- 15 hours ago
Category
🗞
NewsTranscript
00:00I'm going to start on this merry question. How close are we right now as we look at Russian
00:05interference into NATO airspace with jets and drones? How close are we at this point
00:09to conflict with Russia? Thanks Tom. When we did our first funding round, I remember in 2021 when
00:17defense wasn't that popular to be honest amongst investors and no European VC wanted to touch us.
00:24We had a slide in our pitch deck that said that we thought there was going to be a galvanizing
00:28event and that was going to be an invasion of Ukraine by Russia which I guess we got right
00:34but what we got wrong was the timing because we thought it was going to happen in 2025 and as we
00:38all know in the event it happened at the beginning of 2022 and ever since then we believe it's rational
00:44to be cautious and pessimistic because the reality is that it's too hard to predict right now and
00:51how close are we right now to another conflict maybe at the NATO eastern flank? The date that
00:57we're hearing everywhere is that we all have to be ready and operational by 2029 at the latest. So
01:03when we think about systems that we're developing you know either in prototype form or going into
01:08mass production we always have that day 2029 in mind. Okay 2029. NATO countries have now committed to
01:15five percent of spending, five percent of GDP. How is that, that's going to take some time to trickle
01:21through but it's starting. How is that changing defense innovation both here in the UK and on the
01:28continent? So first of all it's good to see those increases and they're now really increases.
01:32It's particularly visible right now in Germany but also a number of the Scandinavian countries.
01:38The reason it matters is not just because of the pure increase in money. I think that matters as well.
01:42There is a very specific reason it matters for innovation and for what we call defense technology and
01:49the new technology that's now required in defense which is based on autonomous systems and on AI and
01:54software. And the reason is that existing defense budgets tend to be entirely locked up already in
01:59legacy programs. So even if as a country you wanted to reallocate that money you actually find it
02:05politically really difficult because it usually means cancelling projects and programs that a lot of
02:11jobs relying on. So it doesn't tend to happen. The thing that now is happening again with Germany
02:17probably being the best example of the additional funding becoming available is that you can start
02:23new programs and new ambitious programs without cutting old ones. And this is why I'm worried about
02:30countries that don't significantly increase the defensemen right now because what makes it very
02:36hard in that particular case is to actually introduce new technologies because it's so hard to cut
02:43legacy programs. And so that links to a question around procurement which can be a years long
02:48process. Is that still a roadblock? How much of that funding is going to come through to defense
02:54tech like Helsing and other startups versus the legacy primes? It is happening now. Again in reality we're
03:01seeing a lot of tanks still being bought and artillery systems being bought and I think that's okay up to a
03:06point. And that point is where it can't cut too much into defense spend, the new defense spend.
03:12So what we're seeing right now again in the countries I mentioned earlier is that there is now a real
03:16commitment to a spend money on new systems, new autonomous systems what we call autonomous mass
03:22but be also do so in a much faster way than previously. So you know 18 months ago I would have
03:28complained a lot about the speed of procurement. It is actually getting a lot faster now.
03:34There's a lot of focus on supply chains right now. This conference is focused on sovereignty as well
03:39and we're thinking about rare earths and just today or unless 24 hours the US and Australia
03:43inking a deal around minerals and rare earths. How vulnerable are you at Helsing and others within
03:48the defense space to a dependence on rare earths from China? How do we make more resilient supply
03:54chains here? How much of a gap is there? Yeah I think we all share a certain vulnerability
03:59obviously across different countries and across industry. Defense isn't the only industry that's
04:04dependent. It is true that rare earths are a problem but I think the problems actually start
04:08elsewhere. There's a whole number of components that you can source in limited numbers from all
04:14kinds of different suppliers and countries. What we've learned especially now that we're going to
04:18mass production for some of our systems it's one thing to build dozens or hundreds of something.
04:23It's another thing to build ten thousands or hundreds of thousands and that kind of scalable supply
04:28chain particularly when you make it if you want to make it resilient is still a challenge.
04:33It feels like there are hundreds of drone companies being started up here in Europe. I mean obviously
04:41almost on kind of on daily basis and I think according to some reports there are over a hundred
04:44different drone startups. So as we think about the capital that's going into defense tech
04:48is that rational? Are we seeing a bubble in when it comes to drone startups here? How are we thinking
04:56about whether or not that fund flow specifically into drone makers is rational? I think there's definitely a
05:01bubble right now and the reason for it is what I said earlier in 2021 not a single European VC in
05:08particular wanted to touch defense. Now everyone has moved into defense and you know including the
05:13crypto caravan that's moved from crypto now into defense and is trying to spend its money there.
