Best of LinkedIn: ICT & Tech Insights CW 39/ 40

00:00:00: This episode is provided by Thomas Allgaier and Freeness, based on the most relevant LinkedIn posts about ICT and tech insights from CW-ThirtyNine and Forty.

00:00:09: Freenes supports ICT enterprises in the form of delivering precise ICT market and pricing intelligence that analyst subscriptions and existing databases cannot provide.

00:00:19: you can find more info in the description.

00:00:21: You know i was thinking about this scenario The other day.

00:00:24: imagine your uh acquiring a company for Let's say three billion dollars.

00:00:28: Okay, that's a pretty massive acquisition right?

00:00:31: Massive but then right in the middle of post merger integration You realize their entire digital infrastructure is well.

00:00:38: It's basically fundamentally compromised.

00:00:41: Wow.

00:00:42: and the craziest part Is The threat hasn't even fully materialized yet But the underlying cryptography That they use is essentially this ticking time bomb And it has baked into literally every microservice you just bought.

00:00:54: Yeah, that is exactly the kind of reality check we are dealing

00:00:58: with today.

00:00:59: Absolutely.

00:01:00: And so for those of you listening this deep dive Is built exclusively For You!

00:01:04: You Are The Digital Transformation and Tech Professionals Who Are Actually Steering Enterprise Architecture.

00:01:10: They're

00:01:10: the ones in the trenches

00:01:11: Right...you don't need us to explain the basics.

00:01:14: So our mission today is really to just cut through all the noise, you know past the hype cycles and look at The actual architectural reality

00:01:22: which is shifting so fast right now.

00:01:24: It really does.

00:01:25: Today we're looking at this massive shift in quantum computing moving from lab experiments street two industrial assembly lines.

00:01:32: We're gonna cover the immediate financial debt of legacy cryptography.

00:01:35: That's a huge one.

00:01:36: Oh yeah

00:01:37: Plus were getting into why?

00:01:40: Negative day exploits completely changes the operational baseline for cybersecurity.

00:01:45: And finally, how edge architecture is basically actively fighting back against cloud centralization.

00:01:51: So I mean The most logical place for us to start Is probably the hardware layer right?

00:01:56: Specifically this massive shift happening Right now in quantum computing.

00:01:59: Yeah let's Start there

00:02:01: Because if you just look at mainstream media coverage You'd think Quantum this theoretical contest, like a bunch of scientists arguing over who can stack the highest number of raw qubits on a whiteboard.

00:02:12: Right!

00:02:12: The classic Qubit Arms Race

00:02:13: Exactly.

00:02:15: But the operational reality is entirely different.

00:02:17: now...the focus has violently shifted away from theoretical math and straight into industrial manufacturing.

00:02:24: And we are actually seeing this reflected at very high levels in national infrastructure planning Like Zlatko Meinov pointed out the new national roadmap from US Department of Energy recently.

00:02:34: Oh yeah, I saw that!

00:02:36: The metrics they are using have completely changed haven't there?

00:02:38: Totally changed.

00:02:39: They aren't measuring success by Qubit count anymore.

00:02:43: Thank goodness for that.

00:02:44: Right...The mandate now is strictly about delivering scientifically useful machines By twenty-twenty eight.

00:02:50: It's entirely focused on measurable utility Like what can a machine actually do in real

00:02:56: world?

00:02:57: Which means industry has finally acknowledged its actual binding constraint.

00:03:01: Which is?

00:03:03: It's no longer the physics, you know or the algorithms it's the manufacturing capacity.

00:03:08: Ah right building the actual things at scale

00:03:11: exactly and The talent migration happening right now Is a perfect indicator of this.

00:03:16: if you look at top-tier quantum talent like me here best car for example

00:03:21: What did he do?

00:03:22: well He recently left a very traditional Quantum development trajectory to join skywater technology.

00:03:27: okay so as semiconductor foundry

00:03:29: Right.

00:03:30: He joined specifically to lead their quantum foundry solutions, so the absolute top minds aren't just going into theoretical labs anymore to run experiments...

00:03:39: They're moving in to foundries to solve actual supply chain and fabrication constraints

00:03:44: Precisely!

00:03:45: Because I mean shipping wafers that actually work is incredibly difficult when you are dealing with non-standard materials at this kind of microscopic scale.

00:03:54: It's custom fabrication but it has hit a meaningful velocity.

00:03:58: Yeah exactly If the manufacturing base cannot scale up the yield of these specialized, you know superconducting or photonic chips.

00:04:06: The entire quantum industry basically stalls.

00:04:08: it doesn't matter how good the math is.

