Adam Patti
The AI Supercycle: A New Way to Invest in AI
Adam Patti has been in the ETF market since 2001, when he was running a division at Time Warner called Fortune Indexes that created a stock index based on the Fortune 500 and partnered with State Street to launch the FFF ETF. He founded Index IQ in 2006 as one of the earliest ETF issuers (roughly the 18th to launch), focused on packaging institutional-quality liquid alternatives in an ETF wrapper. He sold Index IQ to New York Life in 2015, helped roll out their ETF platform globally, stepped away from the industry for a few years, and then partnered with John McNeill and DVX Ventures to create VistaShares.
On this episode of Behind the Ticker, Adam walks Brad through AIS, the VistaShares Artificial Intelligence Supercycle ETF. It takes a "bill of materials" approach to AI investing, focusing on what actually gets built (semiconductors, data centers, cooling systems, infrastructure) rather than the applications that sit on top.
The Problem With Existing AI ETFs
Patti's critique of the existing AI ETF space is pointed: most don't give you the exposure their names promise. Look under the hood and you're getting the Mag Seven for the vast majority of your exposure, with some Netflix, Tencent, and Salesforce thrown in. These are massive companies making hundred-billion-dollar AI investments, and they'll likely be significant consumer-facing players once AI applications mature. But that's not how to play AI right now.
"Without the infrastructure, there's no consumer-facing applications," Patti argued. "Without developing the compute to take these AI models from the training portion of development to the inference portion and beyond, there's no innovation, there's no ability to create that value for consumers." AIS is designed to invest in the infrastructure: data center components, semiconductor supply chains, and the physical systems that make AI possible.
Supply Chain Analysis and Bill of Materials
The portfolio construction starts with supply chain analysis. What are the drivers of AI? Where is the investment going? What components are required to build AI infrastructure? Then comes the bill of materials: how much does it cost to build an AI data center? What percentage goes to cooling systems, fiber optic cable, racking, power systems? On the semiconductor side, what are the GPUs, VRAMs, cooling systems, and other components that go into building the chips themselves?
Holdings are weighted first by their contribution to the AI supply chain. If cooling systems represent 25% of a data center build, companies in the cooling bucket get a 25% aggregate weight. But the analysis goes deeper: just because a company sells cooling systems doesn't mean it's selling to AI data centers. The team examines project pipelines, AI-specific revenue, and actual customer relationships. "You need to understand what are the projects that they're working on, what is their pipeline looking like, how much revenue are they actually getting from AI projects," Patti explained.
The Investment Committee
On top of the rules-based core sits an active overlay managed by an investment committee of four people, split into two groups. The "practitioners" are John McNeill (former president of Tesla, former CLO of Lyft, GM board member) and Sunny Madra (president of Grok, one of the leading AI companies). Their job is to identify risks and opportunities before others see them. "Sunny is building data centers around the world," Patti noted. "He knows who the players are probably before most people do because he's working with them." The portfolio construction side includes Patti and Professor Robert Whitelaw, former dean of NYU Stern's undergraduate college and former chairman of their finance department. The rules-based core is rebalanced twice a year, but the active overlay allows rapid adjustments in a space that moves fast.
Super Cycles Beyond AI
VistaShares' thesis is that four super cycles are happening simultaneously: AI, robotics, electrification, and biotech. Each is a technology revolution on the scale of the internet. They chose AI as their first product because it's the most active and investable right now, with the clearest supply chain to map. Future products will address the other super cycles, and the bill-of-materials methodology is designed to extend to those areas as well. Consumer applications will eventually become a significant piece of the AI supply chain, but not today. Infrastructure comes first.
Key Takeaways
- AIS uses a "bill of materials" approach, weighting holdings by their contribution to the AI data center supply chain (semiconductors, cooling, power, infrastructure), not by market cap. Revenue from actual AI projects is verified.
- The investment committee includes John McNeill (ex-Tesla president, GM board) and Sunny Madra (Grok president) as practitioners with real-time visibility into AI infrastructure deals.
- Most existing AI ETFs deliver Mag Seven exposure with AI branding. AIS deliberately focuses on infrastructure components, not consumer applications.
- Adam Patti founded Index IQ in 2006 (sold to New York Life in 2015) and has been in the ETF market since 2001. VistaShares is his second ETF company.
- Positioned as a 3-5% growth equity satellite. VistaShares plans to extend the super cycle methodology to robotics, electrification, and biotech in future products.
