- 16 hours ago
"Europe has the fundamentals: world-class talent, a large market, and growing regulatory influence. Awareness has shifted decisively: from Brussels to boardrooms, sovereignty is now understood as a strategic imperative rather than a political slogan.
Yet a gap remains between ambition and execution. Investment levels still lag behind global competitors, market fragmentation continues to slow scale, and coordination between public and private actors is not yet optimized.
This session brings together industry leaders to focus on what comes next: how to operationalize sovereignty as a driver of competitiveness. From smarter capital allocation and industrial policy alignment to cross-border collaboration and infrastructure buildout, the discussion will explore how Europe can translate intent into execution and turn sovereignty into a tangible business advantage."
Yet a gap remains between ambition and execution. Investment levels still lag behind global competitors, market fragmentation continues to slow scale, and coordination between public and private actors is not yet optimized.
This session brings together industry leaders to focus on what comes next: how to operationalize sovereignty as a driver of competitiveness. From smarter capital allocation and industrial policy alignment to cross-border collaboration and infrastructure buildout, the discussion will explore how Europe can translate intent into execution and turn sovereignty into a tangible business advantage."
Category
🤖
TechTranscript
00:04all right good afternoon everyone welcome to this session my name is Ruben schaubroek senior partner
00:11McKinsey and we're going to talk today about sovereign AI before I let my panelists introduce
00:18themselves maybe just to frame a little bit the conversation we're having right sovereign AI so
00:24far had been a conversation mainly in Brussels policymakers and so on but that has radically
00:28changed over the last year now it's on the agenda of every boardroom and just to give you a bit
00:34of a
00:35sense of the some of the numbers and the magnitude of this change if you basically look at the total
00:41AI global AI spent we actually think that 30 to 40 percent of that is subject to sovereignty
00:47requirement so that's a massive amount it also means it's a massive market that's why we think
00:52the sovereign market is actually 500 to 600 billion of USD right same time we also see benefits we did
01:00some calculations at McKinsey we think there could actually be a benefit of over 400 billion euros
01:04just for Europe alone now what's the issue the issue is execution and that's what I want to talk
01:09with my panelists about execution on two fronts demand why do we need sovereignty exactly right
01:15what are the benefits what is the real ROI and some issues on the supply side right how do we
01:21bring
01:21sovereign solutions in Europe are still quite fragmented how do we get the right level of
01:25investment so to help answer those questions I'm joined by an amazing panel thank you for for doing
01:31that of course we have Catherine just now who's at Airbus leading digital driving a lot of AI
01:37transformations including in manufacturing we have Tanuja who's at AWS leading EMEA and we have Michael
01:45Beck at Ericsson corporate officer at Ericsson and so I think in this combination we have the perfect mix of
01:51technology strategy and policy expertise to actually have that debate I'll just ask first the panelists
01:57to just give a bit of an upfront reflection right as you reflect on sovereignty sovereign AI what is
02:04the one thing you think this audience should take away or was the one thing which is on your mind
02:08I mean
02:08Catherine probably the definitions that we have in others for digital sovereignty is twofold on one side
02:16it's about keeping an absolute control on the access to our most sensitive information be it military or
02:26intellectual property or trade secrets and the second aspect is to secure the business continuity so
02:35basically nobody can have a key switch that will prevent others to access to that information so that's basically if
02:44if we have to remind one thing I would like you to remind this definition and maybe the second
02:50thing is that one size does not fit all means that not everything even at Airbus needs to be
02:59completely sovereign so we can and it's all a question of balance between when do you need sovereignty when you
03:07don't and to select what is the most appropriate technology to deal with that business
03:12the new job what do you think and I think Catherine stole my thunder no but I gosh could not
03:21agree more
03:22the the very 30 second answer that I would say to you is the definition of sovereignty is very ambiguous
03:30and it really
03:31does depend on whether which industry you're in what workload the use case the application I mean it
03:37really is not one size fits all actually to use Catherine's words when I talk to our customers across
03:42EMEA what they say to me is they want scalability of platform security performance cost and I think all that
03:50translates into choice so my my our point of view is that sovereignty should be an and and not an
03:58or actually
03:59and you can in fact meet the requirements that you have for those very stringent workloads and the rest of
04:05your
04:06workloads without having to give up or compromise on the quality and performance of your services so we
