2026: The year the AI bubble bursts?

0:00Hello and welcome to the very first Sifted podcast of 2026. I'm your host Amy Lohan Sifted editor. And to kick things off today, I am joined by my colleague, Martin Coulter, Sifted's news editor, to discuss what we think will be the big topics this year and what you, our dear listeners, should be watching out for. Martin, hello. Hello and happy new year to you. Apologies in advance if either Martin or I sound a little bit sniffly. It's currently about minus two degrees in London, so we're both suffering a little bit there on our first day back in the office. I will be taking appropriate Kleenex break.

0:40Lovely. So, Martin, do you have your crystal ball at the ready? A crystal ball, tarot cards, tea leaves. I'm ready to go. I've got a list of predictions, thanks to the team that we've put together. Lovely. So, I guess the big, big question is, when, if at all, do we think the AI bubble will burst? What bubble? We don't know. So depending who you talk to, some people say this isn't a bubble. Some people say it is a bubble. And some people say it is a bubble, but it's a good kind of bubble. We'll get into it in a minute, but I think it's unlikely there's going to be some extinction-level event, famous last words, where Google and Meta and OpenAI suddenly fall apart because everyone wakes up one day and realizes that AI is useless or fundamentally unsustainable.

1:35As far as Europe goes, I think there's a chance that some of the buzzier companies that we've written about may not be able to fulfill their ambitious targets and either pivot, get bought up or shut down. But time will tell. Yeah, I think ROI, return on investment, is going to be a bigger word this year, isn't it? Because I think there were all these reports that came out last year and people are starting to question really, we spend all this money on AI, but what actually do the big enterprises, the big organizations that use it really kind of get out of it? And I guess feels like the past two years maybe have been very experimental and there's lots of new tools out there and forward thinking companies are very much encouraging their staff to try everything out.

2:24But I think this year it will be much more sort of like, OK, which one's actually working? Which ones do we not need to use anymore? You know, how is this actually making you more efficient in your job? And I think the companies that aren't really demonstrating that, the AI startups will have a tricky time, won't they? Yeah. A lot of churn, I imagine, in their customers. Yeah, I think last year there was a lot of momentum definitely inside legacy companies to get to grips with different AI tools. There was a rut, I think even speaking for myself, to try and get to grips with them and work out how they could improve productivity, how they could save you time, how they could save you money.

3:02um and i think this year you're probably going to see people moving from experimentation more into accountability and actually working out how tools like chat gpt or otherwise are helping you meet kpis and saving you money um i don't it'll be interesting to see the thing about But AI applications, which I think is different, is you sort of need a degree of original thinking in terms of how you use them. This is where like prompt engineering came from. So there's this kind of onus on you or your staff to actually work out how you can use an AI app and what its potential is. You know, what you can get it to reveal about your company.

3:50You're not going to hire someone because they're really good at using Slack. But you might hire someone if they're really good at working at how to use ChatChapuT. Yeah. And then I think another big question is, as maybe we start to see some players pull away and some fall behind in terms of their customer acquisition, is whether we see some mergers and acquisitions happen and whether perhaps the big tech companies might become a choir of these smaller ones. I think that makes probably the most sense in the vertical applications where a bigger AI company that isn't, say, already specialized in legal tech is a good example, might buy a legal tech company to have that specialisation and we might see that in other areas.

4:43Do you think we're going to see, do you think that's a likely scenario for this year? Yeah, I think so. I think there's definitely going to be more consolidation. A big theme last year revolved around the growing number of either acquisitions or reverse acquisitions, which has really picked up in the last couple of years, which is where a big tech company, for example, wouldn't necessarily buy a startup outright. but it would sign some kind of licensing deal or it would effectively hire all of the top talent. Talent is kind of one of the key drivers of the market at the minute because it's in such short supply, and that's where you see Meta and others paying such a huge amount for AI scientists, et cetera.

5:25Outside of M &A, I think the other thing that we might start to see is cloning or basically copying, and we'll kind of find out how well defended the moats around some of the buzzier European AI startups, whether this is Eleven Labs or a Synthesia or a Lovable, and whether the market that they've targeted they can defend against if an Apple or an Amazon or a Meta decide they want to get in on that kind of product offering. Yeah, and on Lovable, I guess, probably the company of the year 2025 in European tech. We saw, Sifted saw a leaked pitch deck from the company, which suggested that this year, 2026 was all going to be about margins for them.