Listen to the full conversation on Spotify, Apple Podcasts, or YouTube.
Full Transcript
4,034 wordsMachine transcribed from Brad Roth's conversation with Adam Patti, with speakers identified automatically. Timestamps link to that moment on YouTube. Lightly cleaned, otherwise unedited.
Behind the Ticker is brought to you by UX Wealth Partners. If you're a TAMP user and you're sick and tired of the legacy technology they are run on and you want more customization and flexibility, as well as an AI-driven model marketplace, UX Wealth Partners is your destination. On top of that, they have institutional trading. So if you are an ETF issuer or an SMA provider looking for outsourced institutional trading, UX Wealth can also be your destination. So check out uxwp.com to find out all the ways UX Wealth Partners can help grow and make your practice more efficient.
Welcome to Behind the Ticker. Today we have on Adam Patti. He is from VistaShares. And we were talking about AIS, which is their ETF, the VistaShares Artificial Intelligence Supercycle ETF. It's very, very well crafted, very solid team. They're looking more at what they call the bill of materials and pure exposure approach to investing, where they're investing more in all of the things it takes, like infrastructure, data centers, semiconductors, rather than just traditional AI applications or some really overweight exposure to those MAG7 names.
So they're really looking underneath the hood. It is very well thought out. And I think they're going to have some success here in the space. So I will let Adam tell you everything about the product. So without further ado, please welcome Mr. Adam Patti.
Hey, Adam, welcome to the show. Thanks for having me.
Read the full transcript (44 more sections)Collapse transcript
So before we get started, why don't you take a little bit of time and tell everybody a little bit about your background and how you ended up founding VistaShares. Sure.
I've been in the ETF market dating back way back to around 2001, where I was running, among other things, a division over at Time Warner called Fortune Indexes, where we created a stock index based on the Fortune 500. And we used that index to create an ETF with State Street. The triple F was the ticker. That was way back in the early days of the ETF market. But that gave me the first taste of the market and the growth prospects and how interesting the ETF wrapper was for investors and how efficient it is to use in a portfolio. So fast forward, I ended up taking that knowledge and founding a company in 2006 called Index IQ. We were one of the early ETF issuers.
I think we were the 18th or so that launched. And we focus on liquid alternatives, trying to package institutional quality alternative investment strategies in the ETF wrapper. Built that up, sold that to New York Life in 2015. Was at New York Life, rolling out their ETF platform globally for a few years. And then I left and spent a few years out of the industry and was dying to get back in and had the good fortune to meet up with a guy named John McNeil and his DVX Ventures company. And John and I ended up partnering to create Vista shares and create high quality products for advisors, but using a more thoughtful approach to the thematic space.
Yeah, no, it's great. And we're going to talk about AI today a little bit. But before we get into kind of geeking out on the ETF and AI, I always like to ask, what do you like to do for fun? Any hobbies?
You know what? It's an interesting question. I have three kids. One's going to college, one's in college, one's going into high school. So I'd like to say I have hobbies. But I think all my hobbies have been usurped by my three kids' constant demands for whatever they happen to be doing at the moment. That combined with a startup company is taking up all my time. But I used to like to work out and lift weights and play tennis. So hopefully I'll get back to that soon.
Yeah, right now you're just an unpaid Uber driver, I would assume. Exactly. So let's talk about the ETF you guys launched, which is AIS, the VistaShares Artificial Intelligence Supercycle ETF. So we're seeing a lot of thematic AI ETFs starting to be launched. So how does this differentiate itself from maybe some of the other AI-focused ETFs in the space? Sure.
Well, first, just talking about the supercycle concept, just to take a little bit of a step back. part of what DVX Ventures and John kind of came up with this brainstorming session early on to try to figure out how to create a better exposure to some of these disruptive trends. And what's interesting is that if you look at what a supercycle is, it's really a long-term disruptive trend founded in the technology revolution, right? So the last one was, of course, the internet. That was a single supercycle that, of course, changed the way we work, that we live our lives, communicate. But interestingly now, there's really four supercycles that are occurring at the same time.
It's AI, it's robotics, it's electrification, and biotech. And they certainly play off each other. But each one of those is really changing the world. So we chose to focus on AI as our first kind of foray into the market. And the problem in the market is that, and this is really a problem across thematic ETFs generally, they don't necessarily always provide investors the exposure they think they're getting. So just because something's called artificial intelligence doesn't necessarily mean the companies you're buying in the portfolio have much to do with AI or at least aren't the key drivers of the space.
