FidelityConnects: Live with Mark Schmehl

[00:03:40] Pamela Ritchie: Hello, and welcome to Fidelity Connects. I'm Pamela Ritchie. The momentum behind AI continues to redefine how companies operate yet we're seeing a notable shift in sentiment as investors reassess the pace and the durability around software and AI applications. At the same time powerful new use cases keep emerging from automation that's slashing operating costs to models that are reshaping how global firms make some of their decisions. Are the massive infrastructure investments finally nearing some real returns, and with global adoption accelerating could the next wave of opportunity come from beyond the US tech giants? Happy to say that joining us here today to break it all down is portfolio manager Mark Schmehl. Mark manages several Fidelity funds including Special Situations, Global Innovators, and the Canadian Growth Company fund on which he's marking the 15th year as portfolio manager. A friendly reminder that our discussion here today features live French, Cantonese and Mandarin audio interpretation so join us in any of those languages. Welcome, Mark, great to see you. How are you?

[00:04:50] Mark Schmehl: I'm great, Pamela, how are you?

[00:04:51] Pamela Ritchie: I'm very well, thanks. It's early on the west coast. You're on coffee number...?

[00:04:57] Mark Schmehl: This is coffee number two.

[00:04:58] Pamela Ritchie: Yeah, that seems reasonable, seems reasonable. Great to see you. Thank you for joining us here. Let's begin with a little bit of the discussion of safe havens around the world just because that seems to be what's grabbing a lot of attention. Your fund, in fact, tries, you, try not to pay attention to a lot of the macro things that go on. You stay focused, stick to your knitting. Is it actually a safer haven to be investing in some of the things that you're looking at because they don't take into account what's going on?

[00:05:27] Mark Schmehl: In a strange way I think that that is actually true. I spent the last two weeks literally talking to everyone in the space who does AI, AI investing, the AI model companies, you name it, and it's really independent to a certain extent to what's going on in the Middle East with oil and the economy and politics. It's one of those things which it's so transformative to everything that we are going to touch that the companies are going to invest. The biggest, most capitalized companies in the world are the ones doing the investing and they really don't care about the oil price going up 10 bucks or 15 bucks. This is a fundamental generational investment thing so they're just going to invest. So it is safer to be in this space than it is in a lot of other places in the markets.

[00:06:31] Pamela Ritchie: That's kind of fascinating. You are known for, when you need to, selling, moving on, taking the sell strategy that moves quickly. Your investments move quickly by the very nature of them. Have you sold a whole lot or you've been having a lot of turnover of late because markets certainly have.

[00:06:49] Mark Schmehl: I am trading less than normal, way less than normal. I'm not doing anything. Typically in a volatile period I'll be moving the funds around, trying to take advantage of ... I'm not doing anything.  I was talking to a colleague the other day, he's like, why are you doing so well? I'm like, I have no idea. I think it's just that I want to own the things I want to own because they have the strongest fundamentals and those fundamentals aren't being impacted by what's going on. Now, if this goes on for a year, well, okay, maybe we'll have a different discussion but at the moment I just see zero, like zero impact to anything I'm investing in.

[00:07:27] Pamela Ritchie: It's fascinating. Bring us up to date a little bit on all the things that get discussed within AI. We won't be able to hit all of them. Are the models at this point, they're all being fed, they're all being made smarter, they're all moving along at rates within companies to the point where it sounds like some of them are getting to be accretive in sort of the halo, the high, the heavy asset and so on. Big companies that need to be transformed. Just bring us up to date sort of that world, and then there are many other stepping stones as well.

[00:07:56] Mark Schmehl: It's interesting. I had this discussion with OpenAI just last week, we had dinner.

[00:08:01] Pamela Ritchie: What did you eat?

