Polar Capital Technology: performance, SpaceX and Al maximalism

Polar Capital Technology lead manager Ben Rogoff reflects on the stellar returns the investment trust has delivered in the last year and the outlook for the future in the AI era.

He explains why being an AI maximalist is so critical, responds to views of the trust as ‘benchmark-aware’, and discusses buying SpaceX and the end of the cloud/smartphone world.

• Our Services: SIPP Account | Stocks & Shares ISA | See all Investment Accounts

Dave Baxter, senior fund content specialist at interactive investor: Hello and a very warm welcome back to our Insider Interviews series. I’m Dave Baxter here at ii, and today I’m joined by Ben Rogoff, lead fund manager on the Polar Capital Technology Ord (LSE:PCT) Investment Trust. Ben, many thanks for joining today.

Ben Rogoff, lead fund manager on the Polar Capital Technology Investment Trust: Thank you.

Dave Baxter: The trust has enjoyed some really stellar returns both in absolute terms and versus the competition in the last year or so. What has driven that, and would you expect that to continue further? 

Ben Rogoff: Well, it has been a very good period for the trust and for the tech sector generally, but most importantly for the AI story, which has just continued to make huge gains. When you look at the performance of 2026 to date, we’ve had a very strong first half of the calendar year because we’ve seen the agentic part of the AI story take off.

So, if we went back and looked at the short potted history of the AI story, we’re kind of four years old. The first three and a bit years we’ve spent basically having to argue that this was going to be really important, and this is why people were investing and there were going to great returns and products down the road, but those things didn’t exist. The products were inferior or imperfect. There was really very little in the form of revenues to talk of. 

Then, in 2026, we’ve seen the explosion of revenues at the frontier labs, at Anthropic, at OpenAI. We’ve seen continued growth in capex, and we’ve really started to see the models get to levels where you can see the actual revenues coming from coding and other use cases that we’re seeing. 

And now, more recently, the excitement around agentic AI. And that’s resulted in, as I say, a kind of reappraisal of the AI trade and a very strong period of returns for stocks that are exposed to that theme. We have probably more exposure to the AI theme than most. We describe ourselves as AI maximalists, and we’ve really been of that view now for three and a half years. 

How do I feel about it? Well, I’m obviously really pleased that, so far at least, the positioning of the trust has allowed us to capture a good portion of the returns. Look, everyone knows that past performance isn’t a guide to future performance. It’s worth just repeating that. 

Since the summer, actually, there’s been quite a lot of buffeting, and I think that investors should be aware that with new cycles, that’s typically what happens because we’re talking about something that’s going to happen, or at least the revenues have arrived, but the real promise of AI remains to be seen. 

As a result of that, and because of that uncertainty and that time horizon, there’s going to be more buffeting along the way. The way I like to describe it is, the faster you travel, the more bumps you get on the way, so the turbulence you should expect. And I think that that’s where we are. 

Going forwards, we’re still very positive about AI. In the end, outside of that volatility, and again, I can’t extrapolate past performance, but what I can say is that the growth profiles of the companies we’re investing in look good. There’s a general level of overall scepticism about AI that is healthy. We’re not at a point, in my opinion, that we were in the late 1990s when everyone’s in the trade. We are not putting what we would describe as peak multiples, peak price/earnings (PEs), on peak earnings. There’s a lot of strong earnings growth that we’ve seen in the portfolio, but the market has been actually fading the forward PEs as the earnings have gone up, and I think that reflects a sort of general level of skepticism that an industry that’s gone from nothing to $1 trillion of capex can continue to grow. 

We think it can, we think that models can continue to get better, they open new market opportunities, and we’re not really even imagining yet what AI might achieve once we start to see real evidence of non-human scaling. 

• SpaceX shares tipped to rocket after latest results
• Polar Capital Technology: the AI stocks I’m excited about in cognitive era
• Anthropic IPO: what UK investors need to know

Dave Baxter: What stands out to me is that the trust has much higher exposure to Asia than many of its rivals. But, interestingly, you still do have around 67% of the portfolio in the US. How do you expect that mix to shift over time? 