00:04:10: But on the flip side to that...the math is actively trying bail out hardware constraints right now.

00:04:14: What

00:04:15: do you mean?

00:04:16: Well we are seeing software optimizations drastically reduce burden on physical processors.

00:04:23: Josh Cunningham recently highlighted this massive algorithmic leap from Oak Ridge National Laboratory.

00:04:28: Oh, with the Lugo algorithm?

00:04:30: Right.

00:04:30: Exactly!

00:04:31: The Lugo Algorithm.

00:04:32: They managed to reduce the number of quantum gates needed for Quantum Phase Estimation by more than ninety-five percent.

00:04:39: I mean that is just a staggering optimization

00:04:41: It really is.

00:04:43: they went from needing around two million gates down to just ninety one thousand.

00:04:46: Wow

00:04:47: So we are extracting massively more computational utility out imperfect, noisy hardware that we can currently manufacture today.

00:04:55: And that kind of optimization right there is what bridges the gap between today's noisy intermediate scale quantum era and actual fault-tolerant quantum computing?

00:05:05: But let me push back on the broader implications here for a second

00:05:08: though.

00:05:08: Sure go ahead

00:05:10: If QuantumGateLogic is getting this highly optimized... ...and the computations themselves require so little direct power.

00:05:18: Doesn't quantum computing eventually just solve the massive energy-grade crisis we are seeing right now with AI data centers?

00:05:24: Like, We just trade our GPUs for QPU's.

00:05:28: Yeah I mean it is a really compelling idea on paper but it completely ignores the brutal physics of deployment environment.

00:05:35: Oh!

00:05:35: The cooling...

00:05:35: Right.

00:05:36: Shiraz Nantagma brought up the reality of facility power here.

00:05:39: superconducting quantum processors have to operate at roughly ten to twenty millikelvin

00:05:45: Which is what, just a fraction of degree above absolute zero?

00:05:48: Exactly.

00:05:50: So sure the quantum chip itself might dissipate almost zero power during a calculation but maintaining ten millikelvin that requires massive dilution.

00:05:58: refrigerators extensive microwave control equipment huge cryogenic infrastructure.

00:06:04: Right.

00:06:04: so the facility power required to keep the ambient environment that cold basically wipes out any computational energy efficiency you may gain from the quantum state.

00:06:13: exactly.

00:06:14: So we aren't solving the AI data center power crisis with quantum.

00:06:17: We're just trading an extreme heat dissipation problem for an extreme cryogenic cooling

00:06:22: problem.".

00:06:23: That makes sense, at least with current superconducting architectures.

00:06:27: there really is no free lunch on The Power Grid?

00:06:29: No free lunch at all!

00:06:31: Well if the foundries are actually successfully building these assembly lines for Quantum and the algorithms are getting ninety-five percent more efficient that means the expiration date of our current encryption has moved from a theoretical someday to a very concrete soon.

00:06:46: It really did!

00:06:47: The math is changing, which brings us the second big theme – post-quantum cryptography or PQC?

00:06:54: Right and this exactly where that three billion dollar M&A scenario I mentioned earlier comes into play because PQc isn't future IT roadmap anymore….

00:07:04: No it's happening now...

00:07:04: …it is present day hidden business debt.

00:07:08: Craig A. Linton framed this well around the Harvest Now decrypt later reality

00:07:13: Right, the idea that they're just hoarding our data.

00:07:16: Exactly!

00:07:17: Nation states and advanced persistent threats are actively intercepting and storing encrypted enterprise traffic today.

00:07:23: right now

00:07:24: Because storage is so cheap They can pull down petabytes of encrypted data Park it in a datacenter somewhere And wait for quantum hardware to mature enough To run Shor's algorithm and break keys

00:07:36: Precisely.

00:07:37: So if you are transmitting intellectual property, sovereign data or long-term strategic plans today using standard RSA or ECC algorithms.

00:07:46: You just have to assume it will be plain text in a few years

00:07:50: Which means your crypto agility planning is already late.

00:07:53: Oh very late.

00:07:54: and the readiness data backs that up entirely.

00:07:56: And honestly?

00:07:57: It's pretty bleak.

00:07:59: Steve V shared an assessment of about three hundred fifty organizations globally.

00:08:04: Of the organizations that actually mapped out their quantum migration timelines, yeah.

00:08:08: Twenty-eight percent are already out of time based entirely on their own internal assumptions.

00:08:13: Wow Right?

00:08:14: Their plan to migration takes longer than they believe they have until a crypt analytically relevant quantum computer comes online.

00:08:22: That is terrifying!