[00:08:03] Mark Schmehl: Oh, it was Italian and it was not that great as corporate dinners generally go. The point I wanted to make to them, and I've made the same point to Anthropic as well, if they want to pay for their compute capacity they need to show the world how fast they're growing and what their business looks like. The only way to really do that is to go public. I think this is one of the disconnects we're currently seeing. If I look at some of the memory stocks, for example, they trade at 2, 3 times earnings. These are companies that are growing really quickly, have great balance sheets, lots of cash, and they're trading at 2 times earnings. The market is saying, this is it, it's all over. You shouldn't buy these things, it's the peak, it's the peak, it's the peak. These companies cannot go out and invest in capacity to grow their business. The only way to convince the market that it's not the peak is the model companies have to show what their business is doing. I can see it. I'm a private investor, I see their numbers. It is astonishing how fast some of these companies are growing. It looks like Anthropic is growing 30% week-over-week, week-over-week.

[00:09:16] The math is so crazy that it's almost going backwards now. The problem is that the stock market, or the public markets, aren't seeing this. They just don't see it. They can't see it. And they don't see where it's going. In order to get the public market willing to say, hey, we have to invest in all of this compute infrastructure you need those private companies to go public and convince the market that this is not a peak, because so many investors are trained to say, no, this is the peak, it can never get any better than this. It's like wait, no, it's not.

[00:09:53] Pamela Ritchie: I mean, part of the ability to get exposure to those companies, which none of us can see except for you, is via your funds. Why don't you just talk a little bit about the private part of the funds that you invest in. You've spoken about how well you can see what's going on back there but at the same time what portion does it make of the fund itself?

[00:10:16] Mark Schmehl: It's funny, the two Canadian funds it's about 6.5%, almost 7% of the portfolio now. The same sort of exposure in the US, the closed-end funds, are getting four times the NAV, which is crazy.  Americans are chasing these things like you wouldn't believe. Canadians [indecipherable] my funds can get this exposure for free, which I think is an amazing deal for anyone who wants to buy Canadian Growth Company or Special Sits. They just have so much, Anthropic and SpaceX, all these things, all the fancy stuff that you just can't get anywhere else. They're fantastic. I think Anthropic is about 2% of all my funds right now. SpaceX is similar. I think the prospects for both companies still look fantastic, especially for the AI companies. Oh, one of the things the AI companies ... you can use your model to do is pick your bracket. I don't know if you read that today, it was in the Wall Street Journal.

[00:11:21] Pamela Ritchie: No, I didn't.

[00:11:21] Mark Schmehl: Claude apparently has an excellent chance to win the Wall Street Journal March Madness Pool.

[00:11:27] Pamela Ritchie: Yes, I did see that, actually. That's one of the many things that you could do, one of the many developments. Well, it's super interesting. Things are moving very, very quickly and the things and the ways that we can incorporate it into our lives but it's really the companies themselves. Do you get to the point where you're ... you're investing in companies at 2 times earnings, why would you go anywhere else? I guess the question is ... because it's cheap. It's like a value play. Other companies that are putting it to work, that's happening? The accretive piece, the productivity piece, I mean, this seems to be ... what do you see?

[00:12:01] Mark Schmehl: Absolutely. Companies are definitely... especially in coding. I mean, if you were a software engineer you don't code anymore, like, you just don't. You use Claude Code or you use Cursor or you use Codex. It just does all the work for you, all the grunt work that you don't want to do. Folks are now experimenting with the agents that are showing up on our desktops like Claude Cowork. We've been messing around with Claude for Excel. We are all going to use this stuff and we're going to use it aggressively. It's very funny to watch as it rolls out. The first phase is, this doesn't work. The second phase is, I hate this, this is going to take my job. The third phase is, holy cow, I can do a lot of stuff. I guess the next phase after that is companies are like, great, you can do a lot of stuff, let's do more. We're seeing this journey from the executive suite which is we're going to use AI to improve efficiency and then we're going to fire people, and then they go, wait, we're not actually going to fire people, we're just going to do more stuff with it. I think that the last part of that journey will be now we can grow faster.

[00:13:07] That's a big part of the thesis I have for the whole economy barring what's going on in the Middle East, which I can't quite figure out why we're doing this, but I think that the economy will grow faster with AI as an input because labour has always been the big bottleneck, right? There's only so much you can do. I think there's still only so much you can do with the big physical things but with respect to office work and intelligent work or whatever you want to call it we're removing a bottleneck which is human productivity, and we are going to be able to grow faster. I think the economy's going to actually grow faster and have a higher speed limit for a period of time as we start implementing AI throughout our ecosystem of apps and processes.