05:19I think in some ways it doesn't matter it creates noise in the market obviously but there is a lot of
05:25money I think that's not going to see a return. And the reason is that I think people use their muscle
05:30memory and their experience in VC which is all around founder quality which is of course important
05:36but also product market fit which is important but is only a small part of being successful in
05:41defense. If you want to build a prime in defense you need to understand procurement, you need to have
05:45government relations teams, you need to be able to you know think about even security clearance,
05:53understand export law etc. And these are all subject matters that are quite different from
05:57product market fit. So if you just go for drones, drone manufacturers that look that they have a
06:02good program or good product that might be useful, you're probably not going to see a return because
06:08there are too many of them right now. And there won't be that many programs of records for large drone
06:12purchases either. So there's going to be consolidation when it comes to the drone startup. So is that what
06:17you foresee and you're going to be active in that space? So there'll be some consolidation. I think
06:21there'll be a lot of drone startups and in fact defense startups that are going to disappear again over
06:26the time. We have a Cambrian explosion at the moment which is fine. I think it's to be expected
06:30but probably 80% of those companies aren't going to make it. 80% are not going to make it.
06:36Talking of your acquisition strategy, if you can just outline that for us because you recently made
06:41an acquisition in Australia for underwater drones. So you've got the air drones underwater and then
06:46you've got your systems around fighter jets which will get onto autonomous fighter jets.
06:49What is the acquisition strategy at Helsing? You're looking at full stat. You've gone from software
06:55through hardware. What does Helsing look like in five years in terms of the bits that are missing
06:59now and what you hope to? Yeah I mean Helsing's focus is software and AI. That's how we started
07:04the company. That's been initially our pure focus. What we've now done over time is layer in hardware
07:10expertise where we required it. A good example is the HX2. It's our strike drone where we needed to be able
07:16to design something where we could absorb a lot of the complexity of the hardware into software
07:22and then start thinking about designing this for mass manufacturing. That didn't exist in the market
07:27and whenever that's the case it doesn't exist then we build hardware ourselves. If we find that
07:32hardware exists even in prototype form that may be useful for us we either partner with companies
07:37or in some cases we acquire them as well. Blue Ocean is an example of a company that we acquired.
07:42We've been developing together with them an underwater glider called SG-1 Fathom which can stay
07:48underwater for several weeks and uses what we call a large acoustic model. So a big neural network
07:53to scan you know the underwater acoustic signals for nefarious objects. I touched on on air warfare
08:02there because you've done some test flights recently with your Centaur system in terms of unmanned or at
08:08least there was a pilot but the controls taken over by your software system with the Saab grip and E.
08:12How much progress has been made? When do we start that seeing that operational?
08:17Yes so Centaur is a system it's a reinforcement learning based agent a little bit similar to what
08:22you would have seen in AlphaGo and AlphaStar with DeepMind that we've trained through self-play
08:29against each other to essentially become as good as a human fighter pilot in what's called BVR
08:34beyond visual range air combat and that system achieved human performance last year and it's since
08:41surpassed it and it's one of those areas where we're incredibly proud that the team around the
08:46board which we call Area 9 has been able to build a system in record time and actually show that in
08:53Europe we can build extremely advanced AI systems that can compete worldwide and that actually have
08:58implications for how we should think about crude air combat which we think over time is going to
09:02become more and more dangerous because you'll have you know drones coming up against you but also
09:07large programs that are currently being set off like for example FCAS which is still a crude aircraft
09:13program where probably over time that is not the kind of system that you want to have. Well it's
09:17interesting you touched on FCAS so this is this program being developed between France and Germany
09:21for a new fighter jet but they're at loggerheads essentially and there's corporate reasons for that
09:25maybe some political reasons for that as well what's happening there if Germany and France can't get
09:30together and build a fighter jet in this moment what does that tell us? Well I think there's still a
09:35political will to work together obviously and you know Europe needs to work together to build these
09:39systems and we actually have a rich heritage of being able to build systems together the Eurofighter
09:44Typhoon is one example the Tornado is another and there are many others as well in the particular case
09:49of FCAS obviously the companies involved at the moment are not able to see eye to eye that's also
09:54happened previously and needed a political intervention at the moment that hasn't happened yet so at the
09:58moment it looks like the what is your vision of what air warfare looks like in five ten years time? I think
10:03it's pretty clear that it's going to be largely uncrewed and the reason is that the AI just makes
10:08faster decisions it's able to find new strategies in ways that humans can't and I think in a number
