Latest / Best of LinkedIn: Digital Transformation & Tech / Best of LinkedIn: Digital-Gipfel 2026

Transcript

Best of LinkedIn: Digital-Gipfel 2026 · Best of LinkedIn: Digital Transformation & Tech

• 0:00
  Speaker 1: This episode is provided by Thomas Allgaier and Frennus, based on the most relevant LinkedIn posts about digital Gipfel twenty-twenty six.
• 0:07
  Speaker 1: Frenness supports ICT in tech providers with AI ecosystem strategy by delivering independent vendor assessment build versus buy analysis an ecosystem intelligence that prevents expensive mistakes and positions the provider's competitively.
• 0:20
  Speaker 1: you can find more info in
• 0:29
  Speaker 0: coming out of the Digital Gipfel twenty-twenty six in Berlin.
• 0:33
  Speaker 1: Exactly, and if you are navigating digital transformation right now The sheer amount of noise coming out these summits can feel completely overwhelming.
• 0:41
  Speaker 0: Oh absolutely it's just a wall buzzword sometimes
• 0:43
  Speaker 1: Right.
• 0:44
  Speaker 1: So our mission today is to cut through all that fluff.
• 0:47
  Speaker 1: We're taking raw insights shared by people actually building this stuff And we going still exactly where AI infrastructure and digital policy heading.
• 0:57
  Speaker 0: Yeah understand where this ecosystem is going, you first have to look at the massive shift in the tone.
• 1:04
  Speaker 0: At The top?
• 1:05
  Speaker 1: The vibe was definitely different This time around.
• 1:07
  Speaker 0: it really was.
• 1:08
  Speaker 0: I mean if you read through the post from attendees the conversation has violently pivoted From what should we do too?
• 1:15
  Speaker 0: how fast can We physically build this
• 1:16
  Speaker 1: right?
• 1:17
  Speaker 1: the impatience Was practically radiating off-the-screen.
• 1:20
  Speaker 1: like You.
• 1:20
  Speaker 1: Look at the government's new ki nation twenty thirty agenda.
• 1:24
  Speaker 0: Yeah, Benedict Bowman broke that down.
• 1:25
  Speaker 0: And
• 1:26
  Speaker 1: it's not just big task lists anymore.
• 1:29
  Speaker 1: the government actually laid out eighteen concrete measures with hard metrics
• 1:33
  Speaker 0: which is a huge step up.
• 1:34
  Speaker 1: It Is they are aiming for eighty percent of large firms in fifty five percent of SMEs to adopt AI by twenty thirty.
• 1:41
  Speaker 1: so They're finally moving toward actual impact logic.
• 1:44
  Speaker 0: well and that shift to Hard Metrics.
• 1:47
  Speaker 0: it represents a necessary maturity I think.
• 1:50
  Speaker 0: But what's really driving those metrics is this new cultural rallying cry.
• 1:54
  Speaker 1: The Team Yes thing?
• 1:55
  Speaker 0: Exactly, Digital Minister Karsten Wildberger was pushing this hard.
• 1:59
  Speaker 0: he essentially demanded that the tech ecosystem join team yes
• 2:03
  Speaker 1: which Michael Kaldige expanded on perfectly.
• 2:05
  Speaker 1: He pointed out that Germany has been trapped in this exhausting cycle of endlessly debating theoretical risks.
• 2:11
  Speaker 0: You know yeah debating it before a single line of code as even deployed.
• 2:14
  Speaker 0: right
• 2:15
  Speaker 1: Colleagues said we need to just retire the team-no over regulators.
• 2:19
  Speaker 1: We need Mutsurluka, The Courage To Leave A Gap.
• 2:22
  Speaker 0: Just launch the product find the flaws and fix it on the fly.
• 2:25
  Speaker 1: Exactly Speed is a feature now not a bug.
• 2:29
  Speaker 0: But hold on I have to push back a little on those metrics.
• 2:32
  Speaker 1: Okay lay it on me.
• 2:33
  Speaker 0: If the big twenty thirty target is fifty five percent SME adoption Doesn't that mean?