[00:13:55] Pamela Ritchie: Can it save us from sort of the short term ... higher interest rates are being discussed and certainly priced in--

[00:14:02] Mark Schmehl: Hundred per cent.

[00:14:03] Pamela Ritchie: --and the entire inflation story? Can it power us through the other side?

[00:14:07] Mark Schmehl: No inflation. This is the stupidest thing I've ever seen. I think that folks are like, oh, yeah, we're going to have inflation because oil is up, and that has historically been true, but we have this productivity boom coming at us which no one has seen before and it is going to make every office worker 5 to 10 times more productive. The numbers you see from the software engineers, they've got like 6 or 7, or in some cases 17 agents running. It's all being limited by compute power not by their imagination. We are going to be so productive. There's going to be zero inflation. This whole rates are going up thing is total garbage. Oil's going up, yes. Labour is the biggest part of any inflationary cycle every single time. Labour prices are not going to go up, they're just not, because we're going to have AI in here substituting for labour and there's just no way. For me, there's no way interest rates go up. I'm betting my entire fund that there's no way interest rates go up. I do believe in higher structural energy as a result of what's going on in the Middle East and that may or may not be good for energy companies. I think that the overall economy is just going to grow faster with lower inflation for a long time. That's my bet.

[00:15:25] Pamela Ritchie: That's an amazing bet and it's incredible. You're smarter than most of us so we'll see how that goes. It sounds right. How, ultimately, on the energy compute side of things, I mean, if interest rates don't go up then it's, in theory, easier for any company to continue borrowing to build that out and to get there along the track that they think will be tricky and bumpy and maybe take a while, but it's not going to be suddenly upended by higher interest rates. Is that right?

[00:15:52] Mark Schmehl: I don't think so, no. I really do believe ... and also it's interesting, the energy thing. Energy is an engineering problem. When you're building a data centre you are solving for your energy prices and what you're trying to get is more tokens per unit of energy. As you implement new technology, whether it's CPO, whether it is the latest Nvidia machine, whether it's better networking, everything is a result of token per watt. That token per watt curve continues to come down. I don't think investing in utilities, electricity, I mean, while we need it I think we're going to continue to solve for it. Like your fridge, your fridge is so much more energy efficient now than it was 50 years ago. That's going to be true for every data centre built. I don't think energy's going to be as big a rate limiter as I believe the market currently expects, but I could be wrong on that, because every part of the engineering compute cycle is designed to reduce that energy component.

[00:16:55] Pamela Ritchie: Okay, that's fascinating. An area of your portfolio has always, I think it's usually, or usually included gold. That's been kind of a fascinating thing to watch over the course of the last year and very Canadian. It's powered the TSX, certainly. Is that something that you can continue ... I mean, is there a way that you hedge within the portfolio? I mean, that's not always ... you're not really a balanced fund but at any rate you do use that sometimes.

[00:17:21] Mark Schmehl: It's very difficult. I run all these different products and mandates. In Canada gold is a huge chunk so I have to own it. I have always owned some in my innovation portfolios as well as a hedge, it sort of trades inversely to NASDAQ, except now it doesn't.

[00:17:39] Pamela Ritchie: No, it doesn't. It's a risk [crosstalk].

[00:17:42] Mark Schmehl: It no longer does that. I've been sort of reducing that as a component of that strategy in my tech funds. With respect to Canada, I still own a lot of gold. I don't think that the rationale for owning gold has changed in any way. Government deficits worldwide. There's so many reasons to own it but I think that it has now started to trade with technology, so its old reason to own it as a diversification tool is gone. It's less compelling to me than it was and as a result I own less of it than I used to. But I will always probably own a little bit of gold because it's just a very interesting asset class for me.

