Best of LinkedIn: Deutsche Telekom's & NVIDIA's Industrial AI Cloud

00:00:00: This episode is provided by Thomas Allgaier and Frennus, based on the most relevant LinkedIn posts talking about Deutsche Telecom and NVIDIA launching an industrial AI cloud.

00:00:09: Frennus supports ICT enterprises with market and competitive intelligence, decoding emerging technologies, customer insights, regulatory shifts and competitor strategies, so product teams and strategy leaders don't just react, but shape the future.

00:00:24: Okay, let's unpack this.

00:00:26: We're diving into a major initiative that's really stirring things up in the digital transformation and tech

00:00:31: world.

00:00:31: Definitely.

00:00:32: We're talking about Deutsche Telecom and NVIDIA launching the industrial AI cloud.

00:00:36: And

00:00:37: this is aimed squarely at giving Europe a sovereign industrial backbone.

00:00:41: Exactly.

00:00:42: And what's fascinating here is how fast this all happened.

00:00:44: I mean, from concept to launch and the sheer executive muscle behind it.

00:00:48: We're going to synthesize the LinkedIn chatter from CEOs and industry leaders to really pull out the top trends and insights you need to know about this new European AI factory.

00:00:57: So let's start there.

00:00:58: with the scale, the speed, and the location.

00:01:02: The conversation on LinkedIn was just electric, defined by these huge numbers and incredible speed.

00:01:09: What are the anchor figures that really set the stage?

00:01:12: Well,

00:01:12: the scale is immense, and it really signals a serious, serious commitment.

00:01:17: We're looking at an investment.

00:01:18: sighted by several people, including Benjith Benny, of one billion euros.

00:01:22: A billion euros, so about one point two billion dollars.

00:01:24: Right,

00:01:25: and the centerpiece of that is the capacity.

00:01:27: We're talking ten thousand high-performance NVIDIA GPUs.

00:01:30: These are the top tier systems.

00:01:32: So not just any GPUs.

00:01:33: No,

00:01:33: specifically the DGX B- two hundred systems, you know, the premium stuff for massive AI model training, plus specialized RTX Pro.

00:01:41: six thousand servers.

00:01:42: Okay, but I have to stop you there.

00:01:44: We hear billions announced all the time in the cloud world.

00:01:48: Is one billion euros really enough to challenge the U.S.

00:01:51: hyperscalers who, you know, spend that much quarterly?

00:01:54: Yeah.

00:01:55: Is this a real infrastructure challenge or more of a high-profile niche play?

00:01:59: That's a great question, and I think the sources suggest it's very strategic.

00:02:02: It's about what that billion buys in the context of Germany.

00:02:05: Yeah.

00:02:06: Dr.

00:02:06: Ferry Obelhasen and Anshuman Ja, they both noted this single deployment is expected to boost Germany's entire AI computing power by Uh, roughly fifty percent.

00:02:18: Fifty percent overnight.

00:02:20: Basically overnight.

00:02:21: Just immediately shifts the whole landscape.

00:02:22: Patrick Gompelman quantified the power at zero point.

00:02:25: five exit flops.

00:02:26: Which is, what, five hundred quadrillion calculations a second?

00:02:29: Exactly.

00:02:30: So the point isn't to match the biggest clouds globally.

00:02:34: It's to force a structural change in the price and access for sovereign compute, right, in Europe's most critical industrial sectors.

00:02:41: That fifty percent national boost, that really puts it in perspective.

00:02:45: Now, the timeline.

00:02:46: Munich is the anchor, but the speed of delivery has been the thing everyone's talking about.

00:02:51: Oh, yeah.

00:02:51: Tim Hutkes, the CEO of Dorches Telecom, he noted they went from concept to launch in just six months.

00:02:57: He called it a sprint, not a stroll.

00:02:59: And the service horizon is Q one, twenty twenty six.

00:03:02: Six months for a project this complex with a new data center and ten thousand GPUs, I mean.

00:03:08: Yeah, it's just unheard of in this kind of regulatory environment.

00:03:11: Michael P. Unger specifically commented on this.

00:03:13: He noted the intense speed and.

00:03:16: And the attention to detail required from teams at both NVIDIA and Deutsche Telekom, it really speaks to this profound sense of urgency in the European private sector.

00:03:27: Like they need to capture this opportunity before the window closes.

00:03:30: That urgency, it's a perfect lead-in to our second theme, the sovereignty imperative.

00:03:36: Why the rush?

00:03:37: Why this specific German-based infrastructure?