00:08:24: And I think their median threat exposure score was sitting at something like seventy four out of one hundred.

00:08:29: Yeah, that sounds right

00:08:30: and you know the reason it takes so incredibly long.

00:08:32: is that migrating to quantum resistant standards?

00:08:34: It's not just some modular software update?

00:08:37: Right wait?

00:08:38: So let me ask Is this just why two K all over again?

00:08:42: a massive tedious IT update where we just swap out the locks expand some field to move on?

00:08:47: I mean, Y-two K was tedious but it was essentially finding date fields in code bases and expanding them.

00:08:53: It was isolated logic.

00:08:55: PQC is totally different.

00:08:57: How so?

00:08:58: Well Nihar S broke down.

00:08:59: why pqc Is fundamentally different...it's a massive tangled dependency problem tied directly into TLS And the underlying network infrastructure.

00:09:09: P Q C algorithms have different key sizes They have different performance signatures and completely different bandwidth requirements.

00:09:16: Oh

00:09:16: i see Which means it breaks legacy microservices.

00:09:19: Yes.

00:09:20: Migrating to PQC isn't like changing the dead bolts on your doors, It's like having to rip out and replace entire plumbing system in a skyscraper while everyone is still working inside because TLS is baked into every API browser load balancer an internal cloud communication channel you have!

00:09:36: IT IS EVERYWHERE.

00:09:38: And that hidden plumbing is now being priced into massive corporate acquisitions

00:09:43: Right Like that M&A scenario.

00:09:45: Yeah Brian See pointed out how cryptographic debt is secretly impacting M&A pricing right now.

00:09:50: He looked at Ioncore, which spent over three billion dollars acquiring companies.

00:09:54: and when you buy a target company...you inherit their entire legacy tech stack.

00:09:59: You inherit they're hard-coded RSA.

00:10:00: twenty forty eight keys And there TLS.

00:10:02: one point two dependencies

00:10:04: exactly And much of that embedded cryptography cannot simply be patched over the air.

00:10:08: It might require total hardware replacement at the edge or, you know rewriting proprietary key management systems from scratch.

00:10:16: That's a huge hidden cost.

00:10:18: it is smart.

00:10:19: private equity firms and acquirers are now treating PQC non-compliance as a priced risk in term sheets.

00:10:26: If you aren't doing technical due diligence on the target's cryptographic agility You're buying a massive compliance and remediation bill at a premium.

00:10:33: Wow

00:10:34: So PQC is basically forcing us to rebuild the absolute foundation of our network security Because the underlying math has fundamentally changing.

00:10:42: Yeah, but you know The urgency isn't just coming from future quantum threats.

00:10:47: Our current security architectures are already collapsing under AI driven attack speeds today.

00:10:52: Speaking of which?

00:10:54: Are traditional cybersecurity models really taking a beating?

00:10:57: The operational baseline has shifted entirely.

00:11:00: We are moving away from traditional perimeter prevention and patching paradigms because the human response loop is simply too slow now.

00:11:08: Before we jump into that, I do want to quickly mention.

00:11:10: if you're finding these insights valuable for navigating your own architectural strategies make sure you subscribe so don't miss our future deep dives!

00:11:18: the

00:11:25: total collapse of the exploitation window.

00:11:28: Juan Pablo Castro shared some sobering data on this,

00:11:39: And we

00:11:55: just saw the operational reality of this with the critical net scaler flaws recently.

00:12:00: Thomas Masacek pointed out that these were being actively exploited before any mitigation existed.

00:12:05: Yeah,

00:12:05: That was a mess.

00:12:06: It was and he made a point that really reframes patch management entirely.

00:12:11: He said applying a patch Just closes the door.

00:12:15: it does absolutely nothing to remove an intruder who has already established persistence inside your network environment.

00:12:21: It's a critical distinction, and honestly one that many compliance frameworks still get totally wrong.

00:12:26: Right because relying purely on patching is zero day in this environment.

00:12:31: it's essentially like putting a heavy-duty deadbolt on your front door after the burglar has already moved their luggage into your guest room.

00:12:37: That's exactly the dynamic which is why threat hunting an evidence preservation have to proceed.

00:12:42: patching now.

00:12:44: So how do we restructure for that?

00:12:46: Well, Sanjeev Charyan suggested that our entire conceptual framework for security has to evolve to match the speed.

00:12:53: He argues at the classic fifty-year old CIA triad.

00:12:56: you know confidentiality integrity availability is just insufficient.

00:12:59: now.

00:13:00: I

00:13:00: mean it's the absolute baseline of InfoSec but It Is Static

00:13:04: Right.