6:17So this pitch deck said that they'd be aiming for 65 % by the end of the year, based on things like efficiency gains in their model, and that they're also planning potentially things like an app marketplace, which would help by taking a cut from developers who are using them. So I think it's going to be a good year to keep following the Lovables, the Synthesias, the 11 Labs very closely, as you say. Yeah, they raised another 330 million towards the end of last year. As you say, our colleague Freya got a hold of a pitch deck that they were circulating, which seems to suggest they are planning to launch their own app store.

6:58We think you'll be able to access pay for apps that have been built using Lovable. It's kind of interesting because OpenAI actually announced a similar thing, an app directory, last month as well, which had some mixed responses. Overall, I think it kind of speaks to a theme that started to come through last year, which was around how being a great builder of AI models is not necessarily enough now, and it's becoming much more about the actual applications you can build on top of those models, but also the platforms like an app store that you need to distribute them. Yeah, I'm also intrigued to see what this means for sort of the earliest age European VCs because it feels like it's becoming more and more a game in which there are a handful of humongous players and they are raising so much money that there's very few investors in Europe who can actually, you know, follow on in those rounds and get involved.

7:58So I wonder if we'll see less interest in kind of newer AI agent type startups or whether perhaps this year the earliest stage investors will be more about, I don't know, various ways into the infrastructure play or other aspects of AI that are obviously essential to the broader ecosystem them where they can you know still have relevance with really basically a very tiny check yeah we'll have to see how it plays out um it's not obvious that there's going to be quite as much money pouring into ai this year as there was in 2025 um there's been a drought in the ipo market the exit market for a while which means that a lot of investors are waiting for their returns that they can't necessarily reinvest um there's been this kind of flight to quality um among lps so some of the bigger brand names like the sequoias the andreese and horowitz etc they're not necessarily going to struggle to raise funds but if you're a new player or you know relatively inexperienced um it's going to be a lot harder ai has kind of been the shiny object for the last couple of years but with more emphasis on actual returns there's no guarantee there's going to be quite as much cash floating around for that this year.

9:19So what do we think are the new shiny objects of 2026? What are the sectors that we're going to be paying a lot of attention to? Well I think it's still AI but I think it's very specific areas of AI. There's a lot of hype around AI agents at the moment. these programs which in theory will be able to automate all kinds of tasks you will be able to type in a prompt and say i want to fly to poland uh this time of year and for this kind of price and it won't just find them but it may even be able to book them almost automatically if you give a permission to do so and that's just one example of the kind of tasks that these agents can do through booking flights, building apps, booking restaurants, etc, etc.

10:11Other than that, it's AI infrastructure. NVIDIA has kind of led the way and has become the most valuable company in the world off the back of supplying AI companies like OpenAI and others with its chips. We're seeing companies in France, companies in Germany, and also the UK with N-Scale. I think the European tech sovereignty put means that there's a lot more resource being poured into homegrown solutions, data centers being built locally by local companies. So AI agents and infrastructure. Yeah, I wonder if we're also going to see more emphasis on critical infrastructure more broadly. we're recording this what two days after Trump I think shocked the world by airlifting Nicolas Maduro out of his own country and taking him to the US it feels like geopolitical tensions have gone up a whole other notch and if you the listeners if you heard the episode the last episode we recorded with Fiona Murray the the chair of the NIF I think she made it abundantly clear that these things, whether they're undersea cables, whether it's things like data centers, whether it's energy production in Europe, are very, very important, pretty under looked.

11:35And whether it's the surveillance of those things or, you know, alternate technologies, there's a wide open landscape in which, you know, Europe needs to invest in protecting these things and having backup plans for if tensions rack up a whole notch more in Europe or the rest of the world. And so I wonder whether that will feed through to the kind of early stage investment landscape or whether that remains more of play for much bigger industry and government. I feel like there's always a way that startups and VCs play in any of these. You know, whenever there's a lot of money going towards something, I feel like smart, innovative new companies spring up too.

12:21So I'm very interested to see if we have a bunch of that. And obviously that kind of, depending on exactly what that technology is, if it's more hardware or software, might require a very different kind of funding roadmap to, you know, your AI startup, for example. So I think that's a very interesting one to watch. There's also obviously lots of deep tech areas that have been picking up steam for a while. Quantum robotics is something that we're increasingly writing about. has sifted lots of big companies there well-funded you know the top tier the sequoias and whoever investors behind them so I think we'll be we'll be following that a lot and then another area I'm interested in is sort of the VCs who might spring up with slightly new models I've heard of well we've seen an increasing move over the past few years towards solo GPs and I think an interesting thing that started happening towards the end of last year was that some of those um you know individual investors raising funds by themselves were also operators at startups uh which also ties into the trend of you know we know that many of the founders um of some of europe's biggest companies are also some of europe's most active angel investors and we've heard that some of those might be raising more sort of formalized funds.

