77. The Rise and Risks of Sovereign Wealth Funds

0:00Welcome to The Difference Engine, the show for tech founders, investors and innovators.

0:09All right, Jono, you tell me what's coming up today. We'll explore the often contradictory relationship between modern government and the international technology sector. But first, we ask a question that not too many people ask, but you need to know, will sovereign wealth funds drive the future of technology?

0:33Many of us who are inside international politics or investments have no idea what a sovereign wealth fund is. Can you help? What we're looking at, a sovereign wealth fund, is a state-owned investment fund that manages national assets, typically, but not exclusively, from all revenues, trade surpluses, or foreign reserves, or even general taxation, but to invest it for long-term public benefit. Right, so it's like the savings account, when you don't need it day-to-day, and you're making lots of money, let's say, back in the day from North Sea oil, as Norway still is, and we stupidly are not, you sort of put a little bit in the piggy bank, and then you invest it in somewhat riskier ventures.

1:18What's that all got to do with tech? Oh, well, it's interesting because sovereign wealth funds, and we're going to call these SWFs, sovereign wealth funds, they've been in the news recently with Europe, and that includes the UK's investment in the biggest ever European seed fund. And the lucky winners in this were Ineffable Intelligence. but closely followed by the claim that the same 500 million UK AI fund, which of course was set up to back homegrown, inverted commas, technology, has been caught granting UK taxpayer-funded support to a startup based in Silicon Valley, of all places. Now, in this case, the Labour government's sovereign AI fund has issued credits for computing power to a firm called Odyssey ML, a startup developing so-called walled models that can understand the physical environment.

2:25Now, this company is incorporated in the US and headquartered in Menlo Park, which if you've been around tech for a couple of nanoseconds, you will know is in the center of Silicon Valley and the home of Facebook owner Meta, other companies like E-Trade, and of course the Sand Hill Road VC mob, including Kleiner Perkins and Adresen Horowitz. Now, well, some members of the team at Odyssey ML are based in London. For instance, Jeff Hawke, the CTO, and co-founder Oliver Cameron. but Odyssey's other co-founder and chief executive is based in Palo Alto. Additionally, we found out two other startups that received the credits had UK operations that were ultimately owned by holding companies incorporated in the US.

3:22And the startling reality of this is there is in fact no requirement for founders who get computing funding from the sovereign AI fund to base themselves in Britain or even keep their companies here. I do detect some whizzle words here, credits. What do we mean by credits? I think we mean giving money to data centres. Yes, I think we mean giving them pocket money to give to the local shop, i.e., we're the state, we're going to give you a load of money, here's some money to spend with data center suppliers so you've got computing power right so the data center suppliers um you know we're talking about jensen et al are themselves us owned so so what i detect here is money moving from the uk overseas quite rapidly isn't the idea that software uh that um swf sovereign wealth funds should be there to make sure nations don't miss out on new tech categories not to subsidize other nations, or even, God forbid, whoever's funding it becomes a tech leader, not just subsidize the ambitions of other states.

4:27It seems to be a contradiction of the basic idea when you try and put it into practice. Yeah, so this is basically how not to create categories, how to help other people create category leadership. Absolutely. And then there's a strange case of the new boss, who's very much like the old boss, when we look at who's actually benefiting from all this, apart from all the money siphoning off to Jensen et al. Let's just say when it comes to the profits of managing the business of sovereign wealth funds and AI funds in particular, regardless of wins or losses the taxpayer makes, the apple stays close to the tree when you look at the management team.

5:04Let's have a look then at five arguments for and against using software wealth funds, many people do, to drive national category leadership. So the first thing is, and I use the term advisedly, is sovereign wealth funds provide patient capital, allegedly for frontier technologies. Now patient because you don't get your money back for a while. Although in this case, it could be a double entendre, given the sick person here might be Europe's innovation. But then many breakthrough categories such as AI infrastructure, quantum computing, advanced semiconductors, nuclear fusion, synthetic biology, these are very long-term bets.

