportunities. So, I'd say that's what's really exciting right now is that as long as that ROI flywheel continues to improve, I do think that investment cycle is durable.

Dom Rizzo:
 Yeah, yeah. There's a couple different pieces to what could cause a slowdown. When will a slowdown come, right? We're going from a $45 billion AI chip market in 2023 to a $500 billion AI chip market in 2028 to a trillion-dollar AI chip market in 2030. These numbers are stupendous, right? It's almost impossible to fathom, and when you take a step back, you have to say, Okay, first off, where's all the money coming from? Right? And I actually think there's plenty of money. We are accelerating the core businesses of the most profitable companies in history right now. Google Search is under attack by OpenAI, still growing 15% year-over-year last quarter. Meta is growing in the mid-20s, and we're seeing their core business. All that's driven by AI, optimization's getting better and better. We have the fastest growing startups of all time in the forms of OpenAI and Anthropic, right? OpenAI on a path to, you know, $20 billion at ARR (Annual Recurring Revenue). I think over the long run, probably the fastest company ever to $100 billion of revenue. Anthropic, probably high single-digit billions of revenue this year with their breakout hit Claude Code, right? Just let's take the coding market and think about how big it is, okay? There’re 30 million developers in the world. What does it cost each company in the world to have developers? $100,000 apiece? That means the world spends $3 trillion annually on coding intelligence. If we made all those developers 20% more productive because of Claude code, or any of these agents, any of these coding agents, and the companies that help enable that get half of that economic value, where does the economic value accrue? If you can get half of that economic value, that's a $300 billion potential market opportunity. And just to put that in context, all of enterprise software is only $400 billion. So, this is the biggest end market we've ever seen, digital intelligence in the consumer world, in the enterprise world. And then finally, competitive intensities increasing, not falling. Hyperscalar competitive intensity is increasing. None of them can take their foot off the gas right now. So, you put it all together and you say, wow, all the elements are there to keep spending. What ends it? Same thing that always ends it. Capital markets, exogenous shocks, interest rates going up, right? It's A reflexive cycle. Debt, private markets, IPOs are all working together to fund this. And if that slows, the rest of it slows. So, if the capital markets remain open, I think we'll continue to see strong spending.

Jennifer Martin:
 I think you both said economics, capital discipline, and existential.

Dom Rizzo:
There we go. I think that is why you're sitting in your seat, Jennifer. I think ChatGPT would be very proud of your summary right there.

Jennifer Martin:
I made really good friends. I do have a really good GPT. I have a Rizzo and Wang GPT sometimes working for me. So I do think, though, one of the things that you started to hit on, Dom, and Tony, you probably want to comment on this too, is just the AI scaling laws are intact, and maybe you want to explain what that is. But when you think about that topic, what are the most important implications for GPUs, memory, optics, and these data center build-outs over the next two years as these models continue to grow? So maybe because you kind of hit on it indirectly, so maybe let's be more direct.

Dom Rizzo:
 Just real quick, what are the scaling laws? I think the best description I've ever heard is, roughly speaking, if you throw 10x more compute at a problem, you get 2x more intelligent. Okay? You double the IQ of something by throwing 10x more compute at the problem. How do you throw 10x more compute on it? Basically, you throw 10x more money at it. Right? That's basically where we're at today. And we're no longer—I mean, Tony, you really taught me this. You know, Tony Tony's been an unbelievable Nvidia analyst for us. And we are way past the world of just a single GPU, right? These are systems, right? And the systems are GPUs, memory, networking, optics, power, all coming together in these incredible structures that in some cases are a million GPUs, long effectively. And just to put those numbers in context, each GPU consumes roughly the same amount of power as the average American house. So, when we're talking about a million GPU cluster, we're talking about adding the amount of power necessary to power a million-person city. It's like adding Dallas to the grid. Right? And so, I think these scaling laws are not slowing down. And all these different pieces, and I'm sure we can get into the data of each different piece, but all these different pieces are coming together. And that's why we're seeing a broadening out of that economic value capture that we've seen for the past two years.