04:13really believe that you can have control and compromise I'm sorry control without compromise and that's the true
04:21nature I think of delivering sovereign solutions thank you Michael and I think for it for us very much
04:28on the digital side I think it's very clear that every government needs to have a plan for their whole
04:34digital stack that isn't always been the case every government has a plan for for nuclear for roads
04:41and so on but now the digital infrastructure is going to be as critical in some cases more we also
04:48see that
04:48much of this infrastructure is built by private actors and so on so I think we need to ensure a
04:54system where we
04:54have not only sovereign systems but also ensure enough investments to have meaningful systems built and so
05:01on then I think most countries with few exemptions are too small to completely isolate from the rest of the
05:08world so we also have to judge vendors supply change who we can actually trust I think it's it's going
05:15to be
05:15very difficult to live on second best products because you're completely leaving out parts of the world so I
05:21think there is a number of new tasks where this cannot only be handled by regulators it has to be
05:27handled on the
05:27highest level the governments I love the different perspectives and maybe bringing it back to what you hear from
05:33your customers and in your own management team discussions why is there so much focus now what what do the
05:41customers really want to get out of it right and how do you see this evolve if you want to
05:45bring the customer
05:46perspective into the room I'm not sure who wants to start maybe Catherine maybe I can I can start I
05:51think that what is changing in the aerospace industry is that we are really shifting with towards a data
06:01driven environment so the data the software are taking more and more importance and in all the products be it
06:11to be it's helicopter satellite or or aircraft because it is the digital is becoming so important it becomes
06:21a competitive advantage and because it is a competitive advantage you need the sovereignty you need to make
06:26sure that you get the control that nobody can have access to your IP that nobody can I have access
06:32to your
06:33most sensitive data so basically the sovereignty is driven by the fact that the industry movie is moving and
06:40shifting to something that is much more data driven and much more software defined
06:45I think it's super interesting a competitive advantage or it's no longer just a must from a regulatory or policy
06:50perspective but it actually is a market force right you you
06:54I want to lean on the data point as well because um so as we've been you know deploying cloud
07:01solutions for like 20 years
07:02data has always been something that we keep a very firm eye on and you know many of you will
07:08have heard Jeff this morning
07:09talking about security being the most important thing about what we do but AI and generative AI has won the
07:16hearts
07:16and minds of c-suites like no other technology we've seen and so suddenly all the boards of all these
07:23companies
07:24and all the CEOs are suddenly realizing that this technology is truly either magically transformative or highly disruptive
07:32and so they're leaning into this at scale we see it by the way in the report we've published on
07:37unlocking
07:38Europe's AI potential which shows 54% of businesses are using this technology today
07:43but unfortunately only one in five businesses use it in a way where they're really monetizing the value of their
07:52data
07:52so there's two data is the critical asset with AI it's even more important no data no AI really you're
08:01not going to get it
08:02and then I think the second thing is of course AI also creates and has created anxiety around the responsibility
08:09that we all need to take about our own data as well and of course the talent and the next
08:14generation
08:15so I think it's that combination of factors that leads to this why now and why the atmosphere around it
08:22is so much more
08:24deeply sort of in some ways separated but in some ways together as well
08:28and I think one thing we see from our customers is that all data was not born equal some of
08:36that has to be accessed on microsecond level
08:38some can be stored in a data center very centrally so I think a big part we're discussing also with
08:45our customers is where different types of data resides
08:49and what kind of compute do you need in your device very close to a person or a robot or
08:54glasses or what needs to be on the edge
08:57because it has to be somewhere in between and what can really be done in the data centers where you
09:02perform the training and so on
09:03so for us it's also very much around how this now becomes inference basically where the AI models are executing
09:11because we are now planning for a phase where AI is no longer only in the office and in IT
09:16systems
09:17but really in what we call physical AI so robots, cars, connected glasses
09:22they have really strong dependence on the infrastructure around them and the networks to be able to perform in a
09:28safe, secure way as they should
09:29so that comes up for us as a very key component compared to what has been said before
09:35if I can double click a bit on that, I mean sovereignty is such a broad word, right?
