Fears of an AI bubble continue to dominate investor sentiment, but one portfolio manager says the concern is misplaced. He argues the technology rally is widening beyond mega-cap names, supported by falling yields, improving liquidity and an emerging wave of productivity gains.
BNN Bloomberg spoke with Eric Jackson, founder and president of EMJ Capital, who says investors are misreading the current cycle and overlooking opportunities in small caps, AI infrastructure and other second-derivative beneficiaries.
Key Takeaways
- Tech is not overvalued but widely misunderstood, with the strongest upside in AI infrastructure and second-derivative plays.
- Falling bond yields are fueling a rate-driven relief rally that positions small caps for the sharpest rebound.
- The market rally is broadening beyond mega-caps into AI infrastructure, energy and housing as liquidity improves.
- High-quality small caps are down sharply due to rate stress, but Jackson says the underlying theses remain intact as pressures ease.
- Record cash remains on the sidelines, with macro funds waiting for clarity on rates and new opportunities in emerging AI-linked sectors.
Read the full transcript below:
ANDREW: We’ve heard warnings that tech mega caps are fully priced, and there are fears of an AI bubble, but our guest says tech isn’t overvalued — it’s just misunderstood. We’re joined by Eric Jackson, founder and president of EMJ Capital. Eric, always great to see you. Thanks very much indeed. What do you think is the big misconception about tech right now?
ERIC: Andy, I think the biggest thing is the elongated curve that’s taking place with AI investment right now. People are just impatient and constantly wanting to see the immediate payback of these billions that they see going out the door, especially from the hyperscalers into chips. And the truth is that nothing’s wrong — it’s just taking longer from the time you buy the chips and put them in the data centres for the enterprise customers to actually build the applications, to map onto their data sources in their own enterprises, and to show the payoff.
So, you know, take a name like Palantir. I mean, they typically get raked over the coals for being overly expensive on a forward price-to-sales basis. I think people are ignoring the contract growth that they have and how quickly those enterprise accounts are growing — and that all comes before they recognize the revenue down the road. So the billions in revenues are coming. These hyperscalers do not spend billions willy-nilly, and I think that’s what the bubbleistas are missing.
ANDREW: When you refer to an S-curve, what’s that, Eric?
ERIC: Well, it’s, you know, every technology typically goes through this period of slow early adopters buying into new technology — whether it’s mobile phones and mobile ads that go on those phones, or something like AI — and then suddenly the whole crowd catches up and things start to take off. So, you know, we see the same dynamics in adoption of ChatGPT or adoption of Nvidia GPU chips.
And so the question is, from a revenue perspective, how quickly are we going to see the ramp-up in revenue? It’s coming. We’re just still at the very early stages where the ski slope is just tacking up. But make no mistake — the black-diamond steep curve is coming.
ANDREW: And you do say the big, prominent AI names may be fully valued, but the supporting players, such as the electricity generators — there’s plenty of opportunity there.
ERIC: Well, no, I would even point to Nvidia. I mean, they’ve been in the spotlight, you know, with Michael Burry taking shots at them recently. If you look at their historical forward price-earnings multiple, it’s actually below the mean of where it’s traded over these last five years. And so even a name like that, I think, is sort of undervalued and still underappreciated.
I think the market just can’t wrap its head around the sheer magnitude of these quarterly numbers that they’re posting — they just can’t imagine that there’s another gear for them to shift into and start accelerating even more. It’s unprecedented for a company. So a name like that, I think, is still cheap on a relative basis to where it’s traded in the past.
And a Palantir — even though people can’t believe these price-to-sales numbers — again, it’s like a slow-moving curve, and suddenly these enterprise customers that they continue to bank are going to continue to grow.
Energy storage names, though — names like IREN and Cipher, which I own — that are critical parts of building out all the AI data centres that we need. Again, those have been cut in half in two and a half weeks in November. So you are getting a significant discount in some of these names.
And now some of them have rallied back furiously this week. So, you know, if you miss it — if you sit back and say, “Oh, I’m not going to catch a falling knife” — sometimes you can miss. Cipher, for example, has had a three-day, 42 per cent positive rally just this week. So the violent swings back up can really take some people’s heads off.
ANDREW: Tell us one or two concerns clients are expressing to you right now.
