Best of LinkedIn: Cloud Insights CW 47/ 48

00:00:00: This episode is provided by Thomas Allgaier and Frenus based on the most relevant LinkedIn posts about cloud in calendar weeks, forty seven and forty eight.

00:00:08: Frenus enables enterprises with market technology and competitive intelligence for portfolio and strategy development.

00:00:16: This edition is brought to you by our partner, Google Cloud.

00:00:18: Don't miss out on their upcoming event, Google Cloud Live plus labs Geneva on fourth December.

00:00:23: Find the link to the event in the description.

00:00:25: Welcome to the deep dive.

00:00:27: Our mission is well.

00:00:28: pretty clear.

00:00:29: We're giving you that ultimate strategic shortcut to understanding where the cloud market is right now.

00:00:34: We've pulled the top insights from LinkedIn in weeks, forty seven and forty eight.

00:00:38: And what they show is a a really massive accelerating shift in priorities.

00:00:43: It's absolutely a shift.

00:00:44: It's gone from a focus on raw capability to operational control.

00:00:49: For so long, the whole conversation was just about what the cloud could do.

00:00:52: Now, the overwhelming focus is on how you govern it, how you control the costs, and really how you build resilience against, well, against geopolitical and operational risk.

00:00:59: The market's matured and it's matured fast.

00:01:02: We've basically grouped the intensity of the last few weeks into three big themes.

00:01:06: First up, we've got the accelerating reality of digital sovereignty, especially in Europe, and what that actually means for global architecture.

00:01:15: It's moving from policy to an industrial imperative.

00:01:19: And second, we're gonna dig into the, let's call them the painful realities of phenops.

00:01:23: Companies are still hemorrhaging money, and we're looking beyond just tagging problems to the cultural and governance gaps that are causing the leaps.

00:01:31: And finally, we'll look at how AI is fundamentally reshaping cloud infrastructure.

00:01:37: It's pushing us toward what some are calling the agentic enterprise and this massive race for continental scale compute.

00:01:44: So let's jump right in.

00:01:46: The strategic imperative.

00:01:48: sovereignty.

00:01:48: When we talk about digital sovereignty, the feeling we got from the sources, you know, from experts like Christian Schultz and Dario Meisto, is that this isn't a policy discussion anymore.

00:01:56: It's a full-on strategic imperative.

00:01:58: It's moved straight into industrial policy.

00:02:01: It absolutely has.

00:02:01: And that alignment is crucial.

00:02:03: LaTace Bozeman, for example, highlighted how France and Germany are now converging on a shared definition.

00:02:08: And that's effectively signaling a bi-European approach for public sector services.

00:02:14: This isn't just talk.

00:02:16: Europe is moving from, you know, discussing digital autonomy to actually enforcing it through strategic procurement.

00:02:22: But the term sovereignty itself can be tricky.

00:02:25: It often gets mixed up with just, you know, data location.

00:02:29: Susan Brown made a really vital clarification on this.

00:02:31: Yes, she did.

00:02:32: She said sovereignty isn't about the physical location of a data center.

00:02:35: It's fundamentally about the control model.

00:02:38: Why is that distinction so important for organizations?

00:02:40: It

00:02:41: matters because location alone gives you zero protection if the provider is still legally subject to extraterritorial demands.

00:02:49: As Dave Michaels pointed out, the main reason for using a sovereign cloud isn't just local privacy rules.

00:02:56: It's specifically about protecting against those external data requests.

00:03:00: It's a battle of jurisdiction and governance over who has access, who holds the keys, and who runs the operations.

00:03:06: Which of course brings us to the elephant in the room that causes most of this anxiety.

00:03:11: The U.S.

00:03:12: Cloud Act.

00:03:14: We've seen so much continued concern around this.

00:03:16: What exactly is the legal friction that people like Jim Loicano and Chia Kamiri are pointing out?

00:03:22: The core friction is the power of the Cloud Act to compel U.S.

00:03:26: base providers to produce data in a, quote, readable form.

00:03:31: This legal obligation can legally supersede a lot of the technical controls you might put in place, even if you manage your own encryption keys.

00:03:38: So even with your own keys there's a risk.

00:03:40: For regulated European entities, that legal vulnerability, the chance that a foreign government could compel access, it remains a totally unacceptable risk, regardless of where the bits are physically sitting.