5:43They are highly likely not to pay off unless you're extremely patient. So unlike traditional venture capital, which is always looking to return five or 10 years in, sovereign wealth funds can tolerate things like long commercialization timelines while you get market fit. They can deal with cyclical downturns. They can deal with infrastructure scale investment and the fact that they need to return less liquidity because of the patience that we talked about means they can wait things out. And examples include Mubadala over in the Middle East investing in semiconductors and deep tech. Temasek out of Singapore, which has been investing for years on this basis, and that backs biotechs and digital infrastructure, as we know.

6:25And then the PIF, Public Investment Fund, also from the Middle East, looking at EVs, gaming, and increasingly AI ecosystems. The argument is these brand new tech categories often fail, not because the inherent technology is weak, but just because private capital markets are too impatient for a return. And they can do even bigger things than that. I mean, in theory, they can catalyze entire ecosystems, not just companies. So if you think about it, traditional VCs optimize for company level returns and SWFs often optimize for national capability creation. or at least that's how it should be in theory.

7:04So in theory, again, this allows SWS to support a variety of systems and concepts. Universities, managing talent migration, basic industrial policy, part of which is building out of data centers, which we do believe are the engine of the AI economy. Energy infrastructure, also slightly related to data centers because they'd show up a lot of power, regulatory sandboxes, and even procurement ecosystems, which is certainly how the US helped develop its own technology industry whilst pretending it wasn't in any way involved in industrial policy. Now, why does this matter? Well, it matters because category creation usually requires complementary systems.

7:53For example, an AI ecosystem needs compute, energy, search labs, startup capital, and procurement customers. An EV ecosystem needs batteries, charging networks, minerals, manufacturing, and consumer financing. So again, in theory, SWS can coordinate these layers in ways that fragmented market investments cannot. Unfortunately, they may not produce a lot of jobs, but that's a side issue. Number three of the four argument for SWFs is they can accelerate strategic diversification. So commodity-dependent states, think the Middle East, increasingly use SWFs to transition into post-resource economies.

8:38And category creation becomes, therefore, economic diversification and national resilience and even, as we've seen recently, geopolitical positioning. For Gulf states, a technology investment is often viewed as preparation for a lower hydrocarbon, more industrial diversified future. Makes a lot of sense, right? So the Qatar Investment Authority is investing in digital infrastructure. We've seen the largest sovereign wealth fund of all, bizarrely, a lot of people don't know this, is Norway's government pension fund, which has globally influenced sustainability and climate tech capital allocation through its active investments and its active shareholdings.

9:15and much of the activity of Saudi Arabia's PIF. The positive argument here is that SWFs can absorb national surplus capital when it's accreted and then redirect it later into future-oriented productive sectors. What's not to like about that? Unless, if like the UK, you spend the North Sea tax money providing welfare for the former workers of declined industries such as coal mining, and when you look around to see where your savings are, there ain't none there. Less said about that, the better. Now, back onto the positive. So, if you're a country that's actually got itself together and actually built a significant global-sized SWF, there's always the issue, if you are trying to manage your economy going forward, that emerging technology categories, which often end up being the crowning heights of the economy will collapse during risk-averse economic cycles, sort of like now for some technologies.

10:18Well, basically any technology that isn't AI. So if you think about the way technology cycles work, SWS can continue investing when IPO windows closed, and they've been pretty shut for the last few years. Venture markets contract. Again, that has been happening. And of course, something else that's been happening, interest rates have been rising. So if you can smooth this over as an SWF, that continuity can preserve strategically important technologies through difficult commercialization phases. Supporters argue that this counter cyclical role, it's especially important for big picture stuff. Now, what do we mean by big picture stuff?

11:06Climate tech, semiconductor independence, defense adjacent innovation, and of course, advanced manufacturing, on which we all depend for wealth. And I think the defense adjacent innovation rings particularly true in an antebellum where You know, you feel like a war's coming. You better be ready so SWFs can invest there. The fifth argument for SWFs to say they're amazing is that they're necessary because technology competition is already geopolitical. China uses state-guided capital aggressively. We've seen that with Huawei and other vendors. The U.S. deploys industrial policy through mechanisms like the U.S.