00:13:05: So he proposes this CIA squared tetrad adding a fourth pillar Governed Adaptability.

00:13:11: Governed Adaptability.

00:13:13: What does that look like in practice?

00:13:15: It's the capacity for a network to automatically change its security posture, isolate segments and deploy countermeasures at the speed of an AI threat but operating strictly within human governed boundaries.

00:13:26: That makes a lot sense especially because AI is operating on both sides of the perimeter now And you know securing the enterprise AI infrastructure itself introduces entirely new attack factors.

00:13:38: Oh absolutely

00:13:39: Mohammed Syed actually noted a massive architectural blind spot in how development teams are building RAG systems, you know retrieval augmented generation.

00:13:47: Right this is a major issue as enterprises rush to integrate their internal data with large language models.

00:13:53: exactly the mistake Is relying on the LLM?

00:13:57: To understand permissions.

00:13:59: Syed pointed out that access controls Absolutely must be enforced at the point of retrieval at the vector database level.

00:14:06: The architectural mechanics here are just crucial.

00:14:08: Because think about it, if a user queries the system and the vector search retrieves sensitive HR data or financial projections And then passes that raw text into the LLM's context window with instructions like only show this If the user has admin rights It

00:14:23: is already way too late

00:14:24: Exactly!

00:14:25: The restricted data is in the model memory space for that session.

00:14:29: An attacker can prompt inject their weight around the system prompt to force the model output the sensitive text.

00:14:36: You have to filter the vector search results based on user's IAM role, before the prompt is ever assembled and sent into LLM.

00:14:43: Which

00:14:43: perfectly illustrates how definition of a breach is evolving And why our recovery mechanisms completely change.

00:14:51: Patrick S Conway made a sharp distinction here.

00:14:53: He said Cyber Recovery Is fundamentally different from Standard Disaster Recovery.

00:14:58: Well, yeah.

00:14:58: Because disaster recovery is for linear events right?

00:15:01: A power outage a severed fiber line of flood?

00:15:04: you spin up the backup site You restore the data and your operational

00:15:08: Right but cyber recovery faces an intelligent Adaptable adversary who was actively trying to destroy your backups And maintain persistence.

00:15:18: it cannot be a linear process.

00:15:20: So what's the new process?

00:15:22: Conway defines It as a continuous evidence-driven loop protect Detect, clean and validate.

00:15:31: You can't just auto-restore yesterday's backup because yesterdays' back up likely contains the dormant payload the attacker used to breach you in first place.

00:15:38: Right... You have to clean the immutable data in an isolated quarantine environment before it ever touches production again.

00:15:45: That is incredibly complex.

00:15:48: But y'know To survive these real time intelligent threats and to process the massive low-latency data loads required by Enterprise AI, The physical architecture of the network is basically being forced to adapt.

00:16:01: It has too.

00:16:02: We are seeing a massive repricing in rebalancing of wider ICT infrastructure away from centralized cloud dependency And moving toward extreme edge autonomy.

00:16:13: David Lenticum articulated this perfectly.

00:16:16: He argues that the default posture for edge systems should now be local execution.

00:16:21: Local by default?

00:16:22: Yes!

00:16:22: Routing every edge decision, every sensor read and every basic transaction back to a centralized cloud servers has become one of the most expensive and fragile architectural mistakes in enterprise IT today.

00:16:33: Okay wait

00:16:33: let me play The Devil's Advocate on that first sec.

00:16:36: Go for it.

00:16:37: We spent the last ten years centralizing compute into massive public clouds specifically for visibility centralized security and economies of scale.

00:16:47: If we push processing back down to local edge clusters, aren't we just fracturing our network in making it harder to manage?

00:16:54: We are trading cloud latency for a management

00:16:56: nightmare.".

00:16:58: but you are trading centralized convenience for operational survivability.

00:17:03: Local processing guarantees that your automated warehouse or your manufacturing floor, or hospital systems don't instantly freeze the second.

00:17:10: an upstream internet connection drops...or a cloud region goes down.

00:17:14: Ah I see.

00:17:15: so it's about engineering for graceful degradation rather than catastrophic failure?

00:17:19: Exactly!

00:17:20: The compute power at the edge is no longer restricted to just dumb terminals.

00:17:25: You use the central cloud selectively for heavy asynchronous analytics, global state coordination and model training but the real-time inferencing in operational logic that must run locally.

00:17:35: And as we push more powerful compute to these local and regional data centers... ...the physical constraints of facilities themselves are actually becoming the primary bottleneck.