[00:18:24] Pamela Ritchie: Back in November were sort of the first rumbles that the pieces of the stock market that would be disrupted, you know, software and so on and it went through to insurance and every other industry that I'm sure you saw coming six years ago. Within that how much is overdone? This is a question. It's an interesting moment because markets are getting sideswiped for lots of different reasons. A month ago we were talking about AI disruption in everything. Is it an interesting opportunity to go and buy up software? Take us through software.

[00:18:55] Mark Schmehl: No. No, software's a sell. Full stop.

[00:18:59] Pamela Ritchie: Okay. Really. Really interesting to hear that.

[00:19:01] Mark Schmehl: I mean, I don't want to be a doom ... if you're in the software space I'm not trying to be a doom and gloom guy but it is being transformed and the market doesn't know what the business models look like and we can sort of see how the software business model is changing. It's going from labour was the big constraint because knowing how to code was really, really difficult and you had to have all these smart people to do it, we just got rid of that. You're going to replace all of that with compute. You're basically going to pay for tokens and you're going to use that to do your coding. The business model is no longer the same. It used to be really high margins, high terminal value, it is now moving to a low margin commodity business with bad terminal values.

[00:19:46] Pamela Ritchie: Commodity business, so interesting.

[00:19:47] Mark Schmehl: Well, yeah, it's a commodity business. While that's going on the stocks are all sells. They're all just going to keep coming down. It could be a five-year process. I have a few software analysts internally, they've been covering this space for 15 years, they're all like, it's cheap, it's cheap, it's cheap. I say, no, no, it's not the same, it's a different business, just avoid it. I think you can pick and choose within it. Certain companies like Canadian Shopify is not really a software business, it's an e-commerce platform and they get paid by payments so it's more of a payments company. It's sort of being lumped in with software. There's some names within the software space that aren't really software but in general that whole sector is just not worth investing in, at least not at this time.

[00:20:34] Pamela Ritchie: One of the other pieces, and I think this is where you are invested, in sort of pieces of the infrastructure buildout. Then, of course, you get these questions of are we overbuilding, which is part of the peak discussion in AI and so on. Could we possibly overbuild? We go back to the fibre optic cables that were laid and too many and too expensive and still paying it off and so on, is the analogy true?

[00:20:57] Mark Schmehl: It's not a good analogy because we built a lot of that fibre and we didn't have a use for it. It's sort of like Bitcoin. Hey, here's this technology but we don't know what to do with it. With this compute that we're using for AI, these companies could grow so much faster if they had more compute because the use is so powerful and it is being used for so many different things. Every time you install a box it's sold, and it's sold at a profit. This is not the same thing. We should be thinking about token production like oil. You can never have enough energy. You can never have enough intelligence. We will never have enough.

[00:21:40] The question will be, ultimately, what do we pay for it, what kind of return do we get on it, but we're so early in this cy-, we're still early that we don't know what that is. You just know it's going to continue because we need more intelligence more quickly. The fact that the model companies are growing at the speeds of which they're growing at such scale, I mean, Anthropic's a $22 billion revenue company growing 30% week-over-week. I've never seen that. I don't even know what that looks like. I don't know what the market puts on that. What kind of multiple do you put on a company that's growing that fast? They're growing that fast because other companies are using their stuff to do all the things they need to do.

[00:22:21] Pamela Ritchie: And governments, too. Governments are using this, everyone is using this to do what they need to do, whether it's a charity, all levels. Take us into a sector that I know that you've invested in and liked over the years. Healthcare is something that is being disrupted. AI is very good for healthcare for lots of research reasons, you can see why, and within hospital situations, personnel, so on. Just take us through any sector that is not really being affected by this, or choose one that you'd like to tell us about because I'm dying to hear how you see it being deployed in a particular sector well.