13:51We've also heard of people working at, you know, some of the really big tech players who are very active angel investors. We've also heard some of those might be raising new funds, new sort of fund of fund models. I feel like we really are this divergence between individuals or very small sort of emerging managers doing things quite differently, being quite nimble, sometimes being like very hyper-focused on a particular thing. And then the other end of the scale, you have the enormous funds, the index, the Excels, Andreessen, the general catalyst who have, you know, huge, huge, huge funds, lots of different, you know, covering lots of different stages.

14:34And I think the question, this has been a question for a while now, and I think it remains, is what sort of happens to 200 to 500 million euro VC fund on its third, fourth, fifth generation. are they as attractive to LPs? I'm not really sure that they are. I think that's an interesting area to watch what's kind of longevity to some of those more like second tier or funds that fit into a certain size that's perhaps less it's less easy for them to raise capital basically. I think it's that kind of individual personality that you're seeing kind of being brought forward and there's a parallel to be drawn with media as well.

15:18You know, there's a question over newsrooms of a certain size when you as an individual writer can make X amount just writing your own substatic, for example. And I think that cuts across to VCs and you have some cut across with VC media hybrids like 20VC and Harry Steppings obviously being a very successful face of that. So yeah, I imagine it's a trend that we're going to see even more of this year. Yeah, it's a lot easier to raise a 10, 20 million euro fund from a bunch of individuals and family offices, etc. When it comes to the next size up, you are having to tap into the institutionals. And I think they, as you mentioned earlier, you know, they haven't had money back for a while.

16:03And so I think a lot of them think, well, if I can get into a creandum and index, a plural, a general catalyst, I'd rather do that than these other guys. and on media as well we're very much paying attention to the likes of Seb Johnson and the ETN boys doing their sort of video first interviews so we'll be watching that we've also heard of a few others that might be springing up there's what Mike Butcher's doing with Path founders Mike Butcher used to be an editor at TechCrunch so I think that's That's a good thing. It's important for the ecosystem to have different types of media publications, different ways to consume news and information.

16:48I guess a lot of those are very sort of upbeat about European tech, sifted maybe a little bit more down the middle with our cynical journalistic roots. But it does feel like last year there was a lot of buoyancy and excitement around European tech, especially around groups like Project Europe, which kind of came out of 20VC and does hackathons for very young founders around Europe and initiatives like EU Inc, which is obviously trying to get the kind of simplified legal structures around startups that would, or new company formation that would apply across the EU and hopefully make it easier for companies to raise money and do all sorts of things.

17:38But how do you think that's going to pan out? Are we feeling optimistic about EU Inc's prospects or not so much? Yeah, so the EU has moved to adopt what they call the 28th regime, which is what EU Inc. has been campaigning for. The idea is to roll out a single harmonized set of rules for innovative companies across all of Europe making it easier to invest across borders make regulating the entire ecosystem more uniformly easier speeding up the process of registering and incorporating a company with this centralized register. At the moment, the problem is it looks like the EU is planning to roll this out as a directive rather than a regulation.

18:33Putting my EU legislative cap on for a moment. A directive basically sets a goal for a member state to achieve, but they then have flexibility about how they actually enact that. so the problem is opponents say that rolling the 28th regime out as a directive will not achieve the goal at all it may just leave to 27 different countries rolling out their own version of this which then may just make even more paperwork and make the whole thing even more complicated so we wait and see whether or not these criticisms are going to fall on deaf ears or not whether or not it might get changed into a regulation which in theory would make the whole thing a lot more simpler.

19:19Yeah and there's also speaking to members of the EU Inc team towards the end of last year they were telling me that even if that happens they still hope that there might be a situation in which for example France, Germany, Spain and some other country you know all agrees to do the same thing so it might not be all 27 countries agreed to this to start off with but you might get a sort of cohort of countries that all agree to do this so they're they're an optimistic bunch they've been working hard at this for about two years now and they yeah they do hope it's gonna get across and you know it has a lot of support doesn't it from basically any you know big big name in the ecosystem well-known founder well-known investor has you know put on the eu inc cap and taken a selfie and publicly pledged their support on LinkedIn.

