Best of LinkedIn: Defense Tech CW 10/ 11

00:00:00: provided by Thomas Allgaier and Frennis, based on the most relevant posts on LinkedIn about defense tech in CW ten-and-eleven.

00:00:07: Frennus is a B to D market research company that equips product and strategy teams with market, competitive intelligence across the defense industry.

00:00:17: So welcome to The Deep Dive!

00:00:18: And right now literally as you listen to this adversaries are functionally choking off about twenty percent of world's crude oil shipments.

00:00:26: Yeah which is massive.

00:00:27: It is...and crazy part they aren't doing it with some massive peer level navy They're using drone boats.

00:00:34: Boats cost less than like a mid-range luxury car.

00:00:37: Exactly!

00:00:38: So if you're operating anywhere in the defense tech space, whether your coding on the engineering floor or your writing policy are making those big procurement decisions in the boardroom The economic math of your world just completely broke.

00:00:49: It really did.

00:00:51: Today we're breaking down the absolute top trends and You know technical realities that your peers are actively debating across LinkedIn over the past two weeks

00:00:59: Yeah because were looking at an environment where That transition from theoretical white papers to Actual operational reality has just drastically accelerated.

00:01:09: big

00:01:09: time,

00:01:10: right?

00:01:11: So our focus today is to really synthesize the mechanisms behind these shifts.

00:01:15: We're gonna look closely at well The asymmetric threat of autonomous swarms The profound integration challenges of AI on the edge which

00:01:24: is a huge topic right now

00:01:26: it is And we'll also get into the operational validation of collaborative combat and very real industrial supply chain pivoting.

00:01:34: that's happening right now in Europe.

00:01:35: Okay, let's start with that naval disruption because I think it perfectly illustrates A lot of analysis from professionals like Mark K, John J H and Richard William over the past two weeks regarding The Street Of Hormuz.

00:01:48: Yeah that's been a massive talking point right?

00:01:50: And they all point to the rise of cheap scalable uncrewed surface vehicles or USVs.

00:01:56: You know these highly autonomous drone boats...the

00:01:58: mechanism driving this disruption is really fundamentally about the cost exchange ratio.

00:02:03: Okay

00:02:03: break that down for us.

00:02:04: well

00:02:05: adversaries are mass producing these autonomous surface threats for like under one hundred thousand dollars per unit.

00:02:12: Wow!

00:02:13: Yeah, and when an attacker launches a swarm of these low-profile vehicles alongside synchronized missile strikes I mean they aren't trying to outgun a Western destroyer.

00:02:23: right the kid exactly there trying to overwhelm its radar tracking capabilities just exhausts it's physical magazine depth.

00:02:31: Richard Brilliam actually crystallized this dynamic by pointing out that when economics flip like this, cheap mass beats exquisite capability.

00:02:39: Cheap

00:02:39: Mass Beats Exquisite Capability?

00:02:42: That's a great way to put it!

00:02:43: Right

00:02:43: you create the scenario where a multi-million dollar defense asset is sidelined by a fleet of budget built fiberglass boats.

00:02:50: I'll push back on this cheap mass beets exquisite capabilities idea though because i mean a fifty thousand dollars drone boat inherently has terrible range, right?

00:02:59: A limited explosive payload and it usually relies on easily jammed commercial comms.

00:03:04: So are these actually strategic threats to a blue water navy that's highly capable of electronic warfare or they just tactical nuisances getting way too much PR?

00:03:15: No I mean the elevate to a strategic threat precisely because Exactly.

00:03:23: A single interceptor missile fired from a destroyer can cost upwards of two million dollars, right?

00:03:29: So even if your electronic warfare systems jam say half the swarm If you have to fire three physical interceptors To stop their remaining boats before they hit your hull below The waterline well You've just spent six million dollars to defeat a threat that costs the adversary Three hundred thousand to deploy.

00:03:46: That is brutal math

00:03:47: it Is.

00:03:48: do that every day for a month and the Navy runs out of both interceptors And budget, The math completely breaks

00:03:53: which you know explains that fascinating strategic reversal Tim DeZitter has been documenting recently.