Tony Wang:
 I think those are great points and totally agree. I think one thing that I've noticed over the years from covering Nvidia is that, a few years ago, I guess I think 5, 10 years ago, you could hold the GPUs in your hand, right? And now they weigh about the size of a small car. And so, part of that is just that to get that performance, you have to throw more of the system. So more, better interconnect, more GPUs, more HBM memory. Whereas I think previously you could actually get a lot of jumps with the software. And I think we're at a point where Nvidia might need to share more of the economics with the rest of the componentry. I think that's why you're seeing kind of Micron go up a bunch. You're seeing Sandisk, Western Digital, HDDs even, you know, in terms of we're seeing this broadening out of where the economic profit's going, which is pretty exciting because it creates more opportunities in the market that I think are maybe less trodden through versus the last few years.

Dom Rizzo:
 I mean, just some stats that blow my mind sometimes. Memory right now, quarter over quarter, depending on where we've seen the type of memory and how long the contract is, prices are up 50% to 80%, okay? EML lasers and optics are running 20% to 50% higher demand than supply. Packaging, co-As from TSMC, super, super tight, probably the bottleneck right now. Power, huge jumps in power consumption. Infineon gets $15 for every CPU server rack that they sell. They sell the power chips inside them. You know, 15 to 150, say, on CPU racks. That number's going to $15,000 in the GPU racks, right? So huge, huge changes in where economic value can capture. That doesn't mean that GPUs aren't going to crush it going forward. I think they will. I think. ASICs will crush it going forward. And in fact, I think CPUs will crush it going forward. So, I think there's a lot of people who are going to do very, very well. But it's interesting trying to find these different pieces. Where in the supply chain are we going to see that next big jump? And I think that's what the market's trying to figure right now. And I think that's the tough part, but the fun part about being active managers right now.

Tony Wang:
 I mean, just to drive that home point a little bit more, just you remind me, I got a new iPhone. And to get the extra 250 gigabytes, it was an extra $200 on an $1,100 iPhone, right? Just think about like, so it's actually impacting consumer PCs. And so it's just interesting.

Dom Rizzo:
 I don't think people realize the immense impact that memory prices will have on consumer electronics in 2026. So, I can't give advice, but if you're expecting to go buy an iPhone at some point, Tony's problem is about to get worse.

Tony Wang:

They've already locked in it.

Dom Rizzo:

They've already locked in good memory. That's at the good memory prices. Now memory prices have gotten higher and you need more because AI is super… I mean, the technical term is damn memory intensive. And it is incredibly memory intensive. And so… That's why we're seeing such crazy moves in some of these stocks is because economic value is rapidly hitting some of their income statements and resulting in, and frankly, like stupendous EPS growth.

Tony Wang:
 Buy your Nintendo Switch ASAP.

Jennifer Martin:
 I know, gosh. Well, so what I heard, scaling laws are intact and more compute buys better outcomes. So that's still happening. And then both of you really talked about the ecosystem broadening and really the shortages and scarcity that we're seeing across the supply chain and it's impacting iPhones. I'm just kidding.

Dom Rizzo:
 No, but it will. And it will impact... I mean, it really will impact all of all the market.

Jennifer Martin:
 So I think continuing on this, because you had a really interesting comment in one of our meetings, Tony, where semi-cap leaders are also signaling this like 18-month acceleration window. And so there's kind of a question of what gives you this conviction that this timeline, you know, is intact and really what could extend or compress it? Because I think there's something really interesting going on in semi-cap equipment kind of related to what you guys just highlighted.

Tony Wang:
 No, it's a great point. I mean, I think semi-caps have been on quite the adventure over the last few years, right? So initially, when ChatGPT was launched, it was like, oh my God, semi-caps are going to be great. Well, actually, AI only addresses a small segment of the percentage of wafers that are made. So actually, it wasn't that big of a beneficiary from AI initially. But here, like we talked about HBM, DRAM. And then you also had China, which was coming under kind of constraint from the export restrictions on semi-caps. They couldn't ship it to China. And then Intel and Samsung kind of took a backseat for a little bit, but now they're back. So, a lot of things are turning positively for semi-cap, I would say. One, China has normalized, if anything, is actually getting a little bit better for them. And then Intel and Samsung are more front-footed on spending. Memory prices are going through the roof. And then TSMC continues to be really strong. Here, I think that, we're definitely in a semi-cap upcycle. And, we're probably six months into, perhaps what's historically been a 18 to 24 month, how much semi-caps can run. And so, to me, we're probably early to mid innings on semi-caps. And so, to me, those are like kind of the, makers of the picks and shovels. You need equipment to make these GPUs and ASICs. So, I think it's exciting. And it's also a way that you can kind of play the space in a more diversified way with quality companies that have good free cash flow in a capital light.