20:12So, you know, what that counts for in the EU, I don't know exactly. But the ecosystem seems pretty united in thinking this is a good idea. Yeah, massively. You know, I was no Brexit here. I'll let it be known. But if you take any time looking at the amount of bureaucracy or the amount of paperwork that companies do have to deal with in the EU and how things that are relatively simple in places like the US, indeed China in some respects, versus moving or allocating capital in Europe. It's a faff. So hopefully this will make it a bit easier. Yeah. What do we think will go bang this year? You know, where is a lot of money going to have gone and, you know, not much come out of it, do we think?

21:00I reckon some of those sort of AI infrastructure companies we're paying I'm not I'm not saying that I think this company is going to go up in smoke but I'm going to say I think we're going to pay a lot of attention to it is Nscale the the UK company that raised one billion dollars towards the end of last year it's already talking about an IPO it's not that many you know years old it's It's had a lot of investment from NVIDIA and it's one of those companies where, you know, which is in this question of there's a lot of sort of circular deals being done with these big AI companies, you know, big model makers, big chip makers, etc.

21:42And I feel like there's going to be some fallout from that. And a company like Nscale, which is sort of involved in that, I think there might be some drama ahead or certainly some sifted headlines. Yeah. So I think in the last year, so I think two points, both of them related to AI, unsurprisingly. but a number of prominent companies in europe made a big deal out of their arr their annual recurring revenue last year um but we're approaching a point this year where we're going to find out how many of those subscriptions actually stay on um and those numbers could go down quite dramatically and that might have some knock-on effects for valuation or invest interest or even the viability of certain companies um to your point about ai infrastructure i don't know if it will go bang but i'm curious to see how it plays out you have countries like the uk putting nscale front and center um with its kind of ai strategy you have mistral um building a data center and kind of being positioned as the national champion and quite key to France's long-term technological sovereignty or security and these are very young companies um it's an interesting move to put them front and center of your kind of long-term uh national integrity and there's also just you You know, there's a whole list of companies that have raised a lot of money, often multiple funding rounds in one year.

23:29They've grown incredibly fast. You know, they've got very high valuations. And as we saw during COVID, I think there's just the question of whether they can keep up that momentum. And I think there will be some inevitably that can't. and then they're back into down round territory or you know if if there's a viable competitor that investors could also put their money in that's raising at the same time you know are they are they going to be able to to get those deals done so i think we'll see some sort of pull away and some maybe not not have as many funding rounds as much interest in quick succession yeah Yeah, it'll be interesting to see.

24:13Lots of people, if I start talking about European tech sovereignty, start looking a bit sleepy. But I think where it genuinely gets interesting, as I say, is this idea of them being pulled right to the heart of different European countries' kind of almost national security kind of strategy and a question over whether or not these companies kind of become too important to fail. Do you end up with the state effectively subsidizing them? What does it look like for France if Mistral collapses? What does it look like for the UK if N-Scale collapses? There's a huge kind of credibility question in the long term, not just for these companies, but now for the countries in which they were built, which is a new development.

25:04Yeah, and arguably the same might happen if the defence scale-ups play their cards right, if they're really well integrated into some of the big national primes, if, as we were hypothesising earlier, geopolitical tensions ratchet up even higher and countries are spending a lot more of their GDP on defence, then the likes of a quantum systems a helsing uh some of the satellite companies and the space companies um even some energy companies could perhaps find themselves in a similar situation which yeah as you say is not not something we really saw with you know when all the money was going into speedy grocery and um scooters were not not quite as essential from a government point of view we wouldn't have bet the future of the country on the delivery's performance thank god um what about what's coming down the track from the eu and the uk on the kind of policy front is there anything sort of big this year that startups tech companies of any kind investors will be needing to pay attention to uh yeah so in the uk we have um kanishka narayan uh the ai minister who's overseeing the kind of tech strategy going forward he's recruited some ai ambassadors to kind of help steer the ship including tom blomfield the co-founder former ceo of monzo and james wise from balderton is going to be chairing the sovereign ai unit which is recruiting some local investors to try and work out how best to invest government money and And Einstein is going to be recruiting some VCs locally as well for that.

