Tariffs, Trade and the AI Stack (The Derby Mill Series ep 04)
Question on that one, it may be off your expertise, Mark, but nothing’s off my expertise. I’m actually insulted by the comment.
So welcome to the Derby Mills series: Intrepid Pioneers of the Next Economy, featuring discussions with entrepreneurs at the forefront of deploying machine intelligence and brainstorming sessions about where the technology may go at the limit.
Hello everybody, welcome back to the Derby Mills series: the Intrepid Pioneers of the Next Economy. I’m Ajay Agarwal, co-founder of Intrepid Growth Partners, and I’m here with colleagues at Intrepid Growth.
So first, my colleague, co-founder and managing director Mark Nation. Mark has previously led Goldman Sachs in Asia and then headed up CPP Investment Board. Mark, maybe for our listeners who aren’t familiar with the scale and size and scope of CPP Investment Board, if you could just give a quick description of that.
Yeah, CPP Investment Board is the biggest pension fund in Canada. It’s responsible for investing all the money that’s not immediately needed for the 21 million contributors and beneficiaries across the country outside of Quebec. The scale, when I left, was about 600 billion dollars in terms of invested capital, a bit more than that. We had 2,000 people across nine offices and about 65 different investment strategies around the world.
Excellent, thanks Mark. My other co-founder, Mark Shulgin, has had a career focused on growth equity investing, and that is the sweet spot for Intrepid. Mark, for our listeners who aren’t familiar specifically with what growth is relative to other areas of technology investing, if you could just give a brief description there.
Yeah, happy to. So I’ve been investing in growth for now 20 years, and it’s taken different forms, but at the core of it, it’s always been around identifying companies that achieve something called product-market fit. At the venture stage, it often involves backing science experiments and companies that have not necessarily identified who their core customers are going to be. By the time they’ve got to the growth stage, they know who their customers are, they’ve got some initial traction that demonstrates that they have a real product, and adoption is starting to pick up.
At that stage, it becomes about how we help to scale companies so that they become a lot bigger. Oftentimes, for the companies that we’re focused on in Canada and the UK, that means how do we sell into the U.S. market? So obviously, trade and tariffs are a big part of that.
Excellent. All right. And then our special guest today is Professor Mark Bush, who’s the Carl Landegger Professor of International Business Diplomacy at SFS, the School of Foreign Service at Georgetown University. Mark, welcome.
Thanks so much, AJ. Glad to be here.
Okay, so I’m going to ask a couple of questions just to kick us off and then invite Mark Nation and Mark Shulgin to weigh in with comments or questions at any point as we go. Mark, if you can—Bush, if you can—just you used to be an expert on a topic, tariffs and trade, that only sort of real specialists knew anything about, and now in the last six weeks or so, everybody is talking about tariffs and trade. It doesn’t matter what their field is.
So can you just give a sense to open it up here of what’s the core essence of these discussions on tariffs and why is everybody so worked up about it?
Well, for starters, I have to say that I could never have imagined anything remotely like this happening in 2025. It wasn’t that long ago when I would tell my MBA class that I’ll give one session on tariffs and nothing more because rich countries don’t do this. Rich countries do non-tariffs. They’re far more lethal. Even these days, as I go from audience to audience, I almost want to reach out and say, but I do real technical stuff too. It’s way cooler. It’s way more lethal. Please let me also talk about some of the regulatory measures. The shock and awe—and that’s the only way to describe what we’ve seen since really February 1st—has caught the world partly off guard, partly not. Partly not in the sense that Trump 2.0 is not Trump 1.0. The world anticipated a lot of action.
Where I think the surprise comes in is with respect to how much can Trump really pull off through executive orders. Isn’t Congress holding the purse strings with respect to tariffs? Doesn’t that kind of limit his scope of action? Even as we record right now, we’re waiting for the announcement of some kind of plan with respect to what he’s calling reciprocal tariffs. Trump is reaching for things that allow him to bypass Congress and to do so at times on a temporary and at other times on a less temporary basis.
That’s really where much of the action is right now. What are the legal bases for what we’re witnessing? Are they subject to judicial review in the United States? How does American business feel about some of what we’re seeing? Are there concerns that we should have going forward that this isn’t going to end anytime soon?
In other words, if you think about where we started on February 1st, we got a memo calling for a whole bunch of studies. What we didn’t get was the across-the-board stuff. Then you saw Canada and Mexico and China all targeted—Canada and Mexico more minus energy and then China lesser. Then suddenly we just got the revamped Section 232 steel and aluminum tariffs, and then today we’re waiting for reciprocal tariffs. We were scheduled for Saturday to have the announcement of semiconductors and pharmaceuticals getting hit with new tariffs as well as some other military concerns.