00:03:50: So if the legal framework is the core problem, how are the big three hyperscalers trying to address this?

00:03:56: How are they trying to mitigate that risk for their EU customers?

00:03:59: Well, Chris Menatus give a great breakdown of the three distinct strategies.

00:04:03: Microsoft is leaning heavily on its EU data boundary.

00:04:06: They're ensuring processing, even for tools like Copilot, stays within those EU bounds, and that's all backed by customer-managed keys.

00:04:13: So that's more of a technical, geo-fenced solution?

00:04:16: How's Google's approach different?

00:04:18: Google is taking a partnership route.

00:04:19: They're delegating parts of sovereignty to local trusted partners, setting up things like the Sovereign Cloud Hub in Munich.

00:04:27: It's a trust through regional engagement model where the governance is handled closer to home.

00:04:32: And then there's AWS, which seems to be taking the strongest stance on actual separation.

00:04:37: That's right.

00:04:38: Catherine Renz detailed their strategy.

00:04:40: They are creating a completely separate new EU governed cloud.

00:04:46: And the key here isn't just location, it's autonomous operations.

00:04:50: Autonomous how?

00:04:51: It's operated exclusively by EU citizens and designed to function entirely on its own, even if connectivity to the global AWS network is cut.

00:04:59: It's the highest level of physical, organizational, and human separation you can get to try and mitigate those risks we were just talking about.

00:05:06: These are major commitments, and they're already landing huge contracts.

00:05:09: Luke Leon reported that NATO recently awarded Google Cloud a multi-million dollar contract to build a fully air-gapped sovereign cloud.

00:05:17: And we're seeing this shape new tech, too.

00:05:20: SAP, for instance, is introducing its EU AI cloud.

00:05:23: Sebastian Steinhauser made the point that they get it.

00:05:26: AI innovation has to be paired with trusted sovereign governance from day one, or regulated industries just won't touch these new tools.

00:05:34: And on the regulatory side, Europe is definitely stepping up its oversight.

00:05:39: Henover Keunen and Dr.

00:05:40: Apostolis Criticopoulos noted that the European Commission launched investigations under the Digital Markets Act.

00:05:47: the DMA.

00:05:48: To see whether AWS and Microsoft Azure are acting as market gatekeepers, even if they don't hit those formal thresholds.

00:05:55: And that's a huge move.

00:05:56: It signals that the EU sees control over foundational cloud infrastructure as a strategic issue, not just a numbers game.

00:06:03: If you dominate the infrastructure everyone else needs, you're going to face more scrutiny.

00:06:07: It's a trend everyone needs to watch.

00:06:09: This regulatory pressure, it naturally raises the question, what about the local European alternatives?

00:06:14: Christoph Stalmach made a really compelling argument that telcos are maybe Europe's most unrealized sovereign cloud asset.

00:06:20: Why are they so well positioned?

00:06:22: Well, they're positioned because they inherently cover the four pillars of sovereignty.

00:06:27: They control data location.

00:06:29: They know who is physically touching the hardware.

00:06:31: They govern the network flow.

00:06:32: And crucially, they have decades of experience with national compliance and running critical national infrastructure.

00:06:38: I don't have to build trust from scratch.

00:06:40: Exactly.

00:06:40: They've already earned it.

00:06:42: And we're

00:06:42: seeing domestic players like OVH Cloud in France and Hetzner in Germany really capitalizing on this as failure companions.

00:06:50: detailed.

00:06:51: Sovereignty isn't

00:06:52: just about how the US hyperscalers react.

00:06:54: It's about enabling real continental alternatives.

00:06:57: Okay.

00:06:58: So we've established that global scale is being challenged by these new geopolitical borders and the need for control.

00:07:04: But that need for control applies just as fiercely to internal budgets.

00:07:08: Let's switch from jurisdictional friction to financial friction.

00:07:12: Phenoms.

00:07:13: Despite being a mature discipline, companies are still struggling with just basic cost visibility.

00:07:18: They are, and the struggle almost always starts at the data layer.

00:07:21: It's messy, it's inconsistent.

00:07:23: Eric Lam identified six common data traps that actively stop organizations from getting any real value from their spending metrics.

00:07:32: If your data is bad, your insights are useless.