Jennifer Martin:
 Yeah. You want to add anything to this? I mean, you've got a nice overlay. Both of you own some good companies in your portfolio.

Dom Rizzo:
 Well, Tony and I both covered the semi-cap space as analysts. Look, they're very strong business models, very high incremental operating margins, very high incremental free cash flow margins. And so, when you have a surge in demand, you see rapid EPS growth, right? And so, if you think about wafer fab equipment spending, roughly speaking, it's, you know, call it 18 to 23% of total semiconductor industry revenue at any given time. And that number, WFE is the industry TAM, that number is like roughly $100 billion right now. And I think with AI, we see a very clear path to a trillion-dollar-plus semiconductor market. And if you assumed 18% to 23% capital intensity, you could see a world of $180 to $230 billion in WFE at some point in the future. Now, I think Tony's point was really good about as semiconductor intensive AI is, it's actually not that many wafers. It's this weird paradox. And so, what actually drives semi-capital equipment spend is the wafers not the value of the chip, right? And that value goes to Nvidia's income statement right now. And so, I think we'll see how capital intensity plays out over the medium to long run as we hit those incredible numbers of $500 billion of AI chips going to a trillion dollars of AI chips. What does that mean for overall capital intensity? But I think broadly speaking, semi-capital equipment is one of the main beneficiaries of this whole build.

Jennifer Martin:
 Both of you highlighted some conviction from a combination of demand visibility and supply constraints. That's a nice setup. So maybe, Dom, let's switch to the other side of tech, which is application software, which seems to be an easy area to be underweight right now in portfolios. And a lot of it's because the AI value seems to be accruing to the compute and infrastructure. And so maybe you could highlight some milestones that would indicate software can compete with those AI economics.

Dom Rizzo:
 Yeah. Let's take a step back. What do I believe is happening? I think we're in a world where there's a five-year period where hardware has the power over software. And that's through this AI build. We just finished year three of that. So, I actually don't think we're going to see this—of course, you're going to see trades back and forth between the hardware stocks and the software stocks. But broadly speaking, I think we're in this hardware-intensive world. And why is that? Because of the scaling laws that we discussed earlier. When you put it together, we are completely rewriting the way all software is written, and we're simultaneously taking the marginal cost of software development to zero. So, of course, we're going to see radical change in the nature of the applications, right? And so what really matters going forward is not, were you the systems of record for customer relationship management or HR or finance? That matters a lot less compared to what is your data strategy? Where does your data sit? Can we perform AI on that data? And what is the operating system of performing AI going to be? I actually think Palantir has really stood out as the operating system of AI in the future. That being said… Palantir trades at a really, I mean, lofty valuation. So, but just in terms of business analysis, not necessarily stock analysis, Palantir has won this operating system of AI layer at many enterprises. I mean, right now, particularly the killer use case is insurance, but there's many, many different areas where they're starting to win. I also think there's a question of what do the model makers end up doing in this space? Do the model makers like OpenAI and Anthropic and Google with Gemini actually have an innate competitive advantage to the traditional application software companies in writing this new form of software because they understand the model so intimately? And then not to just bring it back to hardware because that's my favorite thing to talk about, but the way these software applications, these AI-native software applications, run is so dramatically different than historical software applications that the way they run changes the hardware layer underneath. And I'll give you an example. I really think 2026 is the year of the CPU. I think AMD and Intel are going to see a major acceleration in their revenue growth because of AI agent orchestration. And what the heck does that mean? It's pretty simple. You are giving the model a task, okay? And then that model is going to perform that task, not once like a human user would, but 1000 times in order to get the best answer. And when it's performing that task 1000 times, opening that web browser 1000 times, opening that web application 1000 times, building out this thing or that thing 1000 times, that is incredibly CPU intensive. And as a result, I actually think we're going to see a world where the traditional CPUs are going to see a renaissance in 26 as these AI agents start to permeate across both the consumer and the enterprise.