Then, in the back of everyone’s mind, there was the end of March, which is when the original U.S.-EU ceasefire is set to expire with respect to the steel and aluminum tariffs from Trump 1.0. That was going to be a wave of tariffs.
Finally, there’s the consideration of what will Trump make of the results of all these studies that are due on April 1st? Will April 1st unleash the floodgates? Will we get a whole new string of tariffs following through on that end? We’ll come back to one of the studies that Trump has asked for because it is disconcerting, but it will trigger tariffs filed by firms, not by the administration.
So, in short, again, Trump 2.0 is not Trump 1.0, but Trump 2.0. There’s a little more razzle-dazzle out of the gates, and while we expected some of the tariffs and the legal bases for doing them to unfold much as they are, it really is just the sheer volume before Congress. There’s a little bit of oil on that, and I think that’s going to be the big mystery.
We are seeing legislation in the House right now to curtail the Canada and Mexico tariffs. We are seeing legislation in the House that would reign in one of the legal premises for the tariffs against Canada and Mexico, namely AIPA, the International Emergency Economic Powers Act, which has never been used for anything remotely like this. We have court challenges that are surely in the works. We have anger and frustration politically. So it’s a lot, and my sense is it’s not going to die down anytime soon—or at least not before April 1st.
Okay, I’m going to ask you two more questions and then hand it over to my colleague Mark Nation. So the first question—and again, just to set the stage for listeners who are not familiar with trade—is at the very beginning of your remarks you talked about tariffs as one small piece. Most of your time is focused on other trade-related issues. Can you just give people a sense of what’s in the other bucket that’s not tariffs?
So beyond tariffs, we have things like non-tariff barriers concerning regulatory measures, and these regulatory measures include things that range from health and safety standards on food and agriculture all the way through to technical specifications on how to build a solar panel.
That’s really where trade gets far more esoteric. That’s where you begin to see things that are also triggering outright bans if you don’t comply with the technical specifications, for example. One subject that I’m sure will come up today is intellectual property, which is ultimately embedded in a lot of goods, but it’s also embedded in services, and we’ve got a lot of issues with respect to cross-border services. AI is really forcing consideration of new categories of traded services, and maybe we’ll talk about Mode 5 as we go through this session.
But the point being, if you were to imagine a base of trade action, you would start with tariffs, and they’re simple but a lot of collateral damage. As a result, you’re not being as politically nimble with a tariff as you might be with a non-tariff, which by and large can be tailored at times to meet the needs of a particular firm or industry. That’s the advantage politically of using non-tariffs. Because you are talking about a measure that could invoke a ban, they’re also just far more lethal than a tax on trade.
As we’ve witnessed since February 1st, Trump seems to have in mind the numbers 10 and 25 in all of the tariffs that he rolls out. Well, a 10% tariff, a 25% tariff—you’re doing harm, but it’s not a ban. And therein lies the delta between a tariff and a non-tariff. You just get far more lethality with the non-tariffs than with the tariff.
In all honesty, I left Trump 1.0 with the following observation: thank god he’s not non-tariff man. I’ll take tariff man. What really drives fear into my heart is the prospect of non-tariff man.
Okay, so again, for people who are sort of just starting to wrap their minds around this, when you talk about lethality—how lethal something is—what you’re referring to is the benefits that are conferred upon, in this case, American firms relative to non-American firms for selling into the U.S. market. So tariffs create a, let’s call it a small wall, which you can climb over, but it’s expensive. The non-tariffs potentially put up an impenetrable wall that you can’t come in at all if you don’t comply with those regulatory barriers.
So all of these things we’re talking about are different tools that a country has to make it harder for non-American firms to compete in American markets.
Right. So my second question, and then I hand it over to Mark Nation, is you talked about, you mentioned food and agriculture, you mentioned steel and aluminum. A lot of what we’re focused on at Intrepid is software. So it’s software companies; sometimes the software is embedded in hardware, and sometimes it’s not. Can you just give us a sense of how do these—let’s call them both tariffs, because that’s what the current administration is focused on right now—but also whatever you think are the most serious non-tariff tools that impact, let’s say, Canadian or European AI companies, largely software, sometimes embedded in hardware, selling into U.S. markets?
So the U.S. doesn’t have a long history of having big tariffs on things that look like software. You have to understand that for the longest time, no one really even knew what kind of tariff code you’d be talking about in the context of software.
Where life gets interesting is when, as you say, the software is embedded in the hardware, and then suddenly you’re talking about potentially hitting the hardware with a tariff. That’s why I alluded to the issue of Mode 5 services because Mode 5 services are a blend in theory of a good and a service.
What you’d be looking for in a Mode 5 regime would be that the service allows the good to come in at a lesser tariff because it changes it, but moreover, that the good allows the service to bypass some of the restrictions that we have on those. So you’re kind of getting two wins through Mode 5 if you’re thinking offense on the defense.
