Andy Leaver, Arqit: AI Isn’t the Disruption Yet. Quantum Unlocks What Comes Next

0:00That unlock moment is going to be when AI and quantum come together. And there's this, I mean, the people listening and watching this go look at how LLNs become LQMs that allow then you to bring that true reasoning and thought process into quantum. Because the power of quantum is that it can come up with all the conclusions at the same time. Yeah. So it can then say, what is the drug that we should pursue to put into trials rather than like, here's a hundred possibilities. And that computing power, it then becomes almost unlimited. And I think that's going to be a real pivotal moment in not just quantum, but AI as well, when you put the two together.

0:35That's going to be a massive unlock. With AI, you can launch attacks that will keep self-modifying and trying to find weaknesses millions of times an hour. It's truly kind of frightening that you're going to be under that sort of scrutiny at some point. You may have seen this in the press. They call it Q Day. And Q Day is the day where quantum can break all-known encryption using Shor's algorithm. It may have happened. We don't know because Notion States may choose not to publicize it.

1:00Andreas Munk Holm:Welcome to the podcast, everyone. Today, I am sitting down with Andy Lieber. He is the CEO of Arquid and also an operating partner and investing partner at Notion Capital. Today, we're going to look at how shifts in technology are really changing how enterprise works. We're going to talk AI, we're going to talk quantum, and we're going to talk about the previous shifts from cloud and how that was framed as a reset. But in fact, some things stay the same. And I'm going to try and get some of that out of Andy, because Andy has really been around and seen it all as a CEO that has taken companies public, as an investor that have helped take companies public, and as an investor that, of course, have done multiple very high profile series A's.

1:47this show is not investment advice and the hosts of this episode may be invested in the funds and companies featured andy welcome to the podcast great great to be here thank you andy i waffled

1:59Andreas Munk Holm:a bit through my intro i want to give you the opportunity to tell everyone who you are how you got to be where you are today so that they know exactly the capacity of the person we're with okay yeah sure hey so let me let me kind of give the short version so i went to university and thought i wanted to be a silicon designer got very excited about uh way back kind of designing microprocessors a little company called microsoft came along and said you shouldn't be doing hardware you should be doing software so i did a kind of very swift exit left spent a lot of time in seattle ended up wanting to do my mba so wanted to do with a large organization i had got to work with GE in the year of Jack Welch, which was quite incredible.

2:41And then the Silicon Valley came knocking on the door and said, come see us. I did, went down there, kind of blew my mind, like I think it would have done for anyone back then, and joined a small little startup called Ariba, who said, hey, we want to get going in Europe. Girls set it up. So I moved back, put the flag in the ground next to the airport, and was on a wild ride then of going public. Joining super early was super interesting. I don't know if people remember March 2001, the NASDAQ went from like 5 ,000 to about 1 ,000 in a matter of weeks. So it was like, should I go running back to big tech and big industry?

3:14But I didn't. I kind of pushed forward. So it put me on the journey of doing five companies joining super early and getting through IPOs. So Ariba, SuccessFactors, Bizarre Voice, Workday, and most recently HordeWorks. And then decided I wanted to give something back. I'm like, where's the big companies in Europe? So that's when I got into investing and advising and my journey with Notion. And then I made the classic mistake of falling in love with a company, Arkyd, who we can talk about later, I'm sure. But I'm now CEO there. And again, one of the things I kind of try to avoid being the CEO of a publicly listed company because of everything that goes around that.

3:51But here I am. I'm the CEO of a publicly listed company.

3:54Andreas Munk Holm:And before we go and talk about technology shifts and how they're affecting and what's happening right now with AI and what you believe is in store for us with quantum. And in that connection, I think we should talk about ARQ in more detail. But before we go there, I want to just double click on your experience taking companies through IPOs, because this is, of course, experience that not a lot of people have, especially not in Europe. If you could just talk through your experience doing that and your core learnings, and literally you can take it from the beginning and then all the way up to your latest company.