11:46Chips Act. and its defence procurement, which is legendary and leads the world. Other nations need institutional capital vehicles just to remain competitive. From this perspective, then, SWFs are not market distortions or corrections. They are a strategic necessity. Ah, they may be. But, of course, there is a case against sovereign wealth funds as a driver of technology category creation. and I'd like to kick this one off with something I think is self-evident. Governments, shock horror, are often poor category pickers. Now, critics, I'm not saying I'm one of these, or may or may not be, they argue that technology category creation depends on experimentation, failure, and what we call decentralised discovery.

12:39You don't slam a whole load of scientists into a room and expect to get amazing things. So in reality, SWFs, because of their structure and because of who tends to run them, may overfund fashionable sectors, right? And that is because they need to be seen to be doing something. So clearly at the moment, there's a danger that SWFs are overfunding AI investments, right? So that is a form of politicization of capital allocation, i.e. they are not investing for pure economic or technological advantage. They are investing to be seen to be doing that. And another sort of rather nasty aspect of this is that it's really mistaking national prestige for market demand.

13:34And because they are so big, they can create entire ghost ecosystems, which appear to be alive, but they're not actually doing anything. So just so that you, dear listener, doesn't think we're just going off or one or miss, if you think historically, if you've been around for a few years, some of the overinvestment waves we've seen are, for instance, smart cities were all the rage. and metaverse infrastructure, some clean tech cycles, I would say, and various national champion projects. And just have a look around at what's happening to some sovereign wealth funds, I'm not naming any names, who are cleaning house at the moment in terms of what it is they really want to get behind.

14:23So the concern in all of this is that bureaucratic incentives reward visibility and scale, I look at my big stuff, rather than the genuine private market fit, top-down management by civil servants just doesn't work. Yes, absolutely. The second point against sovereign wealth funds is that state capital paid for by you and I can distort innovation markets. This is where governments and markets collide. And large SWF funds can, of course, inflate valuations. They crowd out private discipline. So, you know, if you're getting lots of a spigot of government money, why would you apply that much discipline?

15:08And then they encourage capital inefficiency as a result of that and sustain weak companies longer than markets otherwise would. Thinking something close to the multiverse there. When abundant sovereign capital enters immature markets, startups optimize for fundraising and tell stories about fundraising instead of focusing on customer value, which is death for category creators. Critics argue that this contributes to excesses in late stage venture markets, which could and should have been creatively destructed by then. Mobility of startups and has seen some AI infrastructure speculation. perhaps very close to home here in the UK, which will never and could never take off.

15:54Yes, indeed. And this is undoubtedly linked to the idea that in the wrong hands, the strategic motives behind what an SWF is supposed to do is actually in conflict with innovation efficiency. So, I'll give an example. So, as a solver in wealth fund, your political drivers may be national employment goals. So it's all about creating jobs. Now, that may be at odds with technology efficiency. Geopolitical leverage. Do this because it's going to be good for our country, but it will also help us in our relationship with another country or power group. Soft power, often misrepresented and misunderstood, helping other countries do things.

16:51Again, if we develop in this way, it may help somebody or some regime that we're trying to influence develop a better relationship with us. Of course, there's just general domestic industrial policy. By the way, industrial policy is usually made up by people who have never, ever been near industry. So these goals are often not aligned with an idea of competitive neutrality, open innovation ecosystems, which we now have been behind some of the major innovations in patient technology in recent years, and, you know, this is a bit finance, but efficient capital allocation, i.e. are you putting your money in the right place to get the right returns?

17:40So for example, technology localization requirements can reduce global collaboration and increase fragmentation. And of course, if you're getting anywhere near defense tech, that is a clear and present danger to use the language of that sector. Whilst we're on the topic, the fourth reason to argue against sovereign wealth funds is that the The political risks that they take can undermine trust. Let's remember, technology categories rely heavily on international talent. If you look at the number of US leaders, category leaders, almost all of them founded by immigrants. Cross-border research. This is why people get hot under the collar about things like the Erasmus situation here in the EU.