When you talk about AI, the one item that is on Trump’s tariff hit list—and let’s just put it out there because everyone’s going to be reading about it—is semiconductors. We don’t know what the number would be. I’ve heard numbers that go as high as 100%.
We haven’t seen anything, but semiconductors were supposed to be part of the February 15 salvo. There had been a presser that Trump gave where he kind of hinted that he wasn’t sure that it was going to be February 15th, and then there was a question about copper and would copper be postponed in terms of the February 15 salvo.
So, still some questions. But let’s talk about your inquiry into the non-tariffs. When you say AI to me, my first thought as a trade person is the European Union’s AI act, and I worry that we are going to get a lot of trade issues, but they’re not going to be tariff-based; they’re going to be non-tariff-based.
Let me give you one: the European Union’s AI act defines AI by high risk, low risk, and then there’s a prohibited category. One big question that comes up in the legislation is: who gets to audit for high versus low risk? Now, this would fall in part under an agreement at the WTO called Technical Barriers to Trade, and the issue is conformity assessment. How would you do conformity assessment? Meaning how would you audit high risk versus low risk, given that compliance is mandatory, backed by a government making this a technical regulation? How would you audit that? Who gets to audit it? Could American AI producers or Canadian AI producers audit themselves?
Or, and here’s our fear, would the EU require that the AI arrive at what they call a national approved body to do conformity assessment? If that was the truth, given that that’s already raising the cost of the exporter because you’ve got to get it to them under uncertainty about approval, what would the EU national body ask for? Does it ask for the data that you train the AI on? Does it ask you for some algorithms? What does it get to do the review?
And then you’ve got this other question in the context of AI, which is the digital trade field. Here, ironically, there may be some good news with Trump because under Trump 1.0, we got the United States pushing really hard on digital trade. Then President Biden pulled the U.S. out of a WTO Joint Statement Initiative on e-commerce, which is the digital trade agreement.
Now, the question is: does Trump go back and finish the deal? Does Trump bring the U.S. back into the fold and get digital trade right? By “right,” I literally mean right because as soon as the U.S. withdrew under Biden, the remaining countries gutted it of a lot of the things that the U.S. had always demanded, including a prohibition on digital taxes, which would be as close to a tariff on digital trade as we’re going to get.
It has fans out there, including India. Also, a prohibition on data localization—the idea being that to do AI, you’d have to be having even as a small entrepreneurial effort some kind of physical presence with a hard drive in the country you’re servicing, etc., and cryptography.
So these long-standing U.S. preferences voiced under Trump 1.0 would be good to follow through on under Trump 2.0. I want to leave your listeners with something they’re not hearing, which is this guy’s going to take the U.S. back on the offense, and that’s good news.
The downside to Trump 2.0 will be that we get what we’re getting right now in terms of tariffs; in other words, closing the U.S. market or at least raising the cost of getting in. We can talk about why he’s doing that and whether it’s going to work, and the answer is no.
But what we’ve not seen for four years under Biden is any offensive effort, and you’re about to see two documents come out here in the United States: the National Trade Estimates report and the Special 301 report, the latter being explicitly about intellectual property and only IP.
Biden did nothing with these documents for four years and neutered it in the fourth year. Now, the question is: what does Trump do with that document? By sheer personality, my bet is he begins to express a lot of anger, including at Canada, which lives on this list. And demands greater adherence to stricter IP enforcement as likely a condition for skirting future U.S. tariffs. So from a U.S. perspective, I think there’s some really interesting action on the offense. As you may have heard just this past week, Vice President JD Vance lectured the European Union on too much regulatory oversight, including on AI, and he’s right to say it.
Let’s see how the EU responds. It’s not just on AI; the U.S. will ultimately have to swing into action on, for example, standard essential patents like the 10,000 of them in 5G. I’m sure it’s going to spiral under 6G. We have currently a China-EU battle on who gets to determine fair and equitable royalties on standard essential patents. This is a concern if you’re Qualcomm, for example. Where do you get to sue Huawei for a violation of a standard essential patent?
Chinese courts have said we decide, and you must exhaust the Chinese court before you pursue any litigation in a foreign jurisdiction. Europe says the hell with that! File the WTO case against China and then replicate what the Chinese are doing with a bureaucracy. Go figure. So all said and done, the U.S. hasn’t been in on this game. What Vance was saying is can we let the market work? That is probably where the U.S. will come out on the issue of both AI and standard essential patents.
Excellent. Okay, Mark, thank you. Let me hand it over to Mark Machin, and then after he’s had to go, Mark Shulgin. Thanks, Jay. Mark, thanks again. Can you just drill down a little bit again on, you know, from the aspects of what we do at Intrepid? We’re trying to support the growth of companies that have machine learning and AI at their core, that are based in North America and Europe.