• 2:38
  Speaker 0: We are explicitly designing a policy That leaves nearly half the economy behind.
• 2:42
  Speaker 1: That Is Very Fair Point.
• 2:44
  Speaker 0: It's like It's like celebrating a marathon finish when half your relay team is still sitting at the starting line tying their shoes.
• 2:50
  Speaker 1: Yeah, Wolfgang Legblinder highlighted that exact structural flaw.
• 2:54
  Speaker 1: He pointed out that forty-five percent of SMEs will remain without AI under this highly touted plan
• 3:00
  Speaker 0: Which is terrifying for supply chains
• 3:02
  Speaker 1: Totally because you're digital capabilities only as strong as your weakest node.
• 3:06
  Speaker 1: If a giant auto manufacturer has stated the
• 3:08
  Speaker 1: R.A.I.,
• 3:08
  Speaker 1: but their mid-sized supplier is still using like, a whiteboard for inventory... The
• 3:12
  Speaker 0: whole system bottlenecks?
• 3:14
  Speaker 1: Exactly!
• 3:14
  Speaker 1: The
• 3:14
  Speaker 1: A.I.'s
• 3:15
  Speaker 1: bottlenecked by the analog supplier.
• 3:17
  Speaker 1: and honestly this makes the financial reality of the agenda even harder to swallow.
• 3:21
  Speaker 0: Oh..the budget disconnect.
• 3:23
  Speaker 0: Curly Hustet & Kirsten Rulf both posted about it.
• 3:26
  Speaker 1: Yeah....the gap between political promises in actual government spending is massive.
• 3:31
  Speaker 1: The state's AI budget is practically a rounding error compared to what big tech is throwing around.
• 3:35
  Speaker 0: The math just doesn't work, I mean training.
• 3:37
  Speaker 0: your frontier model requires billions in capital expenditure
• 3:40
  Speaker 1: And the government is allocating tiny fraction of that Right.
• 3:44
  Speaker 0: You can't just mandate globally competitive KI nation through policy papers If you don't actually fund the compute clusters.
• 3:51
  Speaker 1: So if Europe physically cannot outspend the hyperscalars on these massive foundation models It completely changes definition of digital sovereignty.
• 4:00
  Speaker 0: It has to.
• 4:01
  Speaker 0: Sovereignty can't mean building every single layer of the tech stack in isolation anymore,
• 4:05
  Speaker 1: right?
• 4:06
  Speaker 1: Chancellor Friedrich Merz essentially admitted this.
• 4:09
  Speaker 1: he defined sovereignty not as autarky but As the freedom to choose and switch providers
• 4:15
  Speaker 0: which is a much more pragmatic take.
• 4:16
  Speaker 1: yeah And Thomas Sauer Esegg from SAP posted a great example of this.
• 4:21
  Speaker 1: SAP isn't trying to build arrival to GPT-IV?
• 4:24
  Speaker 0: No, that would be crazy at this point.
• 4:25
  Speaker 1: Exactly!
• 4:26
  Speaker 1: They are partnering with global players like Mistral and Coheir to bring sovereign AI into regulated industries.
• 4:32
  Speaker 0: And thats where the alliance has come in.
• 4:34
  Speaker 0: Canadian Minister Evan Solomon actually posted about the new Canada-Germany Sovereign Technology Alliance
• 4:40
  Speaker 1: Which proves sovereignty is now all about these trusted international partnerships.
• 4:44
  Speaker 0: Definitely You don't have own the model you just need a framework which controls data environment
• 4:50
  Speaker 1: By the way, real quick.
• 4:51
  Speaker 1: If you're finding this deep dive valuable and want to stay on top of these massive tech shifts You should definitely subscribe so you don't miss our future deep dives.
• 4:59
  Speaker 1: The landscape is just moving so fast
• 5:01
  Speaker 0: it really is.
• 5:02
  Speaker 0: but speaking of the landscape Moving Fast we have to talk about the government's role in actually buying This Tech.
• 5:08
  Speaker 1: Oh?