09:39we can talk about data, we can talk about cloud, we can talk about different software, different parts
09:45so again, talking to your customers also internally, what are you most talking about?
09:50is it the data layer or is it, so, and how should companies think about this frankly, right?
09:55maybe Tanuja, I'll start with you on that one
09:57I think, you know, in the end of course things come back to data and infrastructure
10:02but actually what I hear most about is how do we think about the processes in the company differently
10:11how do we completely not just apply a bit of automation here, there and everywhere
10:16which by the way we've been doing for many years, right?
10:19how do we drive those productivity savings because we change the way we work
10:23you know, and in the future we change the way we work between agents, becoming teammates, working with humans
10:30but even if you park that for one minute, I think the ability for insurance companies for example
10:35to completely rethink claims processing, or for a FSI, a financial services business to say
10:41do I really know my customer, or for an airline to really think about customer experience in a very different
10:48way
10:48as we all fly many airlines, I think that is the conversation that's actually happening
10:53and the companies that are investing deeply in rethinking how they work
10:57are the ones getting the most benefit versus a little bit of productivity, lots of pilots
11:03and then there's a question about return on investment
11:06so I think that's more of the conversation versus tech
11:10that resonates, I mean there's a lot of talk about rewiring processes
11:14rewiring processes, to use a good McKinsey word, that is true
11:18but if you take it throughout the stack, right, and you think so indeed the end goal is
11:22we want to compete better, I love that point
11:25we need to rewire what we do as a company
11:27but does every part of the stack need to be solved
11:29and you said something about it in your intro, Catherine, right
11:31you feel like it's a mix, right, it's not every single part
11:35how would you think about that?
11:36I think that, again, not everything needs to be sovereign
11:42but when it needs, you have to do it through the entire stack
11:46because if one of the elements in your stack is not
11:49basically you lose the sovereignty criteria
11:54and that's a deliberate journey that we have started in Abus
12:01so we have recently announced on the AI stack a partnership with Mistral
12:10that's one part of the stack
12:11on data, we are working with Chapvision
12:14on cloud, we are currently having a call for tender
12:20and we will announce in two weeks from now
12:22the partner we have selected for sovereign cloud
12:25then with our colleagues from Ericsson
12:29we have a partnership on the 5G
12:31you have maybe also seen some acquisition recently
12:35in the cyber security defense space
12:39so in the UK, in France, in Germany
12:42so we are building the entire stack progressively
12:46but again, for where it matters
12:50and where it matters is military, IP
12:53and what needs to be certified
12:56because of course when you start to talk about
12:59embedded AI in our product, edge AI
13:03in the planes or in the helicopters
13:06all of this will have to be certified by our authorities
13:10and needs to be explainable
13:12so you cannot just develop something in AI
13:15put it in the plane and make it fly
13:17you will have to go through a very, very sorrow
13:22and difficult process of certification
13:25certainly, maybe Michael to jump to you
13:27how do you think about it at Ericsson?