00:03:41: This is really the heart of the whole thing.

00:03:43: The consistent framing from all the leaders is sovereign AI or made for German Europe.

00:03:49: It's not just about being fast.

00:03:50: It's about trust.

00:03:51: It's about control.

00:03:52: So making sure the infrastructure and the data all that it hears to strict European standards.

00:03:57: Exactly.

00:03:58: It's about countering that dependency on non-European cloud footprints and, you know, all the jurisdictional risks that come with them.

00:04:05: And Julie Reinhardt pointed out something crucial here.

00:04:07: This is a private sector initiative.

00:04:09: No government subsidies.

00:04:10: Right, but it's constantly framed as being vital for the national interest.

00:04:14: Christian Klein, for example, emphasized this is a bold step towards a digitally sovereign Europe.

00:04:20: They partner for the best tech, but they retain total control over the data.

00:04:24: So to really understand why that's so important, we have to look at this distinction that Colin Massen raised.

00:04:29: Yes, he clarified, this is sovereign by origin.

00:04:32: This is key.

00:04:33: It means the physical infrastructure is EU owned, it's operated in the EU, and it's governed entirely by EU law.

00:04:40: It's a direct challenge to the sovereign by contract model that the US hyperscalers offer.

00:04:46: And for a company with sensitive IP, what's the actual Tangible difference there?

00:04:50: well

00:04:51: with sovereign by contract you might get a promise your data will stay and say Frankfurt But the company providing the cloud is still subject to the laws of its home country

00:05:00: I see.

00:05:00: so there's still ambiguity

00:05:02: a lot ambiguity especially for that crown jewel.

00:05:04: industrial data.

00:05:06: sovereign by origin aims to just eliminate that the hardware the Processing the operational control.

00:05:12: It's all on German soil under German law full stop.

00:05:16: That makes the big launch event make so much more sense.

00:05:20: Niels H. Slasier captured the mood perfectly.

00:05:23: He quoted minister Dorothy Baer saying, the time for excuses is over.

00:05:27: And when you have Jensen Huang there saying, the first step is the hardest, you know, the political weight behind this is just enormous.

00:05:34: It's a mandate.

00:05:35: Okay, so the rationale is political, the scale is vast.

00:05:38: Let's get to our third theme, technology, physical AI and the use cases.

00:05:43: How does all this compute power actually translate into applications for European industry?

00:05:48: This is where it gets really exciting.

00:05:50: We shift from bits and bytes to the real world.

00:05:53: Magdalena Jonsak highlighted that the next revolution is physical.

00:05:57: Physical?

00:05:57: Meaning what exactly?

00:05:58: Meaning AI that leaves the screen.

00:06:00: It's not just organizing data anymore.

00:06:01: It's AI that directly runs machines, optimizes entire supply chains, controls factory logistics, and it's targeted right at Europe's strongest players, mechanical engineering, manufacturing, robotics.

00:06:12: And the CEOs shared some of those aha moments, right?

00:06:16: Tim Hutchies and Christine Neckfest-Nicolet gave some really specific examples.

00:06:21: They did take the pharmaceutical sector with this cloud.

00:06:24: AI could calculate optimal temperatures, formulations, batch variables, getting a new medicine market ready in eight months instead of fourteen.

00:06:33: Eight months instead of fourteen.

00:06:34: That's a revolution in itself.

00:06:36: Huge.

00:06:37: And in the automotive sector, AI can analyze sensor data across assembly lines in real time.

00:06:43: It predicts failures before they happen.

00:06:45: They're talking about reducing unplanned stoppages by up to seventy percent.

00:06:49: Seventy

00:06:49: percent.

00:06:50: That's not a marginal game.

00:06:52: that is a fundamental change to how a factory operates.

00:06:55: What about something like aviation?

00:06:57: For aviation and engineering, it's all about digital wind tunnel tests.

00:07:00: This AI capacity means they can run simulations that are incredibly complex.

00:07:05: It can calculate subtle changes to reduce air resistance so a new airplane could be developed with twenty percent less material.

00:07:11: That kind of optimization must rely heavily on digital twins and advanced simulation.

00:07:15: Precisely, and this is all tied into NVIDIA platforms like Omniverse.

00:07:20: Elvira Gonzalez-Soviet noted that developers can now build these super accurate, three-D digital twins of whole factories.

00:07:26: So you can plan and test virtually before you build anything.

00:07:29: That's the idea.