Ben Rogoff: That’s a great question and it just reflects this like pro-AI portfolio that we’ve had in place, and really going to where we see the best opportunities to profit from that. 

You reference South Korea, that’s because it’s a very memory-heavy market: Samsung Electronics Co Ltd DR (LSE:SMSN), SK hynix Inc ADR (NASDAQ:SKHY), we’ve also found some smaller names that I’m not going to share today because I’m really proud of them and I don’t know if others have found them. 

But bottlenecks around materials, chemicals, bits and bobs that you need that you didn’t need in the cloud world, many of the companies that dominate those small niche markets are in Asia. 

That’s not to say you can’t find great semiconductor companies in the US. Obviously, NVIDIA Corp (NASDAQ:NVDA) is, without peer, discussed. But we also have companies like Advanced Micro Devices Inc (NASDAQ:AMD) in very big positions, in the semiconductor capital equipment industry there’s US names that are world -eaters, but there are also very interesting companies in Japan and elsewhere. So, we’re relatively agnostic.  We just want to make sure that we find the best opportunity, perhaps market share gain, or the best priced asset, but, ultimately, the right exposure to the bits of the market that we’re excited by. 

Longer term or medium term? I think that there’s a downward pressure on the US weighting that comes in the form of AI disruption because, let’s be honest, the US has been the dominant geography in a cloud/smartphone world, and I think that we’re coming to the end of that cloud/smartphone world, we’re in an AI era. That doesn’t mean, however, that the US is going to lose, and if you look at where the frontier labs are, they’re all in North America. 

We bought Space Exploration Technologies Corp Class A (NASDAQ:SPCX) in the portfolio in a relatively modest shape. Again, that’s a unique non-fungible asset that is based in the States. As and when this IPO market more fully reopens, and as we get to refresh the US stocks, there is no reason why the US couldn’t be 70% or 75% again of trust assets. 

But, right now, absent SpaceX and a handful of other companies, we really haven’t seen an IPO market with a kind of recharged AI-native list of companies coming public. If that happens, there’s very good reason to believe that we could rebuild the US, but in the absence of that, we will just continue to look for the best opportunities globally. 

Dave Baxter: It’s interesting that you mentioned SpaceX. You did pounce on the shares when they IPO-ed earlier this year. You also mentioned OpenAI and Anthropic. And I suppose an interesting trend has been that companies have made many of their gains in the tech space while they’ve still been private. How tempted do you get to hire a specialist and start investing in unlisted companies? 

Ben Rogoff: A couple of things, I love the word ‘pounce’ because that sounds like we’re all at our desks poised. I mean, we’ve obviously been doing a lot of work on SpaceX for quite a while.  

I think it’s worth reminding people that the trust, since I’ve been managing it, has exclusively focused on listed companies. The reason we do that is because, ultimately, there’s really nothing worse as a fund manager than being stuck in stocks you don’t want to be involved in, and things in the tech sector change very, very rapidly. So, we have a very different approach, a very different level of turnover, a 100% turnover for probably every year that I manage a trust is unusual relative to perhaps other investment vehicles or other managers. 

So, immediately being in a stock that doesn’t trade will sit uncomfortably for us given that our investment process is inherently active. That isn’t to say we’re not aware of the companies or we’re not doing the work. So, we met with SpaceX when it was private and when we had the opportunity to invest in it, we did. We’ve been meeting with OpenAI and Anthropic, as you would expect us to do. Again, we’re a large team running a reasonable amount of money on behalf of people, and so we get access to companies before they’re public.

Should we be involved in them earlier? Well, obviously when you look at what’s likely to happen with Anthropic and OpenAI, and already has, of course, if you think about where these companies are raising money in private markets, obviously we would have loved to be participants in that. But, in the end, you can’t just rifle shoot in the private market. You have to commit a large amount of time, a large amount of resource, and also it is very much a shotgun approach. No one is going to allow you into a deal at the last minute in order to invest in a great business. You have to be there for the duration, through the corporate life cycle. You have to, I suppose, play the game. 