00:03:58: Oh

00:03:58: the Pentagon procurement.

00:04:00: Yes!

00:04:00: The Pentagon is reportedly looking To purchase F-Ten attack drones directly from Ukrainian brand fDrones as part Of their drone dominance program.

00:04:09: Right.

00:04:09: I mean, historically the narrative has always been that Ukraine depends on Western military technology and now western militaries are actively procuring systems that emerge from the Ukrainian ecosystem.

00:04:22: Well it comes down to iteration cycles.

00:04:24: really i mean a traditional western defense prime might take what ten years?

00:04:29: To move a platform from requirements to deployment.

00:04:31: at least ten years

00:04:32: right.

00:04:33: but in ukraine because these systems Are being tested under live fire daily The iteration cycle is measured in weeks.

00:04:40: Weeks?

00:04:41: That's incredible!

00:04:42: It is, Oleg Ryjinski highlighted this when announcing U-Force' mission.

00:04:46: They're deploying combat proven systems specifically the Magura USVs that have been successfully sinking warships along with interceptor drones and their stated goal to make defense a hundred times cheaper than offense.

00:04:59: So by adopting these rapid iteration systems, Western militaries are basically trying to instantly inherit that compressed learning curve.

00:05:06: Okay so if swarms of cheap drones are breaking the economic math of warfare how do you actually fix a broken cost curve?

00:05:12: You have to change how you intercept them Exactly!

00:05:16: Aviv Barzohar made the point that counter UAS or CUAS is no longer an optional add-on right.

00:05:22: it has to be organic to every single unit.

00:05:25: Yeah absolutely.

00:05:25: And we're seeing hardware evolve.

00:05:28: Felipe Albinos shared some really interesting testing data on American Rheinmetall's Autonomous UaSharad solution.

00:05:35: Oh, the short range air defense?

00:05:36: Yeah!

00:05:37: They're taking ultra-short range air defence so radar and kinetic interceptors and integrating it directly onto GM defences tactical vehicles

00:05:45: which is I mean integrating uaShrad into tactical vehicles as a critical stopgap.

00:05:49: So infantry units aren't just sitting ducks for first person view.

00:05:53: drones right but Kinetic interceptors still have a magazine limit.

00:05:57: Right,

00:05:57: you still run out of bullets!

00:05:58: Exactly to truly invert that cost exchange ratio the industry has to move to directed energy.

00:06:05: Craig Gustafson highlighted the maturity of BA systems work in this area.

00:06:08: We are moving way past the science experiment phase of lasers.

00:06:11: Just be clear on

00:06:12: mechanics here for everyone listening.

00:06:14: we aren't talking about like shooting a burst of electricity.

00:06:17: We're talking about directing concentrated photons to physically melt the components or fry the optical sensors in coming drone, right?

00:06:26: Correct yeah!

00:06:27: The mechanism is thermal transfer.

00:06:29: you are focusing light to rapidly heat critical node on target.

00:06:33: Wow.

00:06:34: And the operational advantage is that as long as the vehicle's engine running and generating power, your magazine is virtually bottomless.

00:06:41: Because it just generates electricity?

00:06:43: Right!

00:06:44: The cost per shot drops from millions of dollars to just a few liters diesel fuel.

00:06:50: Directed energy solves the exact magazine depth problem that swarms are designed to exploit.

00:06:54: But you know, operating lasers managing swarms filtering on complex radar data.

00:07:00: it requires an immense amount of processing power.

00:07:02: Oh!

00:07:02: An unbelievable amount.

00:07:03: Right You cannot separate physical hardware from defense tech the software layer running it.

00:07:09: And right now, the integration of artificial intelligence is sparking this massive highly visible geopolitical debate regarding

00:07:16: governance.".

00:07:21: A

00:07:24: prime example of this friction was brought up by Daniel Kirch.

00:07:27: He analyzed the recent tensions surrounding Anthropic.

00:07:30: Ah, yes...the Palantir situation?

00:07:32: Yes!

00:07:33: So reports indicated that Palantirs' platform was utilized to deploy Anthropics AI model, Claude during a classified military operation captured Venezuela's President Maduro.