4:28Sure. I think one of the things you know when companies get formed is it tends to be the CEO and the founding team tend to come from the technical side. Yeah. And they've had often a really great idea, a brilliant idea, and they've tried it with a few kind of early, you know, design with customers or something. And then when they get to the point of saying, okay, we're ready to scale, we're going to raise some money. They don't really understand kind of the person or the type of person they need that's going to help them scale the business from a go-to-market perspective. And I just started to see time and time again, that complete car crash of, you know, they either delegate the responsibility to somebody else and say, just go find someone to lead our go-to-market function.

5:13Or they try to interview themselves and they come back and go, you know what, I've never interviewed anyone like this before. They sound pretty good and i always say to them well listen if a salesperson can't sell themselves and they're not gonna be able to sell anything you know so they've obviously sold you pretty well and then you meet them and and often they're a good person but you kind of say to them hey you've got a thing i think in europe and i'm talking particularly about europe now the kind of one to three the three to ten the ten to thirty and the thirty plus are all very distinct phases of growth for your organization i'm talking about revenue yeah and the person that you need may be different for each part of that stage of growth and their skills may be different and their background may be different and i think in europe we're starting to figure some of that out and we often get over excited when a company gets to 10 million of revenue and we think that's it they've cracked it they're brilliant they're amazing they're going to be like a billion dollar company but to me the absolute valley of death is the 10 to 30 journey i mean so many companies raise a b they go on that 10 to 30 journey and they just get stuck and they can't get beyond 30 and then we have this problem of growth equity or growth revenue in um in europe you know they it's it's a limited pool to raise from and before they know it they end up being either a national champion like hey we're the best at this in this country and they never get beyond it or they just end up saying, hey, we're just going to end up a part of a bigger organization and kind of accept that.

6:43And I just, to me, it's frustrating. And that's the cycle that I really wanted to try and break when I got into venture.

6:49Andreas Munk Holm:And Notion, of course, I've produced a lot of content. And I know you work very closely with Stephen Miller, but you've produced a lot of content on exactly this point of scaling through the different stages and specifically also from B and onwards. Andy, I'd love to ask you if you can talk about the zero in on the 10 to 30 million stage. How do you unlock when you get there and you then get stuck? What are the things that you see that founders really need to focus in on? And what are the recipe to then get it right? You know, when you get to 10, it's really all about new names and champions. So you want to get as many net new name customers on board as possible.

7:28Yeah. The first company that really proved this out with Salesforce, of course, you know, when you can get your net revenue retention really moving, you know, and when you get into the, I mean, the absolute top performing companies, they're kind of 150 % plus. So for every dollar in the original deal, they'll generate a dollar 50 plus in subsequent renewals and lightning the load on the need to get those new names in the door. And I think that division, it's kind of like one sales motion becomes two sales motions become, and they start to think about that. Also going international is a big part of this, because I think any company, if they're really good and have a smart product, could get to$10 million and be a national champion, the best at this in a country.

8:12And I'll kind of quote TechWolf. You know, they said it to me really clearly. They raised their money and said, we kind of wish we'd skipped some of the European expansion and just gone straight to the US. Yeah, which, whether we like it or not, is the largest software market on the planet. And it's one geography, one set of laws, one language. and if you can crack that, the scalability is just enormous.

8:35Andreas Munk Holm:Everyone is talking about how AI is rewriting the rules. I think we are a lot of people that are looking at fabled firms like Notion that have really built their name in SaaS and enterprise and thinking, huh, how are they moving now? How are they manuring this? Because it is a platform shift. A lot are talking about the rules of SaaS being broken. We're using the word SaaSpocalypse and that type of thing. I'd love to ask you if you agree on this foundational belief that everything's changed now, the rulebook is rewritten, and we've got to really readjust everything we do as SaaS investors and as SaaS founders?

9:13Andreas Munk Holm:Or is it a bit overdone by the media? And in reality, yes, we all got to be building with AI, clearly. But other than that, the game isn't that changed. So I was in New York a couple of weeks ago. It was the NASDAQ's 55th birthday. which was really interesting because the vast majority, if not all the companies listed on there, would be perceived as yesterday based upon what you just said. Yeah. And I was talking to, I won't name them because I think they're all in the same boat basically, but they were saying, you know, listen, I'll make these numbers up. I might have 100 million of revenue today, 100 million of revenue backlog and 100 million of cash, but the market value has made 100 million.