13:29I think for one thing that I think is important
13:32is that sovereignty doesn't come in isolation
13:34it really comes with redundancy, with safety
13:38a number of other things that we have been building over the years
13:41in the system that was there also before
13:43that is not becoming less important
13:44so if we are deploying a 5G network for Airbus
13:48of course that has to be completely safe, secure
13:50even before some of the latest requirements
13:52but I think also one thing that is very clear
13:55is that when you talk about authorities
13:57I think they will have to retrain themselves a bit as well
14:00because they are used to very, very much different processes
14:04very much longer lead times
14:06we look at the most extreme cases
14:09maybe deploying civilian technologies
14:11in defence and public safety
14:13we are used to ten year cycles
14:15where we have no product that is living longer than three years
14:19and so on
14:19so I think there is a big burden
14:21that sometimes companies are better equipped
14:24maybe than typically regulators and so on
14:27so I think this will be a dramatic change over the whole system
14:30then also supply chains and so on
14:32is what we get a lot of questions about all the time
14:35where does actually the boxes down to the component
14:39even raw materials come from
14:41so there is a lot of declaration things coming up as well
14:43that we have been handling before
14:45but it is coming up on a ten times bigger scale
14:48Ruben, if I could just add on that
14:50because I think the point about the whole stack
14:52your point about you know
14:54you are going to have to use an ecosystem of providers
14:57to be able to deliver something like a full stack
15:00so I take an example
15:02you know we have customers here in Europe
15:07that have some very stringent requirements
15:09Catherine mentioned a couple of things
15:11you know you have got customers
15:12and for these additional requirements
15:16that have extra compliance needs
15:19we have to really think long and hard about
15:22what could we do differently
15:23to be able to address that
15:25while not sacrificing any of the innovation that you also need
15:30and so we kind of looked around corners
15:33and we launched this year our European sovereign cloud
15:36and the reason we did that
15:38is not because the cloud is not sovereign
15:39the cloud is sovereign by design actually
15:42and many of you who use it know that
15:44and there are technical safeguards
15:46that ensure that the data back to data
15:50is only always controlled by you
15:53and you have the keys to your data
15:55whether you move it or not
15:56but there were additional compliance
15:59and regulatory guidelines that meant
16:02that we had to go beyond our commercial clouds
16:05to deliver this
16:06so this service then becomes disconnected
16:09physically and logically from any other region that we have
16:13it's just ensuring that we can meet the needs of customers
16:17wherever they need these services
16:20and it doesn't compromise on the fidelity of the cloud
16:23because otherwise you end up with 10 services
16:26and that's not going to do anything for you
16:29so you really do have to follow it through
16:30but I will say European partners
16:34whether it's Mistral or SAP or Capgemini or Ericsson
16:39and I think the fact that we have national champions here
16:42part of the supply chain
16:44is really really important
16:45and addresses some of that sovereignty conversation as well
16:48so if you go back to the execution question
16:51right I think we've covered a couple of points already
16:53I think Michael your point in terms of policy
16:56and it can't be a 10 year cycle to do anything
16:59your point Catherine on ecosystems
17:02a question which often comes up is cost
17:05right
17:07often we see that sovereign solutions
17:09still more expensive than obviously mainstream solutions
17:12some numbers some research say 10 to 30% more expensive
17:16how do you think about that
17:17both for yourselves and again for your customers right
17:20is that can you make a business case for that
17:22is that worth it
17:23do we collectively need to figure out a way to bring the cost down
17:25how would we think about that
17:29Catherine
17:29can start
17:31yes there is a business case
17:34just to confirm
17:35and so typically the fact that we can ensure
17:41sovereignty will ensure the business continuity
17:46and stopping the activities of Airbus for one day
17:50as a cost
17:51for two days even bigger
17:54and the more days you pile on the brick
17:58the higher the cost becomes
18:01so yes
18:02the fact that you can secure the business continuity of the company
18:08as a business case
18:11so that's
18:12and same thing
18:12we are in the process to launch the successor to the E320 in the coming years
18:20all what we do in terms of research and development of what will be the innovations that we will bring
18:28in that new aircraft
18:30same thing that intellectual property as a value
18:34so you can calculate a business case but again one size does not fit all
18:41yes
18:41so it's always a question of balancing
18:45the performance
18:46balancing
18:47the cost
18:48balancing
18:49the
18:49the sovereignty
18:51and it's
18:53it's a decision you make for each and every
18:55system in your
18:57in your IT
18:58and
18:59and AI systems
19:01data systems
19:01so
19:02and
19:02you make the decision based on the data that you will have to handle in that system
19:07makes sense
19:08makes sense
19:09Michael any reflections on your side
19:10no but I think for us
19:12we have never had the luxury to sell more expensive solutions because we are European or whatever
19:17we have always competed with on the global stage
19:20so
19:20so I think that element is not there so much in our business