00:07:30: And Roland Bush, the CEO of Siemens, confirmed that their big customers, like Mercedes-Benz and BMW, will be using this infrastructure for exactly that, for those incredibly data-heavy simulations.

00:07:43: This sounds way too big for just one or two companies to pull off.

00:07:47: Which brings us to our final theme, the partnerships and the ecosystem.

00:07:52: So let's start with the core alliance.

00:07:53: what Monica Scheller and Thomas Arzee called, the Germany stack.

00:07:58: The Germany stack.

00:07:59: It's a three-way powerhouse.

00:08:00: You've got Butch Telecom providing the physical layer, the connectivity, the secure data centers, all the security.

00:08:06: Okay,

00:08:06: the foundation.

00:08:07: Then NVIDIA supplies the raw power, the high-performance GPUs in the software layer, and then, crucially, SAP delivers the SAP Business Technology Platform, BTP, which integrates everything into the core enterprise applications where the business data actually lives.

00:08:23: That integration is vital.

00:08:25: And beyond that core group, Carlo Ruiz noted there are over twenty partners involved already.

00:08:30: Who are some of the key early adopters?

00:08:32: It's a very focused list.

00:08:33: Siemens is there, obviously, for the industrial tech side.

00:08:37: EY is on board as a pilot customer, which Dr.

00:08:40: Florian Dick-Riber mentioned, testing it in real-world scenarios.

00:08:43: And

00:08:44: the more specialized players.

00:08:45: You see firms like Agile Robots, Wandelbots is offering their NoVA platform for robot programming, and Quantum Systems is connecting their software for real-time data processing.

00:08:55: Even Perplexity is in the mix for conversational AI.

00:08:59: This brings us to a really strategic point that Sebastian Burroughs made.

00:09:02: This isn't just a new service.

00:09:04: This could fundamentally redefine the role of a telco.

00:09:07: It's a radical shift.

00:09:08: The way NVIDIA is positioning this, it's not about selling more bandwidth.

00:09:12: It's about transforming all that telecom infrastructure, the towers, the data centers, into what they call factories of intelligence.

00:09:19: Factories of intelligence.

00:09:20: I like that.

00:09:21: It's

00:09:21: a critical move for telcos.

00:09:23: Their goal is to aim for the one hundred and seventeen trillion dollar labor economy.

00:09:27: Okay, break that down.

00:09:29: What does that massive number actually mean in simple terms?

00:09:32: It means telcos are shifting from selling access to selling automated outcomes.

00:09:37: That huge figure, the one hundred seventeen trillion, that's the global labor economy where machines and robots are going to generate the next wave of GDP.

00:09:46: So by hosting the AI that controls the machines.

00:09:49: Deutsche

00:09:49: Telecom shifts from being just a pike provider to being a foundational part of creating industrial value.

00:09:56: They're selling the intelligence that makes a factory seventy percent more efficient, not just the connection that goes factory.

00:10:02: That really does change the game.

00:10:04: So finally, how can everyone from startups to the middle stand actually get access to this.

00:10:09: Dr.

00:10:10: Faria Bulkassen detailed a really flexible model.

00:10:13: It's designed to be accessible.

00:10:14: Customers can just book raw GPU power.

00:10:17: You know, the classic GPU as a service.

00:10:18: The basic model.

00:10:19: Right.

00:10:20: Or they can combine that with pre-trained AI models for their industry.

00:10:23: Or they can go for the full managed service.

00:10:26: The posts all emphasize that it's for everyone.

00:10:28: Startups, mid-sized companies, industry giants, and even academic research.

00:10:32: If you enjoyed this episode, new episodes drop every two weeks.

00:10:36: Also check out our other editions on cloud insights, sustainability in green ICT, digital products and services, health tech, defense tech, ICT and tech insights, and artificial intelligence.

00:10:47: So the launch of this industrial AI cloud, it clearly establishes Europe's foundation for sovereign AI.

00:10:53: But the conversation on LinkedIn shows that the core dilemma we started with is still there.

00:10:58: The infrastructure is governed by German law, sure.

00:11:01: But the chips, the firmware, the software stacks, they're global.

00:11:06: So for you, the tech professional thinking about strategy, the key takeaway is this.

00:11:10: Sovereignty doesn't just end at the data center walls.

00:11:13: The real question is, how must European industry invest in building competence and real innovation on top of this physical infrastructure to ensure true technological independence in the next decade.

00:11:23: That's a powerful and essential question.

00:11:25: Thank you for joining us for this deep dive.

00:11:27: Don't forget to subscribe so you don't miss our next edition.