But we have a particular tilt towards Canada and the UK, in particular in Europe. As those Canadian, UK, and European entrepreneurs look at this, what should they be doing, if anything, in adjusting their businesses? Thinking about generally, these are people who are software type businesses. So, we’ll put the hardware to one side for a second. Just so you know, so you’re a software type business.
You know, you’re in legal tech or education tech, or medical diagnosis, more on the software side. Obviously, the U.S. market is the most important market in the world that they can tackle. So as they look at this and they think about their strategic planning for the next couple of years, what advice would you give them as they think about this?
Unfortunately, I’m going to say something that many of them aren’t going to want to hear, which is you’re going to have to get a little political. By that, I mean you’re going to have to start following some of the U.S.-based trade associations and find out what they’re prioritizing and how they’re selling it and the political narrative behind the sale. I’m not saying that you have to have every entrepreneur from Canada or Europe walk the halls of Congress, but you had better be ready to explain what it is that you do, and you had better be appreciative of what you’re about to hear a lot about, namely U.S. national security.
Since everything raises national security considerations and since this administration, like its predecessor, uses national security as a get-out-of-jail-free card for any violation of any obligation, it’s really important to understand what that means. I can tell you from speaking with semiconductor companies and others, a lot of businesses have no idea what this or the last administration meant by national security.
One issue that we should talk about before the end of our session today is export controls because one thing is on the inbound; the other is will we have restrictions on the outbound? I have a feeling that with respect to a lot of cutting-edge technologies, you’re about to see a lot of export restrictions targeting, obviously, Chinese consumers and businesses. But nonetheless, it could generalize to a wide swath of countries, especially if that’s offered as a bargaining chip in whatever Trump’s reciprocal tariffs executive order we see by the end of business today.
So all things equal, if we were having this conversation in the 1990s, we would have talked about let’s get some baseline regulation on the books so that business has less uncertainty and just do what globalization allows you to do with respect to efficiency, and the world is yours. It’s not so clear anymore. Certainly, my advice to entrepreneurs is, while you may think that you can hide from scrutiny by virtue of your size, you’re wrong. It doesn’t mean you have to have a U.S. headquarters; it just means that you might look into having a U.S. presence at some point.
You might look into having some legal footing in the United States at some point. But even before that time arises, you had better get a little savvy with respect to the political machinations of the administration and the geopolitics of national security in particular. It’s just something that has to be taken into account as you think about your software and your contribution to whatever it is at the downstream end-user point. Because all of this is frustrating the big companies, never mind the little ones.
Whether you’re open-source or you’re proprietary, you raise different concerns, but you’re not hiding from national security concerns. You know, just again as the relatively small entrepreneur, I think your advice makes sense. But what does, and this may be getting into too fine weeds, but what does it mean? You know, how much of a U.S. presence do you need to have at that sort of relatively early stage of building your company?
You know, sort of four, five, six years in, you’ve got a staff of people and engineers, and you’ve got a couple of anchor clients in Canada, the UK, or Europe, and you’re very focused on the U.S. market. Just having, you know, having your sales team and a U.S.-incorporated entity for the distribution of your software and services into the U.S. Is that a sufficient way of doing things? You said you don’t need to report your whole company to the U.S., but is that something that you need to consider over time?
You know, I really like to point out to businesses asking this question that at some point, you’re going to have to have a place where someone from Congress can go see what you do. It really comes down to something as simple as that. That might not be on day one, and it’s going to depend on what it is you do and how much data from where you depend. But ultimately, no, not necessarily headquarters here, but at some point, and maybe not just on your own, but with a couple of other like-minded entrepreneurs, you begin to have a kind of ‘We Work’ government affairs team doing something for you here.
One of those somethings is explaining to members of key House and Senate committees what it is you do to get ahead of the curve rather than being purely reactive. Since everybody is trying to learn AI right now, it’s a great time to get in on the action. Even for American companies that are physically and legally in the United States, you’ve still got to get out there and talk about what you do and make sure that your end user, your consumer, isn’t afraid of you and isn’t afraid that by virtue of what you do, you are opening them up to liabilities with respect to geopolitical and national security considerations.
Again, I said it isn’t necessarily what everyone’s going to want to hear, but there’s a lot of face-to-face interaction, and there’s a lot of walking the Hill. At times, what really matters isn’t so much grabbing the attention of a senator or a representative, but rather their intern and just explaining what you do. Now, not all that has to be done by you. You can join trade associations; you can join various groups that are affiliated with, say, the Chamber of Commerce or something like that.
They’re already made go-to allies, but sometimes you’re a little different than the average yahoo on the block. The more different you are, the more individual you’re going to have to custom the messaging. Very good. Alright, let me hand to Mark Shulgin.