• 5:08
  Speaker 1: The anchor customer paradox yeah
• 5:10
  Speaker 0: Dominic Rapaki introduced us And It was a brutal observation.
• 5:14
  Speaker 1: that Really Was
• 5:15
  Speaker 0: the federal Government Talks Constantly About Being an Anchor Customer for European Tech But Rappacki noted that while they write these strategy papers, federal intelligence agencies are quietly moving over to Google Cloud.
• 5:27
  Speaker 1: Wait the intelligence agencies?
• 5:28
  Speaker 0: The ones who should care most about sovereignty?
• 5:30
  Speaker 1: Exactly!
• 5:32
  Speaker 1: They
• 5:32
  Speaker 0: prioritize immediate capability of long-term ecosystem building.
• 5:36
  Speaker 0: but meanwhile six German states actually signed a real contract for sovereign education platform
• 5:41
  Speaker 1: Right using NexCloud and IONUS.
• 5:43
  Speaker 1: Yes
• 5:44
  Speaker 0: So actual execution is happening at state level While Federal level is just paralyzed by debates.
• 5:49
  Speaker 1: Okay, but I have to challenge this whole sovereignty through partnerships narrative for a second.
• 5:54
  Speaker 1: Isn't this version of digital sovereignty basically just like renting a highly secure apartment in a building owned by someone else?
• 6:02
  Speaker 0: Like
• 6:02
  Speaker 1: sure you have your own locks But if you don't own the underlying infrastructure Can you ever truly control the building?
• 6:08
  Speaker 1: If the global hyperscaler goes down Your sovereign AI goes dark.
• 6:12
  Speaker 0: That is the ultimate vulnerability And it actually ties into the physical reality of all this, which Sehan Adelaide and Andreas Tomobach highlighted.
• 6:19
  Speaker 1: Right!
• 6:20
  Speaker 1: The real bottlenecks aren't just software
• 6:22
  Speaker 0: Not at all.
• 6:23
  Speaker 0: It's energy, grid capacity, permitting for data centers.
• 6:26
  Speaker 0: These AI ambitions are violently crashing to physical limits.
• 6:30
  Speaker 1: We're talking copper & concrete here.
• 6:32
  Speaker 0: Exactly You can have best code in world but if you cant get permits.
• 6:36
  Speaker 0: build a massive data center to power is useless
• 6:39
  Speaker 1: Which is why the ecosystem's pivoting so hard.
• 6:42
  Speaker 1: Since Europe has facing these infrastructure bottlenecks, they are doubling down on where they already have an unfair advantage
• 6:48
  Speaker 0: The physical world.
• 6:49
  Speaker 1: Deep industrial engineering.
• 6:51
  Speaker 0: Yes!
• 6:51
  Speaker 1: The focus of physical AI was massive at this summit... ...the government is planning to invest over three hundred million euros here by twenty thirty.
• 6:59
  Speaker 0: And Tanya Rukert had a great post contextualizing why it matters so
• 7:03
  Speaker 1: much The ninety-five percent accuracy thing?
• 7:06
  Speaker 1: Yeah
• 7:06
  Speaker 0: She pointed out that if a chatbot is ninety-five percent accurate, it's fine.
• 7:10
  Speaker 0: It's little annoying maybe but
• 7:11
  Speaker 1: If an automated driving system is ninety five percent accurate
• 7:14
  Speaker 0: people die.
• 7:15
  Speaker 0: the margin for error Is zero?
• 7:17
  Speaker 1: Exactly.
• 7:18
  Speaker 1: and The scale of this industrial AI has already kind of mind blowing.
• 7:21
  Speaker 1: Holger Wolf highlighted That Schwartz digits in Little are making two point eight billion highly Automated AI driven replenishment decisions every single day.
• 7:31
  Speaker 0: Two point eight Billion.
• 7:32
  Speaker 0: just let us thank you
• 7:33
  Speaker 1: staggering.
• 7:34
  Speaker 1: It's not a human looking at a spreadsheet, it is an AI instantly analyzing weather foot traffic and inventory to route specific pallet into the store.