[00:22:58] Mark Schmehl: We are not seeing the outcomes yet that we expect. We're hearing a lot about we're using it here. Procter & Gamble used it to launch a product, they launched it in, like, I think it was four weeks instead of six months, the time to market for a new product. They used the AI to build the chemistry for the new shampoo and design the product labelling. That's an example where you just shorten the time to market using your AI. We've talked to lots of drug companies who are using it in their research, it's letting them do more research more quickly and screen more things. Is it creating new medicines? It probably will but we haven't seen it yet, it's a long cycle. Governments are using this to do reviews faster. Everything speeds up, I think is the answer. Everything speeds up and things happen faster, products get better faster, as a result your competition gets better faster so you need to race to continue to do all that. It's impacting everything but it's hard to point to, oh, this is this magic AI moment.

[00:24:07] There are sectors of the market where I don't think that AI is going to change a whole lot. Those are those big, dirty physical sectors like, you know, energy is not going to ... it's going to change on the margin a little bit. Mining, transportation, you're still going to need to ship stuff around. A railroad's not going to really change from AI, right? You're still going to ship grain and logs and it's all going to go on a train. There are many parts of the market where AI won't really disrupt it. In some respects I think those parts of market may turn out to be the new bottlenecks because AI efficiency will not improve the speed at which a railroad drives around. If my theory on a faster economy because labour productivity goes up, those physical bottlenecks that you're going to hit in those old, dirty industries that you can't really change that fast. That's sort of what I'm thinking about longer term, those might be the new bottlenecks that we want to invest in. That's probably a three or four year out problem but something I'm thinking about.

[00:25:11] Pamela Ritchie: So X company wanting to launch a product can do that a whole lot faster but getting space on the rail car to get it across the country or continent will still take the time.

[00:25:21] Mark Schmehl: Doesn't change.

[00:25:23] Pamela Ritchie: In fact, as you say, new bottlenecks ultimately. 

[00:25:29] Mark Schmehl: One of the companies I really like is Amazon. When you think about it, Amazon has built this huge physical network of trucks, shipping, and warehouses, and old physical stuff. Hard to fix, hard to make better. AI doesn't really ... I mean, it helps a little bit on the margin but you still need the big warehouse and all the trucks. On the other side, they also sell all the compute. They have this AWS and they're selling all the compute capacity, they're participating in the AI cycle. For me, if you think about big companies that could benefit, they have this big, sort of dirty retail business, this is going to not change much, and then they have this AI AWS business which is going to start growing faster. These are the types of stocks I'm trying to find and it looks compelling to me. Whether that works or not, I don't know, but it really looks compelling as a mix of businesses.

[00:26:24] Pamela Ritchie: That's fascinating, as a mix of businesses. The next question was a bit about hyperscalers but you can take it to other companies as well that are investing and making as part of their business heavy, dirty, whether it's nuclear, whatever it is, they're trying to incorporate power into their actual umbrella of their company itself. Are they still tech companies or are they more mixed use? What is the future of that? It sounds like an interesting place to watch, and to have you watch it for us, but does it mean something for investors that the companies they think are in certain sectors are now almost multi-sector?

[00:27:00] Mark Schmehl: I think each hyperscaler is unique in its own way. I like Amazon because of the retail and whatever. Microsoft has a huge software business which I think is going to zero, but their other business, their hyperscaler business, will be fine. Google, they have a hyperscaler but Google is its own special unique animal that we could talk about for four hours. Each hyperscaler is a little bit different and needs to be thought about differently. I think it's interesting that in many ways they trade as a unit but they're actually very different stocks with different asset mixes.

[00:27:35] Pamela Ritchie: Is it the end of the Mag-7 then because it needs to be?

[00:27:39] Mark Schmehl: I don't think so but I think you need to stock pick your way through it. I don't think you can just own Mag-7. The S&P 500 just owns the Mag- 7. I think we should be able to beat the S&P by owning the right ones of this. So far this year not owning Microsoft has been a really great idea because it's been terrible because of software. I don't know if that continues or not but I'm just saying you can definitely add alpha by picking your way through the Mag-7, which I don't think was clear even a couple of years ago.

[00:28:13] Pamela Ritchie: No, that's really interesting. There are many points to point to but if you were to talk about sort of a new pinnacle, it's when the names that you've mentioned, the OpenAI, SpaceX, and so on, when they go public. Can you just give us a little bit of a runway view on that? That will obviously be an incredible moment when that happens. We don't know when it'll be.