10:13of years time it will be ethically problematic to send humans into the sky against these these
10:18and you touch on ethics there because we're talking about unmanned systems piloted by AI agents
10:24how deeply do you think about ethics about the challenges of human in the loop about who gets
10:30hold of your software systems and your and your hardware? Yeah we think that as a company that's
10:36that's steeped in democratic societies we have a specific responsibility to think about the ethical
10:41side you mentioned human in the loop our view is that the human should always be in the loop for
10:45systems and once you have autonomous mass where you have so many systems that you can't monitor each
10:50one of them individually you should at least have a human on the loop and then you have to make sure
10:55that human is actually an effective human in the loop or on the loop so they need to be able to not
10:59be overwhelmed by the data not be bored and there are all kinds of KPIs that we need to track in terms
11:06of whether a human actually does their job that are currently being established so this is one thing
11:10the other thing that I think we have a special responsibility for is to think about who we make
11:14those technologies available to what about proliferation making sure that those systems are actually
11:19safe and the way we approach it is that we have workshops that include our entire team the company is now
11:25about 900 people and many hundreds hundreds of them actually take part in these workshops regularly
11:30to discuss these topics to come up with guidelines and then to make recommendations to the management
11:35of the company because in particular in Europe you need to have this broad church of people of all kinds
11:41of political backgrounds that can contribute to this kind of ethical muscle as we call it and build
11:46that ethical muscle so that we as companies can make the right decisions do their countries China you're not
11:52selling to obviously Russia you're not selling to are there European countries that you wouldn't sell
11:56to well I know the way we look at this is that they are clear no's and North Korea is another one out
12:01of the list that you mentioned and then they're clear yes as well and then there are some countries
12:06that are you're kind of in the middle where there are edge cases and for those we actually again call
12:11together you know workshops we have an ethics council as well that will make recommendations and then we
12:16decide on a case-by-case basis because quite often it depends on the circumstances and also the
12:20capability you make available okay I want to look back to the supply chain question because you're
12:24in operation in Ukraine have you have you faced as Helsing faced direct attempts by Russia sabotage
12:32across any of your supply chains any of your units and operations yeah one of our component suppliers
12:37was bombed you know several weeks into deployment and that was obviously horrible and tragedy for them
12:44it also meant that our supply chain was disrupted we had to find that particular component
12:47again on the market and this is a this is a regular occurrence it's obviously specifically
12:52problematic in Ukraine okay that's a regular occurrence and when we think about the Russia
12:56risk we think about the CEO of Ryan Mattel as well and his own personal safety is are you still
13:00working with Ryan Mattel what is your relationship with the prime do you see them as competitors or
13:04partners well we see them as potential partners and friends I'm not sure whether the feeling is always
13:09mutual we work very closely with Saab and we also have partnerships across the whole ecosystem but the
13:17reality is that I think some of the primes right now see us as a threat I think initially we were
13:21seen as a maybe sometimes convenient kind of add-on to what their capabilities now I think they see us more
13:28as a threat and but I think that over time will change as well because there is a role for the primes
13:34obviously and you know if you look at Ryan Mattel Ryan Mattel is doing an amazing job you know creating and
13:40providing many of the systems that are required across Europe BAE is another company and I think what we're going to see
13:46over the next few years is that we're going to coexist and we're going to partner again as well
13:49but at the moment everyone is trying to figure out you know how we all relate to each other and you've
13:54before we before we wrap up you you spend a lot of time here in the UK and you you cooperate with
13:59the Department of Defense the UK obviously has a very close security relationship with the US when
14:03we think about sovereignty is that potentially a vulnerability for the UK I I think we still need
14:09a close relationship with the US I think that's pretty clear what we can do and what we should do across
14:13all of Europe by the way not just the UK is to build our own systems and be able to stand on our
14:19two feet so that we reduce some of the dependency obviously but that also we can meaningfully
14:24contribute to the security of Europe and it's pretty clear that we haven't been able to do so over the
14:29past few years it's changing now very rapidly and my prediction by the way is that in the next few years
14:33Europe is actually going to lead even over the US in terms of the deployment of real AI enabled
14:38autonomous systems and so not Europe is going to lead the US in the department I think so I think
14:43it's going to look that way in terms of you know the visibility we have right now in terms of deployment
14:47that's planned of in terms of the volume of these autonomous systems particularly for the NATO eastern
14:53flank I think Europe is going to lead okay
Recommended
19:43
|
Up next
1:58
1:57
1:48
5:25
6:23
7:01
15:02
8:04
5:34
11:44
5:24
1:46
12:24
4:37
2:42
5:19
18:27
6:35
2:17
19:31
6:50
6:03
4:35
Be the first to comment