Yeah, thanks, Mark. And thanks, Mark, for being on. This is really fascinating. I’m curious; I’ve got a couple of questions. The first one is around tariffs on semiconductors and the impact that that could ultimately have on compute capacity within the U.S. And then, you know, I also wonder—we saw this announcement about Stargate, and it’s a huge investment into the compute capacity for the U.S.—how coordinated do you think those two things are? Is it effectively a tariff met with a subsidy of some type?
Then, you know, the second part of that question is what do you think the longer-term implications are for the U.S. and the AI race? It’s definitely a subsidy for economic harm that Trump fully realizes he would be causing with the tariff on semiconductors. That is more broadly his offset. The way to think about what he’s doing to allay concerns about the tariffs is he’s promising businesses, “Listen, I have offsets. I’ll lower your tax rate; I’ll lessen the regulatory burden. I’m doing things that should compensate for the additional cost of a 10% or a 25% tariff,” begging the question, “Why can’t you go down in history as the most pro-growth president and not do the tariffs and do the regulatory slashing and hacking and do the tax cuts?”
But apparently, that question is off the table. So what is the Stargate angle? It’s to compensate for whatever the tariffs are. The question, though, is will it? Interestingly, under Biden, with the effort to pump up R&D for semiconductors, etc., you saw a lot of this new vogue industrial policy initiative that really is not the meat and potatoes of Washington, D.C. coming to the table, a little bit late and rediscovering industrial policy is an interesting affair.
Trump does not believe necessarily in industrial policy and certainly has not been a fan of the Inflation Reduction Act. So you’ve got some trade associations in D.C. trying to defend the Inflation Reduction Act on semiconductors, for example. The effort that Trump has initiated seems to be somewhat in competition with that. I wouldn’t want to chalk it up as entirely based on an offset for whatever the tariffs are coming, and I don’t believe that the offset will be sufficient to reduce the harm if we do get big tariffs on semiconductors.
The estimates are pretty clear; we can’t afford the tariffs. They’re not going to induce any more foreign direct investment than is already happening without the tariffs. You have to keep in mind that really is Trump’s narrative—”Build it, and they will come. Build the tariff wall, and they will invest in the United States.” We call that tariff-skipping foreign direct investment, and it works in Trump’s head, but that’s because he doesn’t appreciate the ever-lengthening supply chains that ultimately drive a product like a semiconductor, which we’ve got inputs into chips that are traveling 11,000 miles around the world before final assembly.
You don’t reassure that, definitely not overnight, and it ain’t free. What’s interesting right now is how many trade associations are trying to explain with detailed math how the numbers don’t add up. Stargate won’t be a pure offset in as much as it fully compensates. But then the question becomes how does Trump react to all of this? Does he lower the tariff? Does he up the subsidy? Does he rediscover IRA? What is the next step?
Clearly, he sees this very much in terms of geopolitical tensions, not so much technical prowess issues. So long as it hurts the Chinese, we may not grow as fast, but hurting the Chinese may be more important than growing fast. Well, it’s so interesting. What do you think, or is there any thinking in the short versus long-term KPIs that he is focused on that will determine whether or not these moves are a success or failure?
Everyone’s bet is that he’s going to watch the market reaction, and he’ll recalibrate given peaks and troughs in the market. Remember, the Republican Party, which is giving him clearly a honeymoon grace period at the moment, has to think beyond the next four years. You’ve got a lot of skittish House reps, in particular, who are beginning to sense that life may not be good in two years when they’re up for renewal.
The correction in the market may be just that—Trump corrects some of these policies given market reactions, especially if inflation numbers continue to be unflattering, as they were for the month of January. How does a guy who was elected to combat inflation start posting numbers that are lesser? My sense is that he’s surrounded by enough bright people that he has to at least take some cues, even if he doesn’t do a full 180.
How long this lasts really depends on a couple of other factors. How do the courts show deference or not? How much domestic litigation, never mind foreign litigation, do we get? What is the political reading with respect to key swing states, in particular, which now must be baffled by the prospect of Canadian crude subject to a 10% tariff, with an anti-retaliation clause in that executive order and Justin Trudeau’s promise to retaliate?
No one takes the 10% number seriously; we know it’s going to go up. It may go up just because Canada is going to respond to the steel and aluminum tariffs. You quickly see how Trump will lose escalation dominance here, and whatever moderation he shows with respect to a given tariff line may be obviated by foreign responses, which is the ultimate in terms of bringing him to the negotiating table.