9:57Okay. It makes no sense. and when you talk to you know i spend time with our banks other banks family officers high net worth etc they've just basically discounted the out years to zero because they believe that ai is moving so quickly that they're just going to be completely disrupted and that future revenue is not guaranteed anymore yeah and then living off cash was kind of the way they think about it so you've got this brutal reality on one side whether it's right or wrong that's what they believe okay on the other side we've got these mega ipos about to arrive you know spacex rumored to be 1.5 trillion ipo there was the leaked open ai cap table you look at it which by the way wasn't amazing returns for a lot of the vcs on there when you look at it you know in in kind of multiple numbers so i think they got it they got in line the amazing thing about that capital though was 20 was the employee stock option pool which is really high so 20 of that company's employee owned which is it's going to be good for a lot of people right but back back to kind of how we we think about notion i i i always think that people overestimate the ability of enterprise companies to absorb new technology okay i've lived that world so i was a ge you know and i was running a large part at a very young age which was crazy a part of our it real estate and just moving to client server when we kind of had the kind of mini mainframe kind of back end previously took years okay so i think we're still very much in the experimental phase where organizations are trying to empower people with tools but they mostly are using enterprise software vendors to to give them that enablement today and i think if you can transition to say we're ai native and we have the unlock and we're going to make i think by the way pricing is going to change a lot you know we we've we've got to been through the perceived we've been through we're going through tokens now and i think there'll probably be another revolution of pricing is just that flexibility is going to be tough for investors as well because they're going to be saying well how do you forecast that how do you think about that when you're in this experimental experimentation world so i just look at companies now and i i look at architecturally how they're changing, how their go-to market is changing, and also the messaging to their existing customers and potential new customers about what that means in terms of their usage of their products.

12:24And often the enterprise will come back and say, that's really great. We'll put that in Q4, or we'll put that in Q1 or something like that. It's not like we're going to do that instantly because they've got a massive workload to get through already.

12:36Andreas Munk Holm:I continue to be surprised at how little excitement there is in this current era around consumer. Because as a consumer, you're super happy every time a new feature ships. And it doesn't break anything for you. The other day, my Spotify asked me if I wanted to mix my playlist. And then I just started doing super cool shifts from one song to the next one. I was like, what is this? And then I was literally AI-generated shift in between two songs. And it sounded really awesome and gave me an experience. I was like, this is completely wow. you don't want any of that happening as an enterprise customer so so i'm surprised that vcs are are more more attentive to what's happening in the consumer space and doubling down there but that's just my take but just just real cut just to cut in real quick sorry and i don't want to interrupt your train of thought but one of the big things that happened if you remember when cloud came in it was the time where people like apple were releasing the iphone and and iPod and things like that, and iPad.

13:39And people were turning up at the office with more compute power in their pocket than they had in the office. Do you remember? So everyone was bringing their own devices and they were like using their own. So bring your own device. And that's where shadow IT all started. Do you remember? Like IT were like, we're going to do an audit of what people are using. And they're like, oh my God, none of these are any of our tools. Yeah. And I think that's happening again. Yeah. You've got shadow AI happening, which is really dangerous from an IP perspective, because just look at this leap from Anthropy where the source code was downloaded, rerun in Python, and kind of made available very quickly on alternate servers.

14:12People are taking company-sensitive information and putting it straight into OpenAI and Claude without their employees having no logic because there's no guardrails because they don't know what the guardrail should be. So we're in this really messy middle right now.

14:26Andreas Munk Holm:I guess this is a perfect transition to talk a bit about Arcade. Tell everyone in the audience, what is Arcade? How did you get involved? I love the story. You shared it earlier, but do it in a bit more detail. You fell in love with Arcade as being one of the investors at Notion and said, wait a second, I know how to take a company IPO. I'll join as a CEO there. And now you're full-fledged on the team and running that. I'd love you to tell everyone about that story because I think it's a cool one. Yeah, let me – I'll tell you about the genesis of the company real quick, yeah? So if you ever get the chance to meet people who were there when the kind of web came into being, you know, they'll tell you that they never really finished it.