And that isn’t where we think the best effort-to-output ratios are. So, we’ve been able to deliver, and I’m really delighted with the returns we’ve able to deliver. We’ve been able to do that 100% in the public market. And when scenarios change, should the AI story change for the better or the worse, we can make changes to the portfolio that are unavailable to those that are in private companies. 

One last thing, if I may, I think it’s just worth reminding people that for every Anthropic and for every OpenAI, and again, I am a bit jealous that we don’t own those stocks in the portfolio, there will be many other companies you’ve never heard of that have perhaps the vintage of those years of private investment that may have just gone off, gone sour. Because those companies were interesting in a pre-AI world, but are much less interesting today. 

So, we’re very happy with the investment universe that we have. We try to make the most of it, and yes, there will be days where we wish we could have been in a private company. 

And then one last thing: SpaceX. You said we’re insensitive or something, I think, to the valuation?

Dave Baxter: Relaxed.

Ben Rogoff: We’re not relaxed about any aspect of investment. I think that we see ourselves, and we’re very hard on the idea, as stewards of people’s capital. I have a reasonable amount of my own money, as you would imagine, in my own portfolio, but even if I didn’t, my job is to try and deliver the best risk-adjusted returns I can. And the only way we can do that is to work extremely hard, be incredibly competitive, and understand the core drivers of outperformance. 

So, just quickly on that valuation point. When Google [Alphabet Inc Class A (NASDAQ:GOOGL)] became public, it became public on 35 times forward earnings, and there was a general level of scepticism that that was good value. Very few people, I’m told, went for the deal. 

In London, we bought it at IPO. It traded on two, two and a half times the valuation of Yahoo at that point, and the rest is history. 

I think that we have a great analyst covering SpaceX. He’s built a fabulous model, and already since the company’s become public, the ex-AI part of that story, so the AI part, the clusters that they’ve been building and the models that they’re building, have been radically re-valued by the market compared to even where we stood a year ago because that’s how fast things are changing in AI. 

So, we are relaxed in a sense that no one can know the future. We also know that the returns to tech companies rarely have much to do with the starting point on valuation. If you’re in the right company at the right time in history and has the right product, earnings numbers that are always modelled in a sort of linear process or progress, ultimately, can be humbled by what actually happens. So, SpaceX is a 1% to 1.5% position, as I say, in a 90 stock-strong portfolio. 

• Earnings beats fail to fire up S&P 500 index
• Sign up to our free newsletter for investment ideas, latest news and award-winning analysis

Dave Baxter: When people try to explain how the trust differs from the competition, they often describe it as being more ‘benchmark-aware’. What does that actually mean in practice, and where might you view yourself as being less benchmark-aware? 

Ben Rogoff: Of course, so active share is essentially how much of the portfolio is unexplained by the benchmark that you follow. So, if we replicated the benchmark stock for stock at the same position sizes, we would have an active share of zero. 

Over time, our active share has tended to be somewhere between 40-50%. The reason for that is that obviously our benchmark is a tech benchmark. So, within that benchmark, there are some phenomenal businesses that we want exposure to. So, we don’t actively go out to try and find names that no one’s heard of. I mean, if we do, that’s great. What we’re trying to do is deliver the best risk-adjusted returns that we can, and if they are in names that are well-known and large parts of the benchmarks, so be it. 

The Magnificent Seven has been well represented in PCT for a long time, when perhaps others were looking for better, more novel, more interesting stories that unlikely kept up with that incredible cohort of stocks.So, 40-50% is the right way to think about it. 

Now, as a company starts to become less interesting, if we think that an underlying company may deliver negative returns, we’ll zero it in our portfolio. So, we don’t own anything that we believe will lose our shareholders money. 