19:23then of course
19:24some solutions when it comes to
19:27redundancy and so on
19:28will
19:29have an extra cost
19:30we have a lot of things like
19:32adding more battery capacity on sites and whatever
19:35where I think we also come into models where
19:38if you compare to other critical infrastructure our networks are built by
19:43private actors basically
19:45is it the regulated business
19:47but when it comes to
19:48securing coverage in areas that is needed by the military or more battery
19:53that is typically not the business case always for the operator
19:56so I think we also need to have a renewed collaboration with the governments on
20:02actually supporting areas that is not a doable business case for the operators
20:07and may not be for a long time and so on
20:09so I think that's
20:10coming back again to
20:11you need to have a plan for anything from
20:14sea cables to satellites to the digital IDs and so on
20:18and work with that for a long time
20:21what becomes costly if you do mistakes and you have to replace things and so on later
20:25that is not so great
20:26so I think that's one of the things we are really pushing now
20:30and I think it's
20:31we get more and more acceptance from
20:33from governments and regulators and so on
20:35that this is as important as
20:38nuclear or whatever you're planning for
20:40one thing that I would caution on the cost side
20:43is I think we all want to try and keep costs down constantly
20:47like I think
20:49you know you need a solution that meets the ROI
20:51one of the things with sovereignty we must be careful about is
20:55if we don't do this
20:57it depends on
20:59rather than one size fits everything
21:02we end up potentially with 27 national stacks
21:06and let me tell you
21:07that is a very expensive solution
21:10it's not even just the funding actually
21:13it's the global consistency gets lost
21:16so I'll give you an example very recently in the Middle East unfortunately because of the conflict
21:21we've had you know our data centers impacted
21:25but what we could do is migrate customers within days out into other regions
21:33why? because we have a globally consistent model within architecture they trust
21:39and I think that's something we should watch in Europe
21:42is make sure
21:43that we always can drive consistent
21:46scalable
21:47standard
21:48and yet
21:49meet the needs of those workloads very specifically
21:52you know
21:53I see Katrin you're agreeing
21:55absolutely
21:57and it's true for each and every business I guess
22:01Ericsson is facing it
22:03Mistral is facing it
22:04every time that they have to deploy and they want to deploy somewhere
22:09basically they have a new set of requirements
22:12that is coming to us
22:13and I was attending a round table with NIV Kunen earlier this year
22:20where there were start-ups and small and medium enterprise
22:25that were explaining that they just don't want to go to export
22:30because it is too complicated to fit with all the standards
22:35so there was a Sweden cybersecurity company
22:39they had a super product but they just keep it in Sweden because it's too complicated for them
22:47to go to each and every accreditation body in Europe
22:51to get their product validated on other markets
22:54so and this is crazy
22:55this is really not helping Europe to be competitive
23:00so I just agree 100% with Stanusha
23:04we need something that is much simpler
23:06and that is completely standard across our 27 countries
23:11otherwise it will be super difficult for the champions we have to scale
23:16and to at some point be in a position where they can be competitive
23:21with other companies from China, from the US, from India and from everywhere else
23:28very clear message
23:30Mike, let me actually go back to you
23:31because Ericsson obviously is doing a lot of the connectivity backbone
23:35so interoperability is something you think about a lot
23:40so how do you think about this as providing this backbone
23:43maybe including the edge and so on
23:46versus the threat of the 27 national stacks
23:50how do you think about it?
23:51No, so telecom is clearly going from a business that has been separate
23:57and now getting into a business that is more or less
23:59I would claim it's the same business now as we talk about AI and cloud for example
24:05it's really workloads moving around and so on
24:07and we come from a very standardized world as you say
24:10we meet the AI business, the cloud business
24:13that is coming from different ways of building things
24:16and more global players and so on
24:19so I think we are unfortunately used a little bit to this national separation
24:24given that you had licenses that were separating countries
24:27but now when you talk about national security
24:29you can hear even in the name it's national
24:31while digital all builds on scale of course
24:34and EU single market if you can really combine that
24:39that is a very attractive market in many ways
24:41but when you separate not even France and Germany are big enough to really fund investment
24:46so I think actually even with this push for national security
24:51I think from the pure scale business
24:53and being able to invest in the way you do in US, China, even India and so on
24:58we need to be able to trust each other to harmonize certain things
25:03then we should not take away the member state
25:05but some of these parts I think we also have to recognize that
25:08without the scale it's going to be difficult actually to be Sweden or even smaller countries
25:14I see a lot of nodding and agreement here
25:16maybe to go back to the ecosystem part
25:20the reality is that ecosystems happen a lot in this space
25:23you've talked about Mistral but also Ericsson and many others
25:26what are some of the learnings?