00:07:35: So what are the biggest traps?

00:07:37: killing organizations right now.

00:07:39: Well, tagging chaos is pretty universal.

00:07:42: You've got inconsistent, non-normalized tags across a multi-cloud setup, which makes it impossible to reliably track spend.

00:07:49: But the one that finance absolutely hates is the fruit salad trap.

00:07:52: Fruit salad?

00:07:53: Yeah,

00:07:54: mixing currencies.

00:07:54: You can't just mix cost metrics in euros, dollars, and yen and then show that salad to finance and expect them to trust it.

00:08:00: You have to normalize the currency or the whole report is compromised from the start.

00:08:05: It sounds so basic.

00:08:06: But I can see how, with hundreds of accounts globally, it gets overlooked.

00:08:10: And this all leads to that master dashboard problem, where one dashboard tries to serve everyone and ends up serving no one.

00:08:16: Exactly.

00:08:17: Engineers need utilization data.

00:08:19: Finance needs forecasts.

00:08:21: Leadership needs high level KPIs.

00:08:23: One dashboard just fails everybody.

00:08:25: And beyond the reporting, you've got these silent cost leaks hiding in the default settings.

00:08:30: Can

00:08:30: you give us an actionable example, something that could be fixed today?

00:08:33: Sure.

00:08:33: Riaz Syod had a great playbook for cutting runaway CloudWatch log costs, which are just a notorious cost drain.

00:08:40: Three simple moves.

00:08:41: First, move production to the Inafo log level, not dehebug.

00:08:45: Second, sample your debug logs.

00:08:47: Don't ingest everything.

00:08:49: And third, adjust retention.

00:08:51: The default is never expi- which leads to massive indefinite storage costs.

00:08:56: Just limiting it to fourteen or thirty days can provide huge immediate savings.

00:09:01: It's instantly actionable.

00:09:02: But we know FINOP's maturity goes way beyond simple config changes.

00:09:06: DSAF Lvivano noted that while providers keep releasing really good FINOP's features like forecasting or EKS cost allocation, the tools themselves often fall short.

00:09:14: the second you introduce human complexity.

00:09:16: And that is the critical organizational insight.

00:09:19: The tools are brilliant at the infrastructure layer, but they don't inherently manage human accountability or cross-departmental budgets.

00:09:27: That organizational gap is often just too big for internal teams to bridge on their own.

00:09:32: And Grant Byram, citing Accenture's work, highlighted that exact point.

00:09:37: Bringing in a strategic partner can accelerate maturity because they bridge that execution gap.

00:09:42: They bring the operational design needed to make the tools actually work in a complex company.

00:09:47: Which brings us back to culture.

00:09:48: Dan W made a crucial observation that true fine ops maturity means moving past just saying, we need a culture change.

00:09:55: You have to specifically understand the behavioral dynamics, the cognitive biases, and the motivations that shape spending decisions inside your organization.

00:10:04: Then you design the culture around that.

00:10:06: So it's not enough to just say, be cost aware.

00:10:09: You have to design the incentives and the workflows that address how people actually behave when they provision resources.

00:10:15: Precisely.

00:10:16: If there's no accountability until the bill shows up, that's a governance failure, not a tooling failure.

00:10:21: And that's why integrating FinOps with established frameworks is so key.

00:10:25: Nicholas Fondrini highlighted connecting FinOps to COVID to provide that necessary governance backbone.

00:10:31: And how does a framework like COVID help FinOps get to that next level?

00:10:35: It provides structure and discipline.

00:10:37: It ensures spending decisions aren't just efficient, but they're auditable, they're aligned with enterprise value, and they're tied back to specific accountability points.

00:10:46: Phenops gives you the data.

00:10:47: Kobit gives you the strategic control framework to enforce the change.

00:10:52: Alright, shifting gears to our final theme.

00:10:54: How AI is fundamentally reshaping the cloud infrastructure of the future, which demands maximum operational maturity.

00:11:01: Aiman Hussein noted that organizations have moved from cloud curiosity to cloud confidence.

00:11:06: Confidence is definitely the keyword.

00:11:08: It signals that modernization is now a strict discipline, not just an experiment.

00:11:13: We're moving way past the simple lift and shift mindset.

00:11:16: To get AI ready, you have to embrace cloud-native design containers, serverless, to get those faster learning loops.