So, while we position size some of those Mag 7 and big benchmark names with reference to the benchmark, so we like this company but we don’t like it as much as the benchmark, so let’s take an Nvidia. 

Nvidia is the largest position in PCT today. It’s an underweight relative to the benchmark even though we can actually own more than 10% in the trust. We could own an index weight up to 15%. The reason we do that is because we think it’s a fabulous business. We think it is very well priced and attractive, and again not investment advice, but that’s just how we’re thinking about the stock today. 

But there are other stories that we might prefer on any given Sunday, and right now we’re overweight AMD relative to the benchmark position. So, this idea of using a benchmark as something to essentially baseline what your position size would be. If you don’t like something, zero it. If you like something, but perhaps not as much as the index, or you have other ideas outside that you prefer, then that’s likely to end up being an underweight position in the portfolio. 

And then your question around where do we actively go out, where isn’t the benchmark? I think that’s an interesting question because obviously the benchmark is very US-centric. So, just about anywhere and everywhere we go outside the US, we find ourselves swimming in relatively, I suppose, un-busy, quiet waters. So, Japan has been a place that we really enjoy looking for companies that are undercovered but perhaps have an interesting angle on the AI story. 

Korea as well, beyond the very largest companies, there are some interesting second liners.In Europe, there have been some more interesting second liners that we’ve been invested in more recently. 

So, in the end, it’s trying to get that combination of the very best the index has, as well as our very best ideas that may or may not be within the index, and trying to bring it together to deliver the best risk-adjusted returns we can. 

Dave Baxter: It’s striking that in the tech sector, we can see some really abrupt massive gains for a stock when a good story comes through. There are many examples I could name, but one recent one is Micron Technology Inc (NASDAQ:MU). Are there any ways in which you can see this happening and capture that momentum? 

Ben Rogoff: Look, in the end, I’m a believer that markets are reasonably perfect for most of the time, or at least the US market is relatively perfect, apart from moments where it just doesn’t work and financial crises [happen], or whatever it might be. So, I think it’s unlikely that we’re going to be able to identify something before anyone else, even though we’re trying very hard to do so, with 12 people and 1,000 company meetings. 

We have a member of the team in Asia probably every month. That’s where the compute clusters, the servers, and all of the sort of nuts and bolts and how they’re being assembled is really being done. So, we feel like we’ve got as much chance as anyone to be in the right places at the right time. Anticipating those is very difficult. 

But to give you a sense of what we are trying to do, say, for example, with the memory trade, I would say that we were relatively early to that, in that we identified that high bandwidth memory was going to become, or would potentially be, a bottleneck as models got better and as more compute clusters needed to be linked, and as the models themselves became larger. So, I’d say we were reasonably early to that story.

One of our biggest wins this year has been SanDisk Corp Ordinary Shares (NASDAQ:SNDK), which is NAND flash memory. So, lower down the memory stack, stuff that you used to think about as storing your photos on your iPhone has now become a critical component in AI systems, particularly in persistent memory and on the inference side of the story. 

Actually, I have to say we were very early to that stock. My colleague, Nick Evans, saw the company, it sounded different, and the business was just moving in the direction that we weren’t expecting. I think that was the beginning of what has been a phenomenal stock for us.

These bottlenecks are a function of the fact that the industry right now, the AI industry, is demand-supply imbalanced. There just isn’t enough supply. A few years ago, of course, the market was worried that we were overspending at 350 billion or whatever of capex, and now we’re completely sold out. 

Bottlenecks will occur, of course, when you’re in a demand-supply balance. The bottlenecks, ultimately, will become resolved as there’s a supply reaction to demand, of course. But then, of course, new bottlenecks. This is a bit like when you build a new road. You can solve that bottleneck, but then a new bottleneck comes along because there’s more traffic going into the roundabout than was there before. 