15:10OK, and there's a couple of things that were missing. Security was missing. Identity was missing. And the kind of one to one communication was missing. Hence, things like WhatsApp coming along and things like that. But a part of that was the PKI infrastructure that we all take for granted. Now, we're sat here chatting away and we're using PKI. it's just working away in the background and we assume it's safe and you know everything's secure quantum was super super theoretical but when people started to think about well what can we use quantum for there was a guy called peter shaw came up with the shaw algorithm and quite quickly he proved that you could break all known encryption using quantum computers yeah i was like wow that's kind of scary you know because we all just take for granted that it's just going to carry on you know being secure and what arkit did was look at well what's what's the gold standard of security today and they looked at what the military do and the military will do things like hey when you're encoding a like a fighter jet people literally have briefcases handcuffed to their wrists and they come along and plug bits of hardware and they said well if we could change that into software what would that look like and that was the genesis of the of the company they realized quite quickly they're onto something and i was just fascinated by the company i mean i was you know it's not often you get to work the software company also has like you know quantum physicists and cryptographers and things like that in it which is you know people i never worked with before so it just kind of blew my mind that there are people that just do these complex mathematical equations proving whether things are secure not secure etc before i knew it it was um so he's run a public company before and kind of the eyes fell on me and i was like yeah they said do you want to do it like not really since you i said well i think i do actually you know there's this there's this kind of young person's voice came out of like i'm quite excited about this so yeah here i am loving it it's every day even though i've been in this industry for a while you learn something new every day it's fascinating tell me a bit about the state of

17:15Andreas Munk Holm:enterprise security today and put it in the context of AI, are we in a safe place in general, do you think? Or are we really seeing a momentous shift that's making everything quite different? There's a number of different lenses to look at that. And kind of let me explain. There's one lens which is all about the kind of agentic world that we're getting into, yeah, where enterprises are worried about if I turn things over to agents, the identity and authentication of those agents is something they should be worried about. And maybe I'm worried about it now. There's kind of that side through AI.

17:54The other side is you may hear these terms like data sovereignty, digital sovereignty, et cetera. So the cloud act in the US, which allows the US government to subpoena an organization to say, we want to look at that data in the cloud without telling the The end owner of that data that has happened, I think has worried a lot of people. So you hear a lot about data sovereignty now saying, okay, I want my data to land in my sovereign territory. And if there's any extension of that, and it moves into other geos, I want to protect that as well. So it looks like my own network, no matter where it resides on the planet.

18:34So if my data does get stolen, then I'm -

18:37Andreas Munk Holm:I literally yesterday did a live announcement episode with Dominic from High Ground, who are doing almost exactly that, on-site security for ensuring that the system stay up and running. And it's exactly about sovereignty and the fact that you can't let anyone else own it than you. Listen, we see that use case all the time. And I think that's going to be a whole emerging new sector, personally, data sovereignty. And the other part that I think we are fascinated by is, so quantum computers today are not available at scale. And more importantly, they're not super reliable. Okay. So the threat today is the harvest now decryblator.

19:22So you may have data that's stolen at scale and you think, well, it's secure because it's encrypted in some sort of way. But that's the harvest now. The decryblator is once quantum becomes available. that data if it's got any longevity and time availability can be decrypted later and used

19:41Andreas Munk Holm:so i think the bad actors in the world a lot of that happening that we actually have a bit of a cyber security bomb waiting to go off because there's a lot of harvest operations now yeah so one of our one of our customers is is sparkle in italy and they're one of the largest providers of undersea cables. So these are the big infrastructure cables that connects continents, countries together. And you see this in the news all the time. Boats come along, they're dragging their anchors, they're pulling up cables. So they're trying to intercept data in motion to say, hey, if we can just suck up this data, store it away, then we can decrypt that later.

20:19And I think there's lots and lots of different levels of worry about this and lots and lots of different levels of worry about when does quantum become available at scale reliably? But I think certainly within the next, well, Google just pulled forward their date by which they think quantum will be available to at scale, decrypt information to 2029. So that's literally three years away that that becomes a reality.