00:11:23: And if you don't make that architectural shift, you just don't see the business value, no matter how great the cloud platform is.

00:11:30: Bebeck Chatterjee's key insight was that the cloud isn't failing, strategy translation is.

00:11:35: Exactly.

00:11:36: The tech stack is often fine, but the business value arrives slowly, or not at all, if the architecture wasn't explicitly built for specific needs like real-time risk scoring or cost-controlled analytics for AI training.

00:11:50: If you don't engineer for specific value, you won't get it.

00:11:53: So if that's the discipline required, what does the future operating model look like?

00:11:56: What are the KPIs for leadership in this AI era?

00:11:59: Benjamin Herman proposed three interconnected KPIs for leaders aiming for twenty-twenty-six, which really defined this agentic enterprise.

00:12:07: First, they need to get over fifty percent true public cloud adoption for their underlying infrastructure.

00:12:12: And

00:12:12: the second.

00:12:12: Achieve seventy percent everyday AI for everyone adoption.

00:12:16: AI can't be stuck in an R&D lab.

00:12:18: It has to be pervasive from HR to sales.

00:12:21: And the third KPI really seems to tie it all together, ensuring every business unit offers its expertise as a chatbot.

00:12:27: That third point is the essence of the agentic enterprise.

00:12:31: It means every department legal finance operations has structured its core knowledge so it can be served instantly by an AI agent.

00:12:39: It transforms internal processes from human mediated cues into AI native knowledge bases.

00:12:44: The efficiency gain is immense.

00:12:47: And to support that vision, the infrastructure investment is just staggering.

00:12:51: We're seeing a continental scale AI infrastructure race.

00:12:54: We really are.

00:12:55: Elijah Zanov and Manish Ranjan documented some of these huge investments.

00:12:59: Microsoft committed ten billion dollars to an AI data center hub in Portugal, just to house over twelve thousand NVIDIA GPUs.

00:13:07: Wow.

00:13:08: Meanwhile, Google is pouring five and a half billion euros into German data centers.

00:13:12: This kind of investment is about creating sovereign, accelerated, AI-ready platforms in key strategic locations.

00:13:19: And the internal architecture has to change to manage that scale.

00:13:22: Aviv Weiss described the private cloud operating model, or PCWIM.

00:13:26: This is more of an organizational pivot than a technical one.

00:13:28: It's a total transformation, where traditional silos compute, network, data stop operating separately, and align as a single unified platform team.

00:13:38: They're goal shifts.

00:13:39: They're no longer just delivering IT projects.

00:13:42: They're delivering integrated AI-ready products to the business.

00:13:46: That focus on unifying the data plane seems absolutely crucial.

00:13:49: I mean, if you don't know where your data is, your AI agents will never hit that seventy percent adoption rate.

00:13:55: Data ops is the key.

00:13:57: Dhruv R shared a great practical example.

00:14:00: They helped a global retailer cut data discovery time by fifty percent just by enforcing unification.

00:14:06: They put everything into a central cloud data catalog and rigorously mapped the lineage.

00:14:11: That shift from data chaos to data confidence is the absolute prerequisite for the agentic enterprise.

00:14:17: That brings everything full circle, doesn't it?

00:14:19: We've got this pressure for local control and sovereignty clashing with the relentless push for AI innovation and global scale, all of which require intense phenox discipline to manage.

00:14:28: That is the central tension, absolutely.

00:14:30: As Konstantin Rolsov noted, cloud sovereignty should be seen as a strategic advantage that builds trust, not just a compliance checkbox.

00:14:39: Organizations have to balance that global capability with geopolitical trust.

00:14:43: Which means every single strategic infrastructure choice you make today is a fundamental statement about resilience and, frankly, national autonomy.

00:14:52: As Ashish Kumar put it, geopolitics and data privacy are drawing new borders in the cloud.

00:14:58: Understanding your Phenops readiness is now completely inseparable from understanding your regulatory risk.

00:15:03: If you enjoyed this episode, new episodes drop every two weeks.

00:15:07: Also check out our other editions on ICT and tech, digital products and services, artificial intelligence, sustainability and green ICT, defense, tech

00:15:14: and health.

00:15:15: Thanks for diving deep with us.

00:15:16: We'll catch you next time.