So if you look under the hood, many of these ETFs out there in the AI space, you're really getting the mag-7 for the vast majority of your exposure with a little bit of Netflix and Tencent and Alibaba and Salesforce.com thrown in there. It's not really – and look, I understand that these companies are making massive $100 billion-plus investments in AI and certainly are players in that space. And we'll certainly be very, hopefully, significant consumer-focused players once these consumer-focused applications become important to people.
But in the meantime, the way to play AI is really by looking at the infrastructure. Without the infrastructure, there's no consumer-facing applications. Without developing the compute to take these AI models from the training portion of development to the inference portion and beyond, there's no innovation. There's no ability to create that value for consumers. So we have really just focused on the infrastructure, in particular, the data center and semiconductor infrastructure that's really driving innovation. And we've come up with a methodology and an investment strategy that we think really captures it very efficiently.
So I was going to ask you about – you guys emphasize a lot this pure exposure to AI super cycles. Is that what you're referring to there, kind of investing more in the infrastructure? Can you kind of elaborate on what that means?
Sure. So to us, there are two important – so just to back up, what we do very simply is we look at the supply chain. So taking AI, what are the drivers of AI? Where is the investment going? What are the components that are required to build the AI infrastructure? And that's really a supply chain analysis exercise. So digging into the supply chain, understanding what each segment of it really means to the growth of the industry. Then looking at the bill of materials. So how much does it cost to build an AI data center? How much does it cost to build a high-performance AI chip?
And then what are the components that go into that? So in the data center side, just using an example, to build a data center, it's cooling systems. It's fiber optic cable. It's racking. It's all the pieces that need to go into building that data center. Similarly, on the semiconductor side, it's not about – for us, it's not about buying the chip companies that are selling the chips to the data centers or other vendors. It's about finding the components, the GPUs, the VRAMs, the cooling systems, everything that goes into building the semiconductor chips themselves. And that's how we're getting to that pure exposure. Looking at the supply chain, overlaying the bill of materials, then looking at the companies that contribute to each segment within that supply chain, then doing a deep dive into the financials of those companies.
Because just because a company is a, quote, cooling company for data centers doesn't necessarily mean that they're providing their services or their products to AI data centers. You need to understand what are the projects that they're working on? What is their pipeline looking like? How much revenue are they actually getting from AI projects? So that's the next level of analysis. Try to really dig in to see how much AI, direct AI revenue is coming from these companies and then building that into a portfolio. Yeah.
So with such a focus on infrastructure, data centers, the semiconductors, rather than, as you kind of put it earlier, some of these AI applications, how do you see this focus maybe evolving over the years as AI adoption matures? we're still, we've moved fast, but we're still in the very early innings. So are you, is this ETF going to continue to kind of follow that bill of materials approach? Or can you kind of see your investment thesis evolving over time?
Well, the bill of materials is the core of it. So we do have, it is an active product and I can explain how we use our active management overlay to create value. But yes, we have the ability within our rules-based process to add additional supply chains. So certainly consumer applications will at some point be a significant piece of the business and maybe even the largest, but not today. Not without the infrastructure required to create those applications. So in the future, we will, probably be bringing out, adding to that supply chain focus and focusing more, bringing that consumer application piece into the, into the portfolio.
The other piece is the energy piece. That's another piece that we have the flexibility to bring in as well. Once that gets to the point where it's more investable and a little more mature.
So you would mention, which I agree with you, a lot of these AI centered focused ETFs right now are heavily weighted. NVIDIA, Microsoft, you guys are much more diversified in terms of underlying holdings. I believe at last look and don't quote me on this. There's, some like 72 companies across various segments of the supply chain. So as you referred to kind of your active approach and rules-based approach, what are some of the key criteria for really selecting these companies and putting them in the portfolio? Sure.
So just, again, going back to the supply chain focus. So you break down the supply chain, you determine what percentage at each level of supply chain goes into building the end product. So on the data center side, call 25% is cooling systems. 2% is racking and cabling systems. So you really need to do that analysis, break down the supply chain, apply that bill of materials to see how important is that segment of supply chain into building the end product. Then finding those companies that are actually have the AI driven revenue within the segment. So, it's, it's, it's a time consuming process. It's heavy in analysis. And the way we've accomplished this is we've built this into a rules-based process that we filed a patent on.