Jennifer Martin:
 Really helpful. It's interesting. Software seems to be sort of in that AI neutral territory right now. And then you made a really compelling case for the CPU.

Dom Rizzo:
 Well, it's very company specific. And I would argue that the valuations are not yet at a point where you have to go buy the software stocks. So software, the index is trading at roughly a 30% premium to the hardware index. Historically, it would trade at roughly a 50% to 70% premium to the hardware index. Now we're in that 25% to 30%, depending on the day, time, whatever. It's still a meaningful premium to the hardware index. Now you can make a ton of reasons why that's the case. Recurring revenue, historically, faster revenue growth. Historically, you would say margins, but you can't even say that anymore because the hardware margins have gotten so good relative to the software.

Tony Wang:
 I mean, I think on EV to sales on a relative basis, the two have traded back and forth in this range. And in December, they kind of hit kind of a peak, maybe hardware had a peak premium versus software. But at the start of the year, that range has been totally busted here. So, it's almost like if you're a reversion to the mean investor or a value investor, you kind of don't have the valuation floor anymore because the fundamentals have been so strong. And if you just look at Twitter, there's always like, it feels like every other week, there's like a new breakthrough in these models. So, it's like hard to fight the narrative, I think. of what these OpenAI anthropic models can do. And I think you made a great point is that the cost of development of software has just gone down so much, so deflationary, right? And you can spin up, you don't need a big software team anymore with LLM. So, I actually, I think Dom's been super right on this hardware call. It's been awesome. And so, I think it's coming through.

Dom Rizzo:
 And so, I think what will change it over time is new companies, right? There will be new companies that are AI native that are accustomed to developing in a world where the marginal cost of software development is zero. And then I think the big question that we debate all the time is, you know, you've heard a lot of this term economic value. Where is economic value gonna accrue? Does it accrue to the model makers or the new age application software companies or the old age application software companies because they have distribution? And the honest answer is it's probably some sort of combination of the three. The last thing I'd say is there's this real crowding out element happening in software. Look at the data-centric companies. They're growing rapidly. Look at traditional application software. They've decelerated. And so if you're data-centric, getting your data house in order, getting your data warehouse in order, you know, your data lake house in order, AI growth enables your business to accelerate. But if you're just like a traditional application software, I mean, where's the money? It's all on data and AI.

Jennifer Martin:
Yeah, and I think, Tony, you had a nice comment about valuations, and I wanted to maybe explore that a little bit more with you because big tech valuations remain reasonable on a growth-adjusted basis. So how do you distinguish between companies that compound and those that appear too expensive to sustain those future returns?

Tony Wang:
So I think there's a few different ways to think about it. So, I think about it as in terms of where are we in the S-curve of adoption for these companies. And so, when they're early cycle, they're going to look really expensive, like a Palantir, right? Because they're going from like 1% adoption to 10, 20. And they're creating a ton of market cap right along the way because of the multiple, the numbers are getting better. I think some of that application software that you were talking about, Dom, is like, those companies have been going up the S-curve for the last 10, 20 years. And so, like Salesforce, Adobe, they are kind of later stage. And so, they're going to trade at lower multiples. And I think a lot of the MAG7 are kind of in this compounding phase of the S-curve, right? So, Google was trading at 15 times, now it's at 25 times, still growing, healthy double digits and resilient to kind of AI, which is actually kind of augmenting its growth.

Dom Rizzo:
 I still think on Google, we need to see where that ends up happening. But unequivocally, up in this moment in time, AI has accelerated the core business.

Tony Wang:
It's been net good.

Dom Rizzo:
Yeah, net good, clearly. to today. What chatbots does to the search interface in three years, I think actually really does remain to be seen.

Tony Wang:
Yeah, and I think on valuation, right, looking at the MAG7, it actually is pretty reasonable to be valued. And a lot of them have gotten less expensive because earnings have been so good. You know, Nvidia trades at 25 times. You know, AMD, sorry, Meta trades at 15 times here. And so, I think it really depends on the company. But, you know, I think the companies that are early S-curve, like they are trading a big valuation because they're growing a ton. So that's what I think we look for is like, not just stocks aren't working, but stocks that have good fundamentals to support it as well.