20:43Andreas Munk Holm:And what does the future then look like in your view? I think we, the big call to action and the US government are probably the fastest moves on this. You're seeing it across a kind of patchwork across Europe of people saying, this is not a rip and replacement, just a greater security. This is like a kind of bottoms up moment of rethinking your entire landscape in terms of making a crypto agile to make sure that you protect it against quantum. Because you may have seen this in the press. They call it Q day. And Q day is the day where quantum can break all known encryption using Shor's algorithm.

21:19Yeah. And that is, it may have happened. we don't know because nation states may choose not to publicize it okay so it may have happened i don't know and i'm kind of 50 50 on this i've heard people actually with security background stand on stage and say it has happened i'm not they haven't got proof though but that's that's

21:35Andreas Munk Holm:going to happen in the next couple years and when you say you're did you call it crypto agile crypto agile yes and that explain that concept to everyone so we understand and then we can unpack how we get there because this is of course important for every single tech company in europe to understand so you've got you you in your estate you will have and i'm talking about enterprises because enterprises grow through acquisition mergers divestitures and it's it's impossible to have unless you're a brand new startup to have a perfect estate of it yeah it will be a mixture of everything and one of the big banks i went to recently i walked ran their data center and data centers are kind of there will be people that offend me saying this though i think they're kind of boring it's like cages with flashing lights right but i got to the end and there in this cage was like two guys kind of hammering away on terminals and i was looking and going and they said oh here's here's the guy that's looking after our fortran code and here's the guy looking after our cobalt code this is a bank a tier one bank okay so you look at that you're like wow you know so we've got tech that will never go into the cloud all the way back to like the 70s that's still running in data centers.

22:49So crypto agile is to say, if you did an inventory of your entire estate across all those different levels of security and legacy tech, what would you need to do to secure that to make sure they all move up to the same level and have the ability to be agile in defending that perimeter? The other part with data sovereignty as well is, you know, some, Germany's been very active on this, and so Saudi Arabia is saying, hey, we only want our data in our country and never to leave our borders. The other way to think about this is with data sovereignty solutions is to say, if our data did leave and it resided on another server, how can we ensure that they never have the keys to be able to decrypt that if it's actually quantum secure as well?

23:32So there's a couple of different ways of thinking about it. And I think most organizations have not even started the discovery. So there's a whole sector that gardener now talk about automated crypto discovery and inventory which is just saying what does my landscape look like from a from an ability to withstand an attack with ai you can launch attacks that will keep self-modifying and trying to find weaknesses millions of times an hour it's truly kind of frightening that that you're going to be under that sort of scrutiny at some point the other part is sometimes if if the cloud act comes in you don't even know if your data has been legally taken or illegally taken.

24:11It's really hard to tell which puts you in an even worse position.

24:15Andreas Munk Holm:You're talking about being crypto agile and getting secure against a quantum attack. Can you get to that? Is it possible? That's what we do. So we do, our way of fixing this is what we call symmetric keys. Symmetric keys means that the keys are never transmitted between the two endpoints so they can't be intercepted and also at the endpoints we can do so that's data in transit data at rest is encrypted and then data in process so we work with people like intel where we put our encryption technology in a trusted enclave on device so that it never leaves the silicon yeah so the points of interception are absolutely minimized it's fully encrypted and if it's ever intercepted and harvested then they don't have the keys anyway to to be able to decrypt it And because the keys never get transmitted, it means they can't be intercepted either.

25:08So that's a replication of the kind of gold standard cryptography I talked about from a kind of military perspective. The complication in today's world comes from the sheer amount. The vast majority of data now is machine generated. It's not human generated. You know, you look at like an ERP system. The data it generates is minuscule compared to like look at Smart City. You know, the data that comes from traffic lights and crossings and things like that. And then you layer on now autonomous vehicles. So in the civil world, you know, I was following a Waymo last night in London, by the way. And it still freaks me out when I look at it and I see no one driving.

25:45But also I keep thinking about the data that's flowing in and out of that, just that vehicle and multiply that by all the cities and all the cars in the world. It's massive amounts of data that's going on. And I think the bit that is kind of the most worrying in today's world is, you know, warfare is turned very much into drone warfare with kind of edge networks and the ability to encrypt that as well and keep those items safe when they are autonomous. They're just doing their own thing, you know. So I just look at it and think the problem of security now is thousands, if not millions fold of what it was when PKI kept along.