So the core of the portfolio is, is rules-based. It's transparent. And we did that because we think it's important for investors to really understand what they're getting so that there's no surprises. understand, why a certain company is in the portfolio, what segment of the supply chain do they fit into, and why are they in there? So that's important. And we rebalance that core of the portfolio twice a year according to the rules. Then, of course, we have our active overlay. So the active management piece is, I think, a critical piece because particularly in a fast-moving, very dynamic, competitive space like AI, things change very quickly. You have to be nimble. You have to have people that really understand the space to identify the risks and opportunities that are presenting themselves on a daily basis.
So we have, our investment committee is comprised of four people. We do it a little differently than I think most fund companies kind of look at it. We have what we call the practitioners, and then we have the portfolio construction side. So on the practitioner side, we have John McNeil, who, the president, former president of Tesla, former COO of Lyft on the General Motors board, an expert in kind of autonomy and AI. We have Sonny Madra, who is the president of a company called Grok, one of the leading, AI companies in the world. He's been in the space for many, many years. Their job is to identify those risks and those opportunities before hopefully other people see them.
So, they'll know a company in South Korea has all of a sudden developed a massive pipeline of AI projects across the world. Or, another company may be falling behind on their technology development in a certain segment. So we need to know that. So we need once that is identified, then it goes to the portfolio construction site comprised of myself and Professor Robert Whitelaw, who Professor Whitelaw was the dean of the undergraduate college at NYU Stern School of Business. He was the chairman of the finance department at NYU, leading expert in portfolio construction. So once we find these risks and opportunities, it kicks over to us and then it's our job to really identify how to affect that change in the portfolio, taking into account all the different risk factors that you want to look at to, optimize a portfolio.
So you alluded to this and this is right in your wheelhouse. So are you how are you weighting your holdings based on contributions to the AI sector in terms of, such as semiconductors or applications or infrastructure? Is that how you're kind of making a base weighting decision?
That's correct. So put aside the applications. That's not part of it right now. It's just the semiconductors in the data center. So, yeah. So we are first we are weighting them first by their contribution to the supply chain. So if it's a 25 percent weighting, if cooling systems are 25 percent of the data center, those companies in the in the cooling system bucket would get a 25 percent weighting. house benefit. Um, okay. And then what we're doing and then we're then we're looking a little deeper at the financials to determine how much revenue are they getting from, um, from AI projects. So let's say in the cooling system, there's $100 million in AI data center projects.
And there's four companies that have met our criteria. And one of those companies has $50 million of the $100 million. We will market share weight them within the segment. So the company that has $50 of the $100 million will get 50% of the 25%, if that makes sense. And the other companies will split the remaining based on how much revenue they're bringing to the table.
Yeah, that makes a ton of sense. And so that kind of core reweighting, is that on that semi-annual? Or are you a little bit more active, kind of maybe looking at quarterly results and reports to kind of tweak here and there as time goes on?
So we do an official rebalance of the core portfolio at the end of December, at the end of June. So that's part of the rules. But then we will certainly be looking at all of this on a weekly basis just to see how things are shaping up. we're not going to make active trades for the sake of being active. We're going to do it if we think there's a significant risk or opportunity that needs to be addressed.
So speaking about geographic exposure, does this go anywhere? Are you evaluating opportunities in emerging markets outside of the U.S. or is it pretty much just domestic?
No, we're looking globally. It's a global product. There's companies all over the world that are providing significant innovation into the space. About 60% of the portfolio is currently U.S.-based, though, followed by Taiwan and China, France, and so forth. So, there's market leaders in different parts of the world, and they need to be captured in the portfolio to get a good representation.
So I just want to get your take on this because I was reading one of your white papers, and it says in there, AI growth of up to 300x by 2032. what kind of key drivers do you think will fuel that growth? And what really, if any, risks do you see that could totally derail it?
Yeah, and frankly, that number might be off at this point. that was based on analysis from, probably six months ago. And you've probably seen all the news on all the CapEx investments being announced. it's absolutely an arms race right now. just a couple days ago, you just came and announced $200 billion. I think $150 of it was already announced, but they added another $50 to that. Amazon, all the Mag7 companies are, putting significant dollars in. So I think the CapEx numbers are even far in excess of what we anticipated just six months ago. And I think a lot of that is driven by just the national security interests around the AI space generally.