• 7:43
  Speaker 0: That is real applied intelligence!
• 7:45
  Speaker 0: And then Christian Sotter brought up Neuro Robotics which is taking this in actual motion.
• 7:50
  Speaker 1: Right they are tracking completely new metric.
• 7:52
  Speaker 1: Time To Skill.
• 7:54
  Speaker 0: How fast do robot learns a skill?
• 7:57
  Speaker 0: Not about programming coordinates anymore but machine figuring out its environment on fly.
• 8:01
  Speaker 0: It's the new ultimate competitive metric.
• 8:04
  Speaker 1: Which brings us to The Brownfield first strategy, OMA-G posted about this and it is such a pragmatic concept!
• 8:11
  Speaker 0: Oh really?
• 8:11
  Speaker 0: Instead of just trying to build shiny new data centers... ...the green field approach you retrofit AI
• 8:17
  Speaker 1: Right.
• 8:18
  Speaker 1: You take millions of existing legacy machines in factory floors And install AI agents into them To make them intelligent.
• 8:31
  Speaker 0: Perfectly bypasses those great capacity bottle next week talked about earlier.
• 8:35
  Speaker 1: Okay, but again I have to play devil's advocate.
• 8:37
  Speaker 1: Are we being overly optimistic?
• 8:39
  Speaker 1: About leading in physical AI?
• 8:41
  Speaker 0: how so
• 8:42
  Speaker 1: well Merkel Ross noted the gap with China In robotics hardware is actually widening.
• 8:49
  Speaker 0: that's true.
• 8:50
  Speaker 1: So it this brownfield first approach?
• 8:53
  Speaker 1: Brilliant pragmatism like turning a fleet of vintage cars into self-driving vehicles.
• 8:58
  Speaker 1: Or is it just a convenient excuse because we can't manage to build the Tesla factory in the first place.
• 9:03
  Speaker 1: Oh
• 9:04
  Speaker 0: That's a harsh way to frame it, but it's a valid fear.
• 9:07
  Speaker 0: It's a dangerous line to walk.
• 9:09
  Speaker 0: You need the pragmatist them to upgrade the legacy assets But you absolutely Can't use it as an alibi for failing to innovate at the frontier
• 9:16
  Speaker 1: right?
• 9:16
  Speaker 1: You still need that team yes execution mindset to build The new infrastructure
• 9:21
  Speaker 0: exactly Less debating, more deploying.
• 9:25
  Speaker 1: So if we pull all these threads together We've got the shift toward hard metrics The evolution of digital sovereignty into strategic partnerships And this massive push to retrofit legacy machines with physical AI agents at the edge.
• 9:40
  Speaker 0: It creates a really fascinating tension.
• 9:42
  Speaker 1: it does and I want leave you the listener With a slightly wild scenario to ponder
• 9:48
  Speaker 0: why like where?
• 9:48
  Speaker 0: This is going
• 9:49
  Speaker 1: think about.
• 9:50
  Speaker 1: If industrial AI is about upgrading legacy machines to make their own autonomous decisions.
• 9:56
  Speaker 1: And the government is struggling to act as a unified anchor customer, what happens when these newly intelligent physical AI agents start automatically procuring?
• 10:08
  Speaker 0: Or what if they start dynamically switching their own cloud providers based on real-time efficiency and energy costs?
• 10:13
  Speaker 1: Exactly.
• 10:14
  Speaker 1: Will
• 10:14
  Speaker 0: the AI itself end up enforcing the digital sovereignty that politicians are currently just endlessly debating?
• 10:21
  Speaker 1: The machines bypassing strategy papers entirely, That is a wild but very real possibility to chew on.
• 10:28
  Speaker 0: It really is.
• 10:29
  Speaker 1: If you enjoyed this episode, new episodes drop every two weeks.
• 10:32
  Speaker 1: Also check out our other editions on Cloud Insights and Sovereignty, AI & Agentec Systems, DefenseTech, Digital Products & Services, Green ICT & Sustainable AI, ICT and Tech Insights, HealthTech!