18:27It requires open capital flows, as we're seeing from China. and how that can affect category creation. And it relies on strategic partnerships. Now, it's quite possible that if you go for a sovereign wealth fund-led investment policy, that can introduce national security concerns, data sovereignty conflicts, and regulatory scraps, as we've seen very recently with the EU and elsewhere and big tech. And this has become particularly visible in telecommunications back in the day, semiconductors presently and AI right now and will be for a long time and the problem particularly with AI is it moves so fast that you're just the governments are just not able, the sub-governmental is not able to keep up with the speed or pace of change required and of course obviously it applies to defense related software.

19:21So innovative management which is needed for these category creators may be operating with one hand tied behind its back in relation to more conventionally funded, commercially funded competition. So number five in this list, you'll be glad to know, this is the tenth, though it's number five, reasons against SWF is that they may reinforce what we call techno-nationalism, right? So a broader criticism of SWFs is that they contribute to a number of things you don't really want if you're going to move forward quickly. The first is fragmented standards. We've seen this right down to what is the standard for the plug that you put your iPhone in.

20:11Competing tech blocks, i.e. one standard is completely adopted in a third of the world. That is going to stop global level innovation. you also see wastage if you don't get this right because you will inevitably see to support this duplication in supply chains and that also has a sort of combining effect that continually reduces globalization so what can happen when you've got this you know sometimes toxic mixture of national ambition and technology development is that you don't create globally interoperable categories you create parallel national ecosystems and of course what that is going to do is reduce innovation efficiency it's going to reduce scientific openness and of course it's going to reduce the economies of scale that otherwise you could create.

21:23Right, so quite a list there of for and against arguments for sovereign wealth funds interfering, or as some people would say, aiding tech innovation and category creation. Let's just recap them if we can. So on the plus side, the reasons why you would think that a sovereign wealth fund would be useful is that they provide patient capital. When that's done right, they can catalyze across entire ecosystems. And that means from a national point of view, they accelerate strategic diversification of your industries. And they can help in downturns like right now to stabilize the funding, to stop people being rash and switching it off too soon.

22:02And they compete with state-backed ecosystems to provide an alternative means for categories to get created. That's all lovely. On the negative side, we all know, we see it today, governments pick categories poorly. They're not a market. They are some ill-informed, normally non-technical people trying to pick out winners. States capital distorts competitive markets. The fact that there's a big funder in the market inevitably distorts the market. And the motives conflict between innovation and efficiency. Lastly, the political risk undermines a lot of trust in sovereign wealth funds. and they can become and have become recently, in our view, political footballs.

22:52And this means that they reinforce techno-nationalism. And as we all know, the last thing you want to do with the category is limit its scope for growth. Oh, yes, the total addressable market question. We always want a big one of those. So what does this tell us? Well, I guess what it tells us is there is a central tension behind all of this. And the debate pro and against ultimately centres on one question. And that real question really is, should transformative technology ecosystems be shaped primarily by markets or by states with strategic capital? Now, the strongest real-world pro-SWF argument, I think, is that some technology categories are simply too capital-intensive and strategically important to leave entirely to short-term or comparatively short-term private markets.

23:50The strongest real-world anti-SWF argument is that governments consistently struggle to identify which technologies deserve sustained support, and political capital allocation often becomes inefficient, right? So how do you use sovereign wealth funds in reality? The truth as usual lies somewhere in the middle. A balanced framework often suggests sovereign wealth funds are most effective when they, for instance, co-invest with private markets rather than replace them. That's absolutely definite because you get the discipline there. They focus on infrastructure and ecosystem layers, i.e. they are enablers.

24:37They and the people that run them avoid micromanaging product decisions because they've been shown to be absolutely hopeless at doing that. And the people that they have put in charge to actually manage the shop end will leave, as we've seen in the UK. They also need to use transparent governance. And by that, I mean they have to be clear about what they are doing and why they are doing it. They must maintain commercial return discipline. them. As soon as they start looking like a money tree, you're lost. And they have to choose to invest across long-term horizons. The market can sort the other stuff out.