ERIC: I think it’s been the volatility in rates and the stubbornness in the rates coming down, and the Fed’s fixation on higher inflation. When’s that going to come down? And just in the month of November, the magnitude of some of the drops in some of the small-cap stocks.
So, you know, we mentioned IREN and Cipher just in the last segment. And I think people buy into some sort of long-term, multiyear thesis — saying that, “Oh, these names are the picks and shovels of the buildout of AI, and I believe they could 10x, 20x, 100x in some cases.” But then when that short-term volatility comes and you see some of these names that you hold in your portfolios — IREN and Cipher both basically dropped, I think, 47 per cent from Nov. 4 till two and a half weeks later — it’s difficult for a lot of retail investors to stomach.
And I try to say that, unfortunately, in tech investing, that’s a feature, not a bug. Even a name like Apple had something like six drawdowns from 1981 until they introduced the iPhone in 2007 of 50 per cent or more. And then even after the introduction of the iPhone, they’ve had four more drawdowns between 20 and 35–40 per cent. So that’s the granddaddy of tech. And within these smaller-cap names, the moves down can be even more vicious.
And unfortunately, you just either have to ride it out. Because if you try to time the market and jump out and think you’re going to jump back in, more often than not you’re going to lose more money that way than if you just stick to your guns.
ANDREW: Just give us a bit more on a couple of those names you mentioned — IREN, what are they all about, Eric? Just remind us.
ERIC: So they’re building from the ground up. They’re buying the real estate out there in geographic locations — including, by the way, Prince George, B.C., but also West Texas — where they think they have a cost advantage. A local power operator — they can negotiate a fixed cost for the foreseeable future to build out an AI data centre.
And so they’re taking on that risk. They obviously believe that AI is not in a bubble and it’s here to stay. And once a data centre is built out, it’s awkward. Sometimes they’re buying the chips from Nvidia to power it up, flipping on the power, and then saying to the hyperscalers of the world or the Oracles of the world, “Hey, who wants to access our data centres? We’ve got the power. We’ve got the chips. We’ve got the actual physical data centres.”
And it’s basically a game of musical chairs now where these hyperscalers need access to that energized data-centre source now — not in three years. So the people who have thought ahead and built these things out are sitting in the catbird seat.
So an IREN and a Cipher have basically morphed from being just a bitcoin miner a few years ago to taking their know-how to build these data centres and positioning themselves at the front of the queue to take orders and power the next generation of AI.
They’re referred to as neo-clouds. IREN and Cipher have three gigawatts each of these kinds of data centres coming online in the next year. So that’s what makes them more interesting, in my opinion, than some of the other players trying to manoeuvre into the space.
ANDREW: And just broadening it out here — a stock you’ve long been involved with, Opendoor, that you like the look of. Do you still favour OPEN?
ERIC: Yeah. It’s 10x since the summer, and I think it can — it’s still a potential 100-bagger from here. What’s most interesting for the Canadian audience is that the guy who was the COO of Shopify — who basically had to do nothing, just stick around, and whenever Toby steps up to executive chairman, this guy, Kaz Nejatian, was in line to become the next CEO of a US$200-billion market-cap company.
Today, he decided to quit and take the job as CEO of what was then a US$3-billion market-cap company. Now, why would he do that? I think it says something about the challenges that Shopify faces in the AI age — and by the way, this is their most important week of the year right now. But it also says something about how well-positioned Kaz thinks Opendoor is to disrupting the whole real estate market using AI.
You won’t need a real estate agent anymore. You won’t need to do showings of your house anymore when Opendoor rules the world, Andrew. So it’s very interesting from a bunch of angles. Obviously, home ownership and the cost of homes is a crisis around the world — in Canada, in the U.S. — and Kaz is looking to make Opendoor the Amazon of real estate. Faster, cheaper, more price certainty of selling your house — when you’ll get the money in hours instead of weeks.
ANDREW: Eric, I think we’ll have to jump. Thank you very much. It’s always great getting your insight.
ERIC: Great to talk to you.
ANDREW: Eric Jackson, founder and president at EMJ Capital.
---
This BNN Bloomberg summary and transcript of the Nov. 27, 2025 interview with Eric Jackson are published with the assistance of AI. Original research, interview questions and added context was created by BNN Bloomberg journalists. An editor also reviewed this material before it was published to ensure its accuracy and adherence with BNN Bloomberg editorial policies and standards.