Dom Rizzo:
My favorite stat just on MAG7 versus the rest of the market is just take a look at what is the PE of the MAG7 and what is their earnings growth rate and what is the PE of the S&P 493 and what is their earnings growth rate. and just do a simple peg ratio, right? Good old-fashioned Peter Lynch, one up on Wall Street, peg ratio. And the PE to growth rate for the MAG7 is cheaper than the rest of the market, even though they trade at a premium to the rest of the market. And why is that? It's because they're growing faster than the rest of the market. And so, as long as that remains the case, I struggle with the MAG7 being the bubble part. There's clearly pockets of speculation in the market. Do not get me wrong, and I could go through a bunch of areas that they are. I just personally don't think it's at the MAG7, or even I would say the large cap tech level. I'd broaden that out to most of large cap.

Tony Wang:
 Yeah, they generate a ton of free cash flow, right? They're self-sustaining, self-funding, and so it makes it better.

Dom Rizzo:
Yes, and, yes, and a lot of that free cash flow is going to Nvidia right now.

Jennifer Martin:
Well, maybe as sector leaders, can you both kind of highlight a little bit of your thinking of the outlook for tech? Can tech continue to lead the market? People are constantly asking us that.

Tony Wang:
Yes, so my outlook is that tech is set up really well. I mean, it's been amazing that we've had kind of 30, 50% up years over the last three years, which is incredible. But I still think that the innovation curve is still early here. And AI is going to be going to supercharge the economy. I believe in this productivity boost increase idea. I think it's inevitable. I like to invest in inevitabilities. But I think it just depends on where's the economic value going to accrue to. And I think it's going to go to part of the system, the componentry. So, I think semi-cap, DRAM, HDDs, SSDs are going to see nice appreciation from here. I still think that the mainstay of the MAG7, they'll still continue to compound nicely. And then I do think software does look a little challenge if you're an expensive kind of late stage software and you're decelerating, I think these LLMs are really going to put a lot of pressure on. And then kind of the Accenture is kind of an interesting battleground stock here because they reoriented themselves to be like a Palantir and become going from AI loser to AI winner. So I think there'll definitely be some changes in terms of leadership. So it keeps things interesting.

Dom Rizzo:
I think everybody wants there to be change in leadership of the market, right? Everyone calls small-cap, mid-cap, international. And the reality is, I personally think, and look, I run a global tech strategy. I have great names all over the world. I'm happy, international was better performing than the US this year. But these companies grow so rapidly, are so well-positioned fundamentally, have the benefit of addressing billions of people. Right now, competitive intensity is increasing between them, which is resulting in a huge hardware build-out. And then the question is, when is that battalion handoff from the hardware to the rest of the ecosystem? When are we going to go from hardware has all the power back to a world where software and internet have the power? And like I said earlier, I think it's a five-year cycle. Let's see how I feel next year. But we just finished year three. And so that economic value, where does it accrue? Potentially, it's hardware for longer. Potentially, it's a transition. But that's the thing that you and I probably spend the most amount of time debating, thinking about, changing our minds about, going back and forth on it. Because this is the fundamental question for every technology investor in the world today.

Tony Wang:
And I think you were alluding to this earlier, but it's not a zero-sum game, and what's amazing is that, I think that the space race for AI is actually going to multiple moons, and it's not just one pond, but you see what Anthropic is doing with Claude, right? They're focusing on coding. Open AI is more enterprise, and so I think it's going to be… Interestingly, TAM expansion, just like semis. I remember we'd always talk, there's always a debate of who's going to win, Nvidia or AMD or Nvidia and Brocka. Well, they kind of all won. All their stock, they're more surfing the same wave, borrow your term earlier, of this cycle. So I think net-net, it is still TAM expansionary.

Jennifer Martin:
So maybe the final question, and it's maybe a good place to land our plane or our spaceship, is there's always one meeting on the tech trip that has the biggest reaction from the team or the biggest buzz. I don't know if there was just one that you want to comment on, but everyone always comes back energized by some of those interactions, and I thought maybe that would be a fun one to end on.