25:25So if you think about it, in this model, sovereign wealth funds function less as central planners and more as what I call market shapers. And that's about reducing friction for categories to emerge and absolutely without fully directing them. You may say this is a bit of a cop-out, I'm sitting in the middle, or I think it's exactly what we should be looking to achieve.

26:00Here we are, as we're getting up to the middle of 2026 in the UK, and we appear to be living through another government in chaos. So what we want to do is to have a look at what's happened with the UK's tech ambition between the landslide victory that the current administration achieved and the current debacle. So the story of the current administration's relationship with Britain's tech economy is at its heart, I think, a story about ambition colliding with dependence. Yeah, you could say that. When our current lot came in just a couple of years ago now, not even, they inherited what they described as a pretty stagnant economy by all accounts.

26:53Productivity was weak, so was growth. And the national mood in the UK was that we'd miss the great platform engines of the digital age. I think it just occurred to people that Meta had stolen everybody's data. But the new administration responded with almost evangelical zeal. Let's go for artificial intelligence, they said. Let's build out data infrastructure, they said. Let's look at advanced computers, and let's look at these to become the engines of national renewal. Very quickly, though, their growth story disappeared. and there was a lot of cautious u-turning and managerialism and what we'd come to expect from i don't know the um civil service class it was just it seemed all just national building rhetoric and at london tech week uh 2025 which we famously asked uh the value of he you know the prime minister framed ai as the mechanism through which britain quote would put money in your pocket well I think we know what happened to that.

27:53It was going to rebuild public services and create opportunity in post-industrial communities. This all looks entirely ironic at this point. It absolutely does. And if you think about it, from the very beginning, the government's technological vision revealed a contradiction and a very big one. The current administration spoke of the language of sovereignty and national capability. but the reality was the machinery of its strategy depended heavily, if not entirely, on American hyperscalers and Silicon Valley capital because the British state no longer possessed national computing champions comparable to the global leaders of early industrial eras.

28:42There was no sovereign cloud giant, no domestic AI infrastructure, no titan capable of competing with google microsoft meta nvidia or amazon or anybody else so labor's growth mission increasingly has become a courtship operation right so we were the unedifying site of of government ministers and advisors cultivating close relationships with large technology firms promising regulatory flexibility, planning reform, energy access, and AI growth zones, like Barnsley, in return for investment commitments, right? So 2025 and into 26, critics of civil society and even parts of Labour's own coalition began arguing that the relationship had become, what's the word?

29:39Oh, asymmetrical or asymmetric. Westminster, for those of you who are not in the UK, that means government and civil service, was no longer merely partnering with big tech, but orientating itself around the priorities of those firms. Now, the fact that the former Deputy Prime Minister Nick Clegg had become the mouthpiece of Meta really only served to fill that suspicion. Yeah well I mean it's not the only one let's not forget Mr Sunak who to give him his dues did actually kick off an AI summit for the UK which did manage to get all the big names including Musk etc over here. He's now the mouthpiece if you want to call it that for Open AI So, you know, what goes around comes around.

30:29And all of these public appearances with Starmer and Jensen Wang, you know, all they ended up doing was praising Silicon Valley investments. And they celebrated these multi-billion pound pledges for data centers and AI infrastructure. We'll recall what's recently happened with OpenAI's Sunderland data center. That got shut down. And instead of which, we get a story in the Times of all places saying, as a replacement, don't look over there, look over here, because OpenAI is going to build some offices in St Pancras. Not the same thing at all. It was sort of sad seeing the government sort of portraying Britain as a democratic and stable gateway for global AI deployment into Europe.

31:10Well, you know, what's new? The UK has been the technical aircraft carrier for the rest of Europe for the entire history of IT. but you know that there's some other weird effects here now there's for years organizations which have taken quite a lot of state money ironically such as founders forum group and tech nation um had attempted to nurture indigenous startup capacity um you know we're talking you know encouraging founders encouraging accelerators regional innovation networks and other scale-up ecosystems that could produce British firms of strategic significance. Yet, it's quite clear that many founders have felt under the current administration, the center of gravity, i.e.