00:17:44: Oh completely

00:17:46: I mean Shah had pointed out Compute architecture isn't even the limiting factor for scaling AI anymore.

00:17:51: It's physical data quality,

00:17:53: meaning that telemetry of the building itself right?

00:17:55: Right

00:17:56: behind every rack of next generation AI servers is an incredibly fragile ecosystem of power distribution liquid cooling loops and thermal management.

00:18:06: shot argues That facilities in modular data centers must arrive sensor ready and connected from day one to optimize The massive physical systems keeping them running.

00:18:16: Well, because the margins for thermal failure are so incredibly narrow now.

00:18:19: Yeah If a cooling loop degrades you don't have hours to fix it.

00:18:23: You have minutes before you see total hardware destruction.

00:18:26: Exactly And while the hardware and physical infrastructure are decentralizing to the edge We aren't seeing an equally massive disruption in the software infrastructure layer that manages all this compute.

00:18:38: Raza Rahman highlighted The tectonic shift happening In the server virtualization market right Now.

00:18:44: Oh man, this is arguably one of the most disruptive operational events in enterprise IT... ...this entire decade.

00:18:52: The Broadcom situation?

00:18:53: Yes!

00:18:53: Following Broadcom's acquisition of VMware….

00:18:56: …the subsequent licensing changes and transition to subscription models have triggered an unprecedented market repricing.

00:19:03: It really has forced every Enterprise architect.

00:19:05: look closely at their hypervisor dependency.

00:19:07: Forced as a right word

00:19:09: When a single vendor controls the virtualization layer that your entire on-premise and hybrid cloud environment runs on, And they just change their commercial terms overnight.

00:19:18: You realize how deep that vendor lock in really goes.

00:19:21: And the friction here is immense!

00:19:24: Migrating thousands of virtual machines off ESXi isn't just a weekend project you know?

00:19:29: Not at all.

00:19:30: It impacts your backup architectures Your disaster recovery runbooks Your storage integrations Your network micro segmentation Everything.

00:19:37: But despite all that friction, Riemann notes a massive surge of enterprise interest in open-source alternatives.

00:19:43: Things like KVM based hypervisors, Proxmox and Open Hyperconverged Infrastructure.

00:19:49: Enterprises are doing the hard math.

00:19:51: right now They're deciding.

00:19:53: taking short term operational pain of a hypervisor migration is strategically safer than remaining entirely dependent on single commercial virtualization stack.

00:20:02: It's total reevaluation infrastructure sovereignty

00:20:06: which actually perfectly ties together everything we've analyzed today.

00:20:09: It really does!

00:20:10: Think about it, We have looked at the industrialization of quantum foundries The immediate threat of cryptographic debt The AI driven collapse Of the patching life cycle And this architectural pivot toward edge autonomy and infrastructure diversification...

00:20:25: ...It is a massive amount of systemic change happening simultaneously.

00:20:30: I mean how should tech professionals listening to this internalize all these moving parts?

00:20:35: I want to leave you with a concept inspired by Ingrid Vassilio-Feltz because it's really easy to look at these disciplines in isolation, right?

00:20:43: Yeah.

00:20:43: It's separate silos.

00:20:45: Exactly!

00:20:45: You have your quantum engineers over here network architects over there cyber security analysts cloud infrastructure teams.

00:20:53: but the ultimate competitive advantage over the next decade won't belong to The Enterprise with the absolute largest data center or the one.

00:21:03: So it's not about owning the individual best-in class components?

00:21:07: No.

00:21:08: It is about orchestration.

00:21:09: The true market advantage will belong to organizations that can finance, integrate and orchestrate hybrid compute Autonomous edge connectivity And quantum safe security architectures as one single resilient and sustainable ecosystem.

00:21:22: Wow!

00:21:23: The silos are completely collapsing.

00:21:25: Exactly For the digital transformation leaders and enterprise architects listening right now, your long-term value isn't just in mastering one specific hypervisor migration or learning one cybersecurity framework.

00:21:37: Your real value lies on how you design the intersections between them.

00:21:41: How do you map your network dependencies to survive a post quantum transition while simultaneously pushing your AI inferencing securely to the edge?

00:21:49: That is the strategic high ground

00:21:51: It's phenomenal perspective.

00:21:53: Are you updating isolated components Or are you architecting an integrated, survivable system?

00:22:00: That's the question.

00:22:01: You have to ask yourself

00:22:02: if you enjoyed this episode.

00:22:03: new episodes drop every two weeks.

00:22:05: Also check out our other editions on cloud insights and sovereignty digital products in services AI and agentic systems green ICT and sustainable AI defense tech and health tech.