25:28how do you actually make that work?
25:29because it's often questioned in many parties
25:31it's complex to manage
25:33how do you make ecosystems work?
25:35anything to share with your audience?
25:38so again, if we take the example of AI
25:42we basically have two different partners
25:45one with Google, another one with Mistral
25:51and basically we have defined our strategy
25:55so we have access
25:58and we share basically the work between Mistral and Google
26:05with Google more on what we call the everyday AI
26:08either for our customers or for ourselves
26:12that is, and Gemini is accessible to all employees in the company
26:17so we had last month, I think it was 96,000 users
26:22unique users of Mistral, of Gemini
26:26where we are using Mistral for more specific use cases
26:32where sovereignty matters again
26:34and for this one, it's a more restricted number of users
26:40mainly engineering colleagues
26:44and the digital team
26:47so we have a strategy
26:49and basically based on the use case
26:52the sensitivity of the data
26:55that we are either using or producing with this AI solution
27:01we decide if it is going on the left to Google
27:04or on the right to Mistral
27:06so it's quite simple
27:08and we have usually
27:10we don't put all our eggs in only one basket
27:14so we try to avoid that
27:17so same thing on the cloud
27:20we have private clouds on premise
27:23we have public clouds
27:26in particular with AWS
27:29and Google again
27:30and we also have tomorrow's sovereign cloud
27:36and we will rotate the workload based again on the sensitivity of the data
27:42it's not that complicated
27:44it's, if you have clear ideas of what are the criteria to go on the left or on the right
27:50basically it's quite easy to manage on a daily basis
27:54I think it's a great illustration of how you actually manage these different parties
27:58and do that effectively
27:59and Ruben, to build on the prioritization that Catherine's doing in the organization based on use case etc
28:08the thing that we've been thinking a lot about is
28:11how do we provide the scalable platforms that allow our customers to be able to leverage any or all
28:19of whether it's models or specific ecosystem players or it's the infrastructure
28:25and the reason I bring that up is because I think increasingly with AI
28:29as you're deploying agents yourself across your organizations
28:34you want to be sure that those agents have the context about you as an organization
28:39by the way both cultural context and compliance and regulatory context
28:43you want to make sure that the humans in your organization are working in the same way as your agents
28:48are working
28:49they should not be working differently
28:50maybe they work through the night but that's a different thing
28:53right
28:53so I think for us it's been what are the platforms that give you the choice
28:59and so if I look at an Amazon bedrock it gives you Mistral
29:02by the way they're a big, one of the big companies we work with on the model side
29:07or Anthropic or OpenAI or DeepSeq or Minimax whatever
29:12and then on top of that you as an organization pick and choose based on what Catherine said about the
29:19workload
29:20and in fact these systems have become so smart they will even tell you based on your use case
29:27which model you should use so that you can start to reduce your cost base
29:32a lot of people say that token costs are going to be the size of their compute costs
29:35so I think for us it's been platform deployment, choice of models, choice of ability to deploy agents
29:43and then ensuring that you have your data security 100% sitting always in your environment
29:52so whatever inferencing you're doing you don't give your data to the model
29:55and that's how we think about the kind of stack
30:01and I think coming back to this where data is executing
30:04I think for us we also have a strategy for what hardware and silicon you use for different parts
30:11because we have, we talk a lot about these big data centers where you do the training and so on
30:17what we see also is that when now these models are executing in the device that could be your mobile
30:24phone, your smart glasses
30:26you're super constrained on energy consumption, batteries and so on
30:30so you have to get compute off to some kind of edge nodes in between
30:35we have our base stations that is also not as power sensitive but very sensitive
30:40so that we actually develop our own silicon to be able to handle that extreme uplink capacity
30:45we didn't do that for sovereignty but that means we have quite good control also over hardware in our systems
30:51and so on
30:52so I think there is a number of things we have to look at from the pure starting point
30:56where there's been an enormous focus on money to build data centers which is important