So what can derail it? look, there's always things that can derail, any sector in the markets, geopolitical risks. wars. anything could, derail all these investments being made. So you need to, of course, always look at that. But from a market perspective, we don't see anything that's derailing this. We think we're very early. There's a lot of heat in the sector. People have been talking about it. But it's really just getting going. We're just at the stage where we need to build the infrastructure to serve the needs of the models, the language, large language models that are being developed. So very early.
Yeah. And, as you know and mentioned, there's a lot of volatility in this space. Like, I know it's not related to your fund, specifically, but I saw what happened to quantum computing stocks just because of one, thing that came out that said they're, another seven or 10 years away rather than five. Right. And they all were down 40, 50, 60% in a handful of days. So we've definitely seen volatility in the space that you're playing in. And does AIS manage risk a little bit in your active stance? Or is it still more of a fundamental reweighting? You're not going to be trading just because there's headlines. So I agree.
Look, there is definitely look, look when DeepSeq came out, right? The whole news about DeepSeq a couple of weeks ago. the whole sector just got clobbered. Of course, it's all clawed back because I, people believe that was overblown. But we don't actively manage for volatility. However, if you look at the performance characteristics of AIS versus the other products in the market, I think because of our focus on the infrastructure versus some of the kind of front end applications, our volatility profile is quite a bit lower on average than most of the products out there or most of the ETFs out there. So and part of that is due to the breadth of the portfolio. Right now, actually, we're even higher than we were, the 72 number was probably before the rebalance.
I think we're at 85 right now. So we have a broad, a broad portfolio of companies across, quite a variety of sectors. And, that contributes to a lower volatility profile.
So with that being, that being said of kind of adding some additional names as this space continues to grow, I'm sure in your analysis, you're looking at. Existing companies in the portfolio to kind of evaluate their durability and and do they still have a competitive advantage in the space? Is that something you guys are doing on an ongoing basis as well? That's part of the constant process.
And that's really about finding those risks. So, we have to make sure that the companies that are currently the leaders, that their technology edge is still intact and that they their pipeline is still being grown from an AI revenue perspective. And then, what new companies are emerging that could take the crown in those specific portions of the supply chain that may be not on the radar of other people. So those are that's that's a critical piece of the puzzle right there.
Find those opportunities. I imagine. evaluating the current portfolio is easy. Trying to keep track of all the new issues and all the new opportunities, just given the growth and the excitement in the space might be the harder part of the equation. Yeah.
And that's why having the practitioners, we think, is a massive competitive advantage that I don't think anybody in the space has. having people, Sonny and John, that are in the space, have been in the space since the beginning and understand the landscape and who all the players are. And frankly, in Sonny's case is involved with building data centers around the world. So, he's they know who the players are probably before most people do because they're working with them.
Yeah. Makes a ton of sense to me. So million dollar question, you guys are sitting down, you're you're selling the fund. You might be sitting down with a financial advisory group or an advisor. Where where are you recommending kind of slotting an allocation to AIS?
Yeah, it's a growth equity product for sure. That's what the exposure provides. So, we I always position it as a as a satellite to your core equity. it's depending on what your risk tolerances are, of course, and what your your clients are looking for. But, it's three to five percent, slug out of your core equity exposure. If you believe in the space and you believe in the long term growth process.
Well, I don't know how is how as an investor and how widely utilized this stuff is becoming that you couldn't be excited about the space or want some exposure. But, Adam, I really appreciate your time before I let you go. I got to ask you, where can people learn more about Vista shares and find all the information on your ETF?
Yeah, our website is VistaShares dot com spelled just how I how it sounds. We have a nice presence on LinkedIn and on X. We are we have a great newsletter on LinkedIn as well. We're putting a lot of effort and time into content. We think content is crucial, educational content around the space so investors can really understand, what it is they're looking at beyond AIS, which is, of course, the ticker of the ETF. But it's really about education and, letting investors understand the space. And then hopefully they come to AIS when they want that exposure.
Well, again, I have a feeling, given your background and given the company, that probably not the first time we're talking. I know you're not going to be able to talk about any future issues that your releases or ETFs you guys may do. But I have a feeling there'll be some other opportunities for us to talk. But again, Adam, thank you so much for your time. Thank you very much. I appreciate it. Thank you. Thank you. Thank you. Thank you.
Daily Market Intelligence
The Signal
Brad Roth's daily market brief — systematic signals, ETF positioning, and what the data is actually showing.
Subscribe Free →