Let’s see how quickly that happens. If you’re looking for one retaliatory strike as a bellwether test of that hypothesis, it will be the EU. Remember, under Trump 1.0, Europe has fully rearmed in terms of trade. They’ve given themselves two legal instruments with which to hit back; they will not be caught flat-footed. That is the ultimate test of Trump’s resolve if Europe starts line-item listing U.S. exports for retaliatory reprisals. Like Canada, Mexico, and China, they do not limit retaliation to tariffs but utilize non-tariff methods as well, including, by the way, like in China, antitrust reviews of U.S. high-tech.
That will be the ultimate bellwether test.
Question on that one—it may be off your expertise, Mark, but—
Nothing’s off my expertise; I’m actually insulted by the comment.
Yeah, so antitrust. I mean, it’s a curious one to me. I mean, certainly, well off my area of expertise, but I don’t quite understand the extraterritoriality of foreign jurisdictions saying, “Well, we’re going to hold up a merger.” I presume it has to have some meaningful part of the business within that territory. I mean, I presume that’s the way we’re so China can hold something up because there’s some meaningful part of the business within that territory or Europe can hold it up.
Why does the U.S. have to be beholden to offshore jurisdiction on something that may just be a domestic consolidation? You know, but it’s not just domestic. Remember back in the early 2000s when Microsoft was under antitrust scrutiny by the European Union? It was all about the IE Explorer web browser.
You had a lot of scrutiny, and Microsoft asked the same question you’re asking: how is it possible that Brussels gets to do this review? The answer was you do enough business here; we have enough of your IE Explorer here. By the way, why are you looking for science behind this all? It’s called politics; look it up. The Chinese have the same story. For them to do antitrust on Google, for them to impose anti-injunction on standard essential patents, how are they able to do this? Because it’s politics.
Yeah, but for example, on the Google example, I mean, China blocks Google and doesn’t use Google. So why, you know, if it’s not a meaningful customer, why does anybody have to pay attention to it? It’s a great threat, and let’s see how this one plays out. Obviously, in part, it’s targeting some of the people who have had Trump’s ear, and that’s not terribly surprising. I got asked by MSNBC the other day, is it possible that Elon Musk will be lit up in terms of foreign governments retaliating for Trump’s tariffs? My point was, of course, but it has nothing to do with Elon Musk heading up Doge. He’s a highly politically salient business owner; he’s going to be lit up in any given retaliatory strike, just like Harley Davidson will, just like Levi’s jeans will, just like bourbon will, for all the same reasons. This is how the game is played, and that’s why countries are loathe to get to the stage of strike counter-strike because it quickly gets out of control.
But we know where the first salvos are aimed—at the most politically influential businesses in that country. Because the logic of any reprisal is you hurt those who have the ear of power, and you force compliance or a cessation of hostilities by getting those to speak up. And that’s what you’re about to see in big-ticket items like what Canada is threatening against the United States: Harley Davidson and 87 1150 as a tariff code for big bikes over 800ccs, right smack dab in Justin Trudeau’s list of retaliatory strikes. There’s a reason for that—it’s good politics.
And by the way, yes, there is a motorcycle caucus in Congress. One other thing you mentioned a couple of times is WTO. My impression, again from a big distance, was that the WTO has been quite out of action for quite a while because of the lack of judges appointed, and it’s not really functioning at all right now, and clearly won’t without the US’s support. So does it matter at all? Is it functional?
It is functional. The issue with respect to judges concerns the appellate body, so we have no review process at the moment. But we have about 20 plus countries that have a workaround for that. To say the WTO isn’t working is a very American-centric perspective. It is true that in the United States at the moment, the WTO has not been used with any interest in the past four years. Trump 1.0 filed six WTO cases, so I wouldn’t expect that Trump does nothing with the WTO even if he doesn’t unblock the appellate body. The rest of the world continues to take the WTO very seriously.
The EU is in a real jam on this one. Their point is, yeah, we’ll negotiate with Trump to get out of tariffs, but whatever we do has to be WTO compliant. We’re not screwing over our other trade partners just to appease Trump. Same with the AI Act; they have made very clear that whatever happens in any rewrites must be WTO compliant. That was their goal as they were drafting it through the European Parliament. It has to be WTO compliant. Ironically, what’s the first thing China does when they’re hit with a ten percent Trump tariff last week? WTO case. What does Canada vow on day one of the executive order? WTO case.
So you have to keep in mind two things about the WTO: There’s the law, which is a very useful language in explaining why you have to come into compliance when you’re the losing country, and then there’s the optics. For whatever is said and done, the US is in a jam on the optics side. It’s one thing to thumb your nose at the WTO, pointing to concerns about precedent and about the self-judging nature of the national security exception, which is standard for the course at the moment here in DC. It’s another thing to thumb your nose at anything the WTO does because the US has ready-made alliances at the WTO, most of which are arrayed against the European Union on agriculture, and we need those.