Dom Rizzo:
 It was, I thought the Palantir meeting was a special meeting. Again, this is business analysis, not stock analysis, right? What is our job? Our job is markets, technicals, business, stocks. But we got to meet, was it, he was the CTO or the chief architect? I forget his title. Who cares what the title is? But basically the main tech guy at Palantir. And his analysis of the way the enterprise software development chain was changing so rapidly, and how they're going in and helping their customers build out their new AI-based operating systems, and how they're unleashing the knowledge capture that's within every enterprise today, right? Think about, you know, they always talk about, you know, your enterprise is your tribe, and people have tribal knowledge of your enterprise. And I always think about our internal, you know, tools that we built over many long periods. There were people who really knew how they worked, right? There's people who understand where is the data buried, how do we access it, where do we use it functionally, where do we not? And then putting it all in an AI native format, I mean, that really blew me away. And I think some of the use cases, particularly in insurance, were really exciting.

Tony Wang:
Yeah, I mean, like they showed us this incredible new module with agents and just kind of—it's interesting to see how they can get customers up and running in a fraction of time when they used to. And so I think it's going to be, makes new age software a lot easier to use because the new language of coding can be English actually. So a lot, you don't want to be that technical to spin up new use cases.

Dom Rizzo:
This is such a great example of Jensen says stuff and you have no idea what he means. And then three years later, you completely understand what he means. Remember Jensen used to tell us the new coding language is going to be English. And you'd sit there and you're, what are you talking about? And then now, it's like, oh, the new coding language is English. We're going to go build AI factories and the output is going to be knowledge tokens. That didn't make sense three years ago. That makes complete sense now.

Tony Wang:
 It's like he's from the future. No, totally. I think there are some really interesting meetings, like Anthropic. I was just impressed by—Great meeting. the breadth of what they're doing and just you can see the vision of the applicability. And then also it's inspirational to be at OpenAI's campus, like going to their headquarters. I mean, first of all, beautiful campus. People are working hard there. But they had one of their Code Reds going on. And so just their response and their aspirations to go after new markets, redefining how search is, not relying necessarily on advertising, perhaps streaming a new one. And I also think Lumentum, and you've covered Lumentum, like that meeting was incredible. I thought they do, as you would say, some of the hardest stuff, but they've never been able to get paid for it. But now these GPUs need lasers. And there's only a few of them. So I thought it was a fascinating tech trip and totally a highlight of our year and top it off.

Dom Rizzo:
 I think last year we titled it AI Forever. Is that the name? And today's, what do we call it?

Jennifer Martin:

Here and now.

Dom Rizzo:

Now, I mean, I think they both apply. Now, AI forever, AI here now, does not mean there's not going to be a cycle. Of course there will be a cycle. Of course we're going to navigate that responsibly for our clients. We're going to use our frameworks to do it. We're going to use our team. Tony and I are going to work tremendously hard debating, and we're going to try to catch this hardware battalion going to other economic values. But hopefully that gave everyone an insight into how we think, how we debate, and why our jobs are two of the best jobs in the world.

Tony Wang:
Love it.

Jennifer Martin:
Well, thank you both for really a lot of great insights on the sector. And really, this probably is a good place to conclude our discussion today. And so, I think we did cover a lot of material in today's discussion, but I think a few huge key takeaways that I heard you guys both talk about is AI has moved decisively from concept to economic reality, and there really is no credible off-ramp for AI infrastructure spending. AI CapEx remains existential for these hyperscalers. So that's just seems to be very front and center. But I think the area that you both spent a lot of great time talking about was talking about the demand for key technologies and whether that was GPUs, CPUs, memory, networking, they remain really strong and the industry is expanding. We did hit on a little bit that software does face maybe a little bit of challenges because their benefits aren't as widespread or not as obvious across everything, you know, we covered a lot of areas. But I think hopefully, most importantly, what our clients took away today is that our tech team provides a really competitive edge across these evolving market dynamics, multiple spaceships, multiple moons. And so, I love that, I'm gonna think about that more. But I sincerely hope that we were able to answer all of the audience questions. And if we didn't get to yours, we'll do our best to get back to you as soon as possible. And a final reminder to please fill out the short survey about today's experience, we wanna hear from you. And again, thank you both. It's been a great discussion and we really appreciate, Dom, thank you, Tony. Thank you. Thanks. And thank you to our audience as well.