So, I think that’s super exciting for us. Can we fully anticipate them? No. But can we react to them fast? Can we be ahead of other investors? I think this is why being an AI maximalist is so critical. If you’re a believer like we are, then you’re going to work that extra hour. You’re going to have that company meeting that doesn’t necessarily yield anything today. But as and when a bottleneck reveals itself to you, you’re now in a position to actually move faster than other investors because you’ve met the companies that might help alleviate or profit from that bottleneck. 

• US midterms 2026: what election could mean for your portfolio
• Stockwatch: the share poised to benefit from a new political alliance

Dave Baxter: What are the next bottlenecks? 

Ben Rogoff: Well, they’re moving all the time, aren’t they? We’re excited through the entire supply chain. Again, go back to this idea that if you have a demand-supply imbalance, there’s bottlenecks everywhere. Interestingly, there seem to be less bottlenecks in the areas that had, I suppose, at the beginning, sold out more than anything else - GPUs from Nvidia. Again, we’ve got a big position in Nvidia and I saw Jensen Huang present just last week, my colleague was in a small group with him, and he remains very bullish on the AI story, but with enough GPUs, that’s not necessarily where the bottleneck is today. 

There aren’t enough CPUs, the chips that you find in regular traditional servers, that are now required to be built into AI racks and clusters in order to permit AI agents to be able to sort of act like humans. So, there’s a shortage there. We’re very exposed to companies in the CPU space right now. Less in memory, where I think because memory prices have moved up so far and so fast, companies like Nvidia are trying to change the specifications of their products in order to alleviate the bill of material pressure from memory.So, again, right now, memory is an area that we’ve slightly de-emphasised.

We’re still very bullish on semiconductor capital equipment, companies that are obviously trying to add capacity need more equipment, but at the same time we’ve got this political desire to bring manufacturing back to the US, because obviously one of the big geopolitical risks is Taiwan and a fault line. So, we like that area. 

So, if you ask me the same question in two months, it might be quite a different answer to the one I’m giving you today, but ultimately that reflects a cadence of product innovation, model improvement that is just genuinely unusual. Again, you can see I’m slightly smiling about this because it creates a really interesting opportunity set for us, but that’s very hard to unlock if you’re trying to do this yourself or trying to do this with one or two people. Again, we have a team of 12, 1,000 company meetings a year and an unlimited travel budget to meet those companies. 

Dave Baxter: And finally our usual question, do you have skin in the game? 

Ben Rogoff: Skin in the game? Well, hopefully you can sense that I’ve been doing this for 20-odd years. I haven’t done really very much else it would seem, professionally, apart from being at Polar and running the investment trust. It’s been the privilege of my life to have managed the trust for this long, and I’m still loving it. 

And we have front-row seats to the AI story, which I think is one of those key civilisational moments. I really do believe that, the agricultural revolution, the industrial revolution, and probably this cognitive revolution. The idea that we can apply machine cognition to problems that humans haven’t been able to solve, or have been able to solve at a price point of cognition that is human. I mean, what a privilege. 

So, do I have actual financial skin in the game? Obviously. I bought the trust the day I joined Polar. I’ve owned it ever since, and I have bought more since. I also have exposure, a reasonable amount, to the company I work for, Polar Capital Holdings (LSE:POLR), and the technology business within Polar is the largest team at Polar, so I guess there are two different ways that I have exposure. 

I also own our artificial intelligence fund [Polar Cptl PLC-Artfcl Intllgnc I Acc GBP (BF0GL54)] and I’ve owned our open-ended fund [Polar Capital Global Tech I Inc GBP (B42W4J8)]. So, as you would expect, I definitely have skin in the game, but even if I didn’t have financial skin in the game, I think it’s fair to say that myself and the team, we’re all-in on trying to do the best we can for the trust. 

Dave Baxter: Ben, many thanks for your time.

Ben Rogoff: Thanks for having me.

Dave Baxter: And thank you for watching. I hope you found it useful and interesting. As always, let us know what you think in the comments, and if you’re a fan, hit the like button, and hit the subscribe button too. Thanks and take care.