31:03but also to how this intelligence should now be used and executed under which regulatory regime
31:09I think we engage a lot in Brussels with this AI act that didn't come out in the best possible
31:16way
31:16I think it's easy to forget that still industrial competitiveness is what gives us the freedom to be secure and
31:25safe
31:25so we need to make sure that we have competitiveness ahead of just regulation and so on
31:33so I think there's a couple of cases where we actually see Europe could be a bit smarter and faster
31:37in the way they support not only us
31:39but the whole industry to make Europe stronger
31:43and if you then think about the value creation right so if you think about what I said in the
31:47beginning
31:48we think there is a 450 billion euro plus potential just for Europe
31:52but there are bottlenecks right and if you would have to pick what the bottlenecks the main bottle
31:57it could be energy it could be talent it could be procurement cycles it could be different things right
32:02it could be policy right what do you feel is the biggest bottleneck and then potentially also what should we
32:08do about it
32:09maybe each of you Tanuja seems to
32:12I would say there's three and I feel this panel might align on this as well but let's just see
32:19I think the one is what Michael just brought up right
32:23and it's a simplified unified I know it sounds like motherhood and apple pie
32:27but a simplified unified regulatory environment that's innovation friendly
32:31today people are spending 40 cents for every euro that they could put in technology into compliance
32:38imagine if you could reinvest that so I think that's one not easy but I think something we need to
32:43strive towards
32:44the second big one is around funding for startups
32:47you said that earlier I saw a panel earlier for startups wow
32:51and they were just like literally their ask is we need you to fund us so we can scale globally
32:58we don't want to be a Sweden only or France only startup
33:02and we just don't have enough public and private funding going into startups
33:07by the way 4 out of 10 startups are considering relocating outside Europe
33:11I'd be very worried about that
33:13and then the third one is scale scale scales because we just don't have enough digital skills
33:19and we need to invest heavily to democratize this technology for people
33:23I worry that people get left behind if they aren't able to use this technology at scale
33:28so those are the three that we worry a lot about but we're very committed as well to figuring it
33:33out together
33:34Catherine will be your list
33:35I will take a slightly different angle so if I look at what is preventing Airbus to go beyond
33:43to go faster, to go bigger, the first limitation is the data
33:49and the availability of the data as well as its quality
33:54quality of the data was not parliament in the past
34:00so we have tons of data at Airbus
34:04but being able to quantify that data and to understand the quality of the data
34:10and if we can build anything on it is another story
34:14so I think that this one is important
34:17I think you mentioned it Michael
34:19compute will be a limitation
34:21maybe not for Amazon but for companies like Airbus
34:27we already see limitation in access to servers
34:32we have seen the price going to the roof
34:36there is limitation in access to the GPUs and so on
34:40so I think that the access and the securization of the level of compute we need
34:47will also be a challenge
34:49and that's why we are also moving to the cloud
34:52because we believe public cloud and sovereign cloud will help us
34:57to bring and to breach that limit
35:01the third one is the talent
35:06but I would not say the digital talent
35:08it's also the business talent
35:10because you mentioned it earlier
35:12what will make the difference is if we are able and capable to transform our processes
35:18I mean if we are just automating what we do today without transformation
35:22it will not work
35:24but to be able to transform it
35:26you don't need just digital talent
35:29and the way for example we work with Mistral is
35:32we have on each of the use case
35:36we have three engineers from Mistral
35:38we have three engineers from my team
35:40and we have three engineers or business representatives
35:43and the most critical ones are these guys
35:46because I can replace my digital guys
35:49they will not like it but I can replace
35:52probably Arthur can also replace the Mistral engineers
35:57but I cannot replace the people who have the business knowledge at Airbus
36:02so for me the three limitations
36:04data, access to compute and talents in the business
36:08available to leverage the new technologies
36:11I love both perspectives
36:13actually complimentary perspectives
36:14Varika what do you think?