[00:28:35] Mark Schmehl: I hope it's this year. SpaceX is its own unique animal and not really related to AI. It's part of the whole Elon complex. Again, that's another 10-hour discussion. I think that OpenAI and Anthropic going public, and if the numbers that you're putting up now continue into that, I think that will be a seminal moment for the market, sort of like a singularity and everyone will go, oh my God. I hope that happens this year because they need more compute capacity and the only way to finance it is to get this market to pay attention to the fact that it's not a bubble, that it's actually needed. You literally need every single token to do all the work that we need to do, it's just astonishing how little the market cares. We need to change that narrative.

[00:29:25] Pamela Ritchie: I'll end with macro just briefly, though, do you think on the other side of whatever it is we're in the midst of right now there's actually an acceleration towards some of these themes? As you say, maybe wiping a bit of the doubt away because this is where the market needs to focus. Is there, again, maybe not a seminal moment but almost a clearing of fog of war and what is actually the structural everything.

[00:29:50] Mark Schmehl: I think so. I think we need to get there. We need to resolve whatever is going on in the Middle East right now, however that's going to get resolved. The most important thing is, and I do tend to believe it, we probably are at artificial general intelligence. We probably will hit it this year or next. We will create a sentient AI. I think you're going to see  and so does everybody in the industry. That is going to change the world in fundamental ways that are way more important than whether or not we are bombing Iran. I think folks need to start to think about it and what does that mean to the market, what does it mean to their professional careers? I think it's going to be a big, big moment and we need to starting thinking about it. I'm thinking about now. I always tell people I live in the future, I'm living in that world where we have this sentient AI and it's a little bit strange.

[00:30:45] Pamela Ritchie: That'd be nice to live in the future. Today's not so great anyway in a lot of different ways.

[00:30:48] Mark Schmehl: No, I know, right? It's really tough out there. It's really tough to read the news these days. Anyway, I think that's what we're going to see.

[00:30:55] Pamela Ritchie: Okay, this may dovetail with this, final question coming in from one of the advisors, how should advisors communicate market volatility declines in a way that builds confidence rather than fear? I actually think you've pointed to a number of those but just to sort of put it into a final thought.

[00:31:12] Mark Schmehl: One of the things, experience helps a lot. When I was a junior portfolio manager I kind of freaked out a lot faster. Having been through the wars many times now I have a lot of scars and I really know what to do. The key is to sort of react quickly and then underreact and think about longer term. Where do you really want to be in terms of your portfolio and your investing, and try and just sail through the storm as best you can making as few mistakes as possible. That's really what I try and do every time, is cut down the number of mistakes that I make as I go through this mess.

[00:31:53] Pamela Ritchie: It's inspiring to speak with you, Mark Schmehl. Thank you for joining us here today.

[00:31:58] Mark Schmehl: Thanks for having me, Pamela.

[00:31:59] Pamela Ritchie: All the best to you. Mark Schmehl joining us today from the West Coast. Coming up on Fidelity Connects tomorrow, you can join our Directors of Tax and Retirement Research, Michelle Munro and Jacqueline Power, for a practical and extremely timely walk through of what you need to know to guide your clients through the final stretch of tax season. Jacqueline Power will join us first at 10:30 a.m., that's Eastern time, for a French language webcast and then she will join with Michelle Munro at 11:30 Eastern time, that's in English.

[00:32:27] On Thursday Fidelity Director of Quantitative Market Strategy, Denise Chisholm, is back with her latest sector thesis and the key signals, boy, we need some guidance on key signals, that advisors need to be paying attention to for the month ahead and months ahead.

[00:32:43] On Friday portfolio manager, Aneta Wynimko. She offers a refined look at the latest shifts in customer taste. She has beautiful taste. We'll hear what she has to say, brand power as well while unveiling the investment themes capturing her attention in the Fidelity Global Consumer Brands Fund. Thanks so much for joining us here today. We'll see you soon. Have a good day. I'm Pamela Ritchie.