So it would be a real mistake for Trump to foo-foo it entirely. Moreover, the action on digital trade would be really hard to replicate one-on-one, nation to nation. We got a ready-made platform at the WTO. My final point is, the more the US isolates itself and is looked at as the outlier in the global economy, the more difficult it will be to realize some of the benefits of the studies Trump called for on day one of his presidency under 2.0, namely, who out there can we strike a trade deal with, who do we like?
We know what the answer is going to be. UK is up first. If I’m the UK and I’m witnessing no deference whatsoever to anything the WTO has by way of obligation or process or side deal, I’m beginning to wonder what’s in it for me. In that regard, just to put it out there to your listeners, if Trump is really pushing to get a USMCA review sped up, or even if he waits till July 2026, if I’m Canada or Mexico, I’m beginning to wonder, is it worth the candle?
I can’t believe I’m saying that because I graduated Queens in 1988, the year that the bilateral was done between the US and Canada that led us to NAFTA and USMCA. We were always told in every course I attended that year that we needed the bilateral to protect us from US unilateralism. I have to say the speed with which Canada went unilateral on Chinese electric vehicles was stunning, completely unfounded in terms of domestic and WTO law, and the speed with which Canada threw Mexico under the bus to appease Trump was shocking and un-Canadian, if I can say it that way.
So, all said and done, I think USMCA has brought the worst out of Canada, and I’m not sure it’s worth the price of admission. I tend to believe what the Conservative Party of Canada said back in 2019: this deal is NAFTA 0.5. There’s one good chapter under USMCA—it’s the sanitary and phytosanitary chapter. The rest of it? Give me a break.
Mark, let me take you back to the line that you were going down with Mark Mason’s question on Canadian software companies. So let’s imagine three different AI companies. One, let’s say, is in an unregulated market, customer service. These are AIs that, you know, a customer comes up with some issue and reaches out on something, and an AI receives text or a phone call or something, predicts the intent—what does this customer want—and then generates a response.
A second Canadian AI company, let’s say, is in a regulated industry like healthcare. So say an AI that predicts whether something seen in a medical image is cancerous or not cancerous. And then third, in a mining industry or oil and gas, let’s say mining, where there’s an AI that is looking at core deposits and predicting what mineral deposits are under the ground. I picked that one because that is associated with exploration and mining of potentially precious metals and things that are relevant to national security.
Okay, so we’ve got one application in customer service, one in healthcare, and one in mining. In what way would your advice to each of those three be different because of the type of markets that they’re competing in?
So for customer service, I’m worried about data privacy for individuals, and I’m worried about the fact that the US is kind of lagging behind the European Union in this regard and hasn’t quite figured out what it wants to do data privacy-wise. There are a variety of issues I could see, but I’m probably thinking that’s the low-hanging fruit story.
On healthcare, I’ve got some concerns because obviously we have a very different system here in the United States, and that system lends itself to, again, data privacy concerns. But it also raises questions about certification of who’s giving what advice and what are the legal implications; because, you remember, this is a very litigious country, and healthcare is certainly not immune to litigation. So you’ve got slightly different questions, but they all kind of track on the customer service AI, but hyped up by virtue of what we have in terms of regulatory oversight, especially where things could involve recommendations on what to do—things that might involve me then going and getting access to a Class 3 medical device. And medical devices, boy, it’s one of the nastiest industries in terms of global trade issues—so nasty that you typically have an annex to our TBT chapter just to deal with the uniqueness of some of their challenges.
Mining—I love the narrative that I could spin about the mining investment because you’re absolutely right; everyone has an insatiable appetite for critical minerals, etc. We got one big problem, though, and I do believe Trump will try and speed up things, but I don’t know if your listeners know this: the US ranks second to last in the world, only ahead of Zambia, in terms of time from discovery to permit. The average is 28 years.
As a result, my big question is: you can discover it, but can you do anything with it if you discover it? On the one hand, if you tell me that you have the ability to authenticate deposits, I love it, and I can tell you exactly the narrative I would want to sell on Capitol Hill to get energy listening to why it is that this all has to be reformed.
But I worry that you’re talking a long time horizon, and I just don’t know how quickly things will work, even if Trump decides to turn his sights on this. Certainly, he has to because of the dependence of the US, like the EU, on China for critical minerals. And now with the reprisal by China to impose export restrictions on some of those, we got to have more security here. Hopefully, security will be defined as North American security, and I would expect that Canadian mines will play a much more prominent role in that story than has been true in the past.
But the Canadians are slightly ahead of the United States in terms of time to permit for a mine, but there’s much greater confidence and far less fear of litigious actions once the mine is up and running in Canada than there is in the United States. So if I’m looking with your AI at Canada and Canada proactively engages with the Trump administration on some critical minerals compact, I’m really liking that story. I still fear time to permit, but I’m liking that story a lot.