36:16I think I agree with all
36:17I think some compliments would be one common definition for sovereignty
36:21I think that would help a lot
36:23because if it's in some countries means data centers on Swedish soil
36:28in some other countries it means nationalities of people working
36:31it will be very difficult and we will lose the scale
36:34so I think that would help quite a lot
36:36the other one I think for
36:38if you compare Europe to many other continents
36:40it's the investment climate when it comes to not really private equity and so on
36:45but generally availability of risk willing capital is one of the biggest weaknesses
36:50we see in startups to scale ups and so on
36:54the reason why you go to US
36:56the reason why it's difficult
36:58and that is not an easy problem
37:00and we try to compensate it sometimes with public money and so on
37:03but if you look at investments like Starlink and whatever
37:06I'm sure that didn't look that great from the beginning as it does now
37:10so we could do these things in Europe as well
37:13but then someone needs to be able to on their own
37:16or supported by government money or whatever
37:19be able to take risk much earlier
37:21because if we always wait until all the business cases are ready
37:24one thing is clear we will be late
37:27so that to me is one thing to add on what we've seen here
37:31I think we have really strong companies in Europe
37:35one thing we are pushing a lot for is
37:37let's take Airbus and Ericsson as one example
37:40but we have world leading players when it comes to innovation
37:44when it comes to the digital systems
37:46we have world leading industries using these technologies using the AI models
37:50I think that's something fairly unique that we could build much more on
37:53when it comes to creating something new instead of copying what is already done somewhere else
37:59I think a great list of perspectives and things to do
38:02I'm going to ask you one closing question
38:04if there's one thing that Europe could do in the next 12 months to accelerate the adoption of software and
38:12AI
38:12what would it be?
38:14what would be the one thing?
38:14let's limit it to one sentence if you can
38:17I think investment friendly, innovation friendly in everything we do
38:23I think that's the key thing here
38:25we will not be able to afford all these good things unless we can keep on being in the forefront
38:30looking out the best solutions and so on
38:33will in some cases make us, I think, less competitive
38:35great, Tanuja?
38:40I'm going to say, you know, think local and global at all times
38:45because you need to have partners who are committed, willing to come in, invest and help Europe grow
38:53don't put walls up, that is not going to help our teams here and our businesses here be globally competitive
39:00great, Katrin?
39:01I tend to agree with both of you
39:03I think that maybe a more focused policy making, investment funding for the startup will also help
39:16I think that we are a little bit too much dispersed in terms of development of the activities
39:25so I think that if we can create that global market, for me it's absolutely necessary for the scaling
39:34otherwise, again, for startups, it's impossible to adapt and to adjust to all the markets that we have in Europe
39:43the 27 countries, so if we are not capable to resolve that problem, I think we will
39:52it will be super difficult to scale and to really go in that direction
39:58and again, I think that most of the private companies are ready to do their share
40:04but we cannot do everything with our own arms, so we need also support from the government in all the
40:15countries
40:16from European unions to help us help them, to make sure that we can be successful
40:23and again, Airbus is a perfect example that when Europe wants to do something great, we are capable to do
40:30it
40:30so we started, we were nowhere on the aerospace industry scale
40:37and today we have built a champion
40:39and the same with Ericsson
40:40so I think that when we bring things together, we can be successful
40:45but it needs also political will to be so
40:49What I remember, it's also a matter of competitiveness, right?
40:52this is not just a matter of regulation or compliance or whatever, this is a matter of competitiveness
40:57and I loved your point
40:58Yeah, but I think that the regulation are in the range of competitiveness
41:03Yeah, exactly, and so we need to enable it
41:05So now it's linked, so if we can simplify a little bit again, as it has been already said
41:11I think it will help everyone in this room, including on this stage
41:15Exactly, and I also remember
41:16If I could say one thing, I think we can't afford to be slow
41:20Like, all of this contributes to speed being a problem
41:24Yeah
41:25Because things are moving really fast
41:27and so we just have to keep up that pace as Europe
41:31Alright, I love the excitement and I love the insights
41:34Thank you so much for taking the time, Mikael, Tanujak, Catherine
41:37Thank you for having us
41:38Thank you for having us
41:38Thank you
Comments