26:58In the EU, on top of the 28th regime that we've discussed, we've also got the ongoing saga of the AI Act. This is something which has been kind of gone over, chewed over in various different forms for the past five years. has been the cause of many arguments across the region. There were going to be some rules around what they call high-risk applications this year, those involved in education, policing. There were lots of rules around copyright with companies like Meta, Google, and some European startups. In theory, they were supposed to hand over the training data, so there would be a much more understanding around what they were scraping from the internet.

27:52Maybe there were copyright claimants who would potentially pursue legal action, etc. But that has now been pushed back, and I think the expectation going forward, based on the amount of noise that businesses both in America but also in Europe have made about the law, and also the changing dynamics, as we've said, with the Trump administration and the relationship between the EU and the US changing. I think we can expect to see the AI Act watered down, tweaked even more, so it probably looks quite different by this time next year. Yeah. We also had a fun prediction from our French tech reporter, Daphne, which you can read on our Sifted Team predictions article, which you can find on the website and we'll drop a link in the episode description.

28:46Daphne pointed out that Emmanuel Macron will be stepping down as president in 2027. And she predicts that we might see a big name from tech run for that rather important role. Her money is on Xavier Niel, the founder of Station F, the big co-working campus, the very active VC fund Kiba Ventures, amongst a million other things and a very, very rich man. So will we see a tech billionaire take the reins of a European nation this time, coming up this time next year? We'll have to see. Let's finish with a bit of a quickfire round. And which company, person, fund will you be tracking like a hawk this year?

29:38So I had to, well, I was going to say Hero Capital, which is the VC firm that Nick Clegg, former Meta executive and deputy prime minister of the United Kingdom, recently joined. But off the back of what you just told me about Xavier Neal, I'm increasingly fascinated. he's co-owner of Le Monde and he's also co-owner of the rights of the song My Way yeah worth having a look at the Xavier Niel Wikipedia page for sure is there a country we think deserves more attention? yeah I think there's a couple I think if I was going to pick one it'd probably be Switzerland they've got a lot of deep AI expertise there they obviously have ETH Zurich, the university where a lot of that expertise is there's a growing number of companies coming out of the country and more investments steadily pouring in and it's also a hotbed, as a result of all that particularly the university system there's a hotbed of talent which actually gets paid reasonably well so I'm curious to see, particularly with Jan now starting his own company whether or not he has any operations, whether he's poaching people from Switzerland.

30:58So that's one I'd be keeping an eye on. And another one, again, across the team, a country that hasn't really had so much momentum for a little while but now really is picking up steam is Spain. There are quite a few AI companies there raising a lot of money, more and more funds on their second, third, fourth, whatever generation. So that's another one we'll be paying a bit of attention to. And finally, is there a major acquisition that wouldn't surprise you? Yeah, there was a lot of speculation last year that Apple was going to buy Mistral, which it's not entirely clear where that came from, but it got taken off the table pretty quickly.

31:45Mistral came out and said, we're not for sale. Don't listen to the rumours, et cetera, et cetera. I think of all the what were once the leading kind of AI lights in Europe, I would not be surprised to see AF Alpha get acquired. It's lost some momentum. There's been a change in leadership, multiple pivots. We wait and see what they do this year. but they have some great talent on board and i wouldn't be surprised if uh a big tech like say an apple or similar thought why don't we buy this for possibly a discount so we'll see and then a final prediction from me i guess is that there's been a lot of money very seemingly hastily thrown into certain companies.

32:38And also there's been a lot of investors struggling to raise, we know, for a long time. And I think in both of those situations it means that there's been a certain lack of due diligence. What are you talking about? We'll have some pretty interesting stories this year when it's proven the emperor has no clothes And that might be that a company was really not worth what anyone thought it was, perhaps fudged the numbers or, you know, just really didn't fulfill people's hopes and might go up in smoke. And then I think also on the investment side of things, we wrote over a year ago now about a fake family office called Gloucester Circus, a family office that was promising to invest in VC funds and seemingly never had the money to do so.

33:33I really would not be surprised if there's a whole bunch more Gloucester Circus type actors out there. And we are hoping to find some of them. So if you have any tips, get in touch. I'm amy at sifted.eu. And that is all we have time for. Martin, thank you so much for joining me. Thank you very much. If you want to hear more from us, please sign up to Sifted's daily newsletter. We will drop a link in the episode description. We'll also drop a link to some of those prediction pieces we were talking about. So have a good read of those. And as always, please rate, review and share the podcast to help us get some more listeners.

34:06This episode was produced by the delightful Maya Durampal-Hornby.