And so, again, you asked me as a trade person; I can find angles on all three. But ultimately, JD Vance and his lecture to the EU is promising for all three, and it’s promising in the sense that the United States has long had this view that our number one policy consideration in trade negotiations is performance over providence. I believe in that. I was on the advisory committee for technical barriers to trade, and that was the way I coined the issue for the US.
It’s how does it perform? For Europe, it’s where does it come from? What is the providence? Who authorized? Who has oversight? If what JD Vance was signaling to his European counterparts was that the US is going back to performance over providence, shedding the Biden preoccupation with maximizing our regulatory policy space by not agreeing to any international obligations, then all three of your examples, AJ, are that much better off. It should be performance over providence.
To Mark Mason’s point, that’s why you’ve got to, every once in a while, walk the hill. You got to show performance; you got to show how this thing works, what it does, what it doesn’t do. You don’t lecture those in Congress on your business model; you don’t give them the esoteric; you give them a demonstration. And that’s why even as an entrepreneur, I show up and I give the demonstration.
For the entrepreneurs out there, collectively come to DC and do that, and walk the hill a little bit, and walk the trade associations a lot. I gotta say it’s a wonderful way of getting the word out. I was advising a Nigerian honey producer; it happens to be led by a Georgetown grad. Most incredible varietals I’ve ever tasted of honey. The first thing we did, because we’ve got to clear real health and safety concerns, packaging concerns, labeling concerns, etc., I said, give me all eight varietals, and I’m giving it to the National Honeys Trade Association in Colorado. We sent it to them, and they said, this is the best honey we’ve ever had in our lives; we can’t believe this! Open the floodgates! Open the floodgates!
We’re still up against the same regulatory challenges, but with a little bit of knowledge out there and with a little bit of guidance, it’s amazing how you can make the politics ooze on by. And that’s why, especially when you’re in the tech domain, you got to be able to show it because, God help you if you have to explain it.
All right, Mark, we’re down to our last four minutes, so we’re going to lightning round. We’ll each ask one question—just a 30-second response. Okay, so I’ll go first, and then Mark Shulgin, and then Mark Machen will bring us home.
So my question for you is: are the CEOs of the companies that Intrepid invests in, if they have a couple of minutes with the Canadian government, the Trudeau administration, or his successor, what is the singular point that the AI software community should bring to the federal government in terms of their response to all this?
Two points: Get off the watch list of Special 301 and improve domestic IP compliance, and appease Trump on that crucial issue. And two, it’s time for the government of Canada to play offense. Stop waiting for Trump to issue ad hoc demands; offer a compact ex ante. Stop waiting ex post. Canada and the UK produce some of the world’s best AI talent.
What are the countermeasures that we should be taking as we think about the floodgates for immigration opening up for the most talented folks to get down to the US? What is the best way for us to incentivize those individuals to stay, to stop the brain drain from Canada to the United States?
A perennial question, Jay. I asked this question when we were at Queens in 2000, and I’ve never heard anyone do much about it. But I gotta say it’s about the opportunities, and it’s about the potential for growth. Ultimately, the universities are going to have to deregulate to begin to compete with their American counterparts as places where people can spend a few years before they go into the private sector with that kind of talent.
And this isn’t a new answer, but it sure is not an answer that has been taken up by the Canadian government. It’s time to offer the same entrepreneurial opportunities in Canada as are evident here in the United States—performance over providence.
Yeah, last question for me is, I’ll ask the same couple of questions that Jay and Mark added, but from a UK perspective. Anything different in the UK? You know, if you were talking to the UK government, you’d say that it’s different than what you said to the Canadian government?
Same answer. It’s time to look—the UK has an epiphany moment. Do the US-UK deal. Make it deep in terms of IP, and make sure that people, on a temporary basis at least, are able to do business in both countries and share the benefits of having done some time in the US and some time in the UK.
And let’s have gains from trade from all that talent. Super excellent! And that’s our show for today. Thanks to all three of the Marks who participated in this recording session: Mark Bush, the Carliff Landegger Professor of International Business Diplomacy at the School of Foreign Service at Georgetown. I’m Mark Mason, our managing partner and co-founder. Mark Shulgin, a partner and co-founder at Intrepid. I’m the final partner and co-founder, Jay Agarwal.
Follow us on the Intrepid Substack at insights.intrepidgp.com, and you’re also welcome to rate or comment on podcasts and subscribe on your favorite platform: YouTube, Spotify, Apple Podcasts, and more. Thanks, everyone, for listening. Thank you, thank you, thank you, thank you!
Yeah, Mark, that was great. The views, opinions, and information expressed in this podcast are those of the hosts and guests and do not necessarily reflect the official policy or position of Intrepid Growth Partners. This content is for informational purposes only and should not be considered as financial, investment, or legal advice.