00:07:45: And just to be absolutely clear, we are neutrally reporting the contours of this debate here.

00:07:51: We're not taking a stance on ethics but the underlying conflict is that U.S.

00:07:55: Department of Defense is pushing to utilize these frontier models for operational planning While Anthropic is pushing back, enforcing their terms of service which strictly prohibit the use of models for fully autonomous weapons and domestic surveillance.

00:08:12: Right in technical friction here's really about API access environment controls.

00:08:16: How

00:08:16: so?

00:08:17: Well a defense agency cannot send classified operational data out to commercial cloud server to be processed by an AI model.

00:08:26: they just can't!

00:08:27: Major security risk.

00:08:28: Exactly, they need the model localized on their own secure servers.

00:08:33: But the moment a commercial AI company hands over the weights of their model to operate in an air-gapped classified environment... They

00:08:40: lose visibility!

00:08:42: ...exactly, they lose the ability to monitor how that model is being used.

00:08:47: Stephen Carrell laid out a brilliant historical parallel to explain this.

00:08:50: actually he argued This isn't fundamentally debate about compute.

00:08:54: it's a debate about control.

00:08:56: We navigated exact same tension during the early days of high-resolution commercial satellite imagery.

00:09:02: Oh, you mean this shutter control policies?

00:09:04: Yes

00:09:04: exactly yeah.

00:09:05: when companies like Digital Globe began launching submeter resolution satellites The capability existed and the commercial market demanded it.

00:09:12: right but the US government recognized the national security implications And they retain shatter control.

00:09:18: So legal authority to limit resolution or restrict imaging of certain locations.

00:09:23: that makes a lot of sense.

00:09:24: It took years to negotiate the governance, the tasking rights and data caps.

00:09:29: The current standoff over AI access in classified environments is that exact same government struggle just translated from orbital mechanics into software architecture.

00:09:38: That's great analogy!

00:09:45: You know, understanding that historical context makes the current landscape much clearer.

00:09:49: And it also highlights a strategic window for other markets.

00:09:53: Nassimiliano Maggi put forward a really interesting presentation on this front.

00:09:57: What did he say?

00:09:58: He argued that Europe should formally invite Anthropic to establish our primary anchor there.

00:10:03: Oh wow!

00:10:04: Yeah His rationale is that Europe already has this thriving ecosystem of defense tech founders who are explicitly building AI to protect liberal democracies.

00:10:14: He pointed to companies like Helsing whose foundational philosophy frames the protection of democratic societies as a civic duty deliberately choosing To only supply Democratic nations.

00:10:24: so Maggie's point is that Europe has this sovereign funding and The specific defense tech ecosystem, but it currently lacks.

00:10:31: A frontier foundational AI model to power it.

00:10:35: I mean It is an aggressive industrial proposal but acquiring or hosting a frontier AI model is really only step one.

00:10:42: True The industry runs the risk of assuming that simply possessing advanced AI translates automatically to operational superiority.

00:10:50: Yeah Eva Sula grounded this conversation with a critical reality check, the bottleneck in defense AI Is not procurement it's utilization and validation.

00:11:01: yeah if you look at the civilian sector we still struggle To make self-driving cars operate With perfect reliability on clearly marked quiet suburban streets, taking an AI model and deploying it into the absolute chaos of a combat zone where there are no lane markers.

00:11:15: Where adversaries or actively spoofing your sensors in were GPS is heavily jammed that as an astronomical technical leap.

00:11:21: you don't just download an AI update to the battlefield.

00:11:24: No You Don't.

00:11:25: The environment Is Actively Adversarial.

00:11:28: This Requires A Profound Shift In Operational Doctrine.

00:11:32: I mean if An AI System Flags A Target Through A Sensor Feed Who in the command structure is validating that data?

00:11:38: Right, who pulls the trigger.

00:11:39: Exactly!

00:11:40: If an automated targeting recommendation leads to a collateral incident how do they?

00:11:44: existing military accountability framework supply?

00:11:48: Introducing probabilistic software into deterministic military hierarchies doesn't automatically solve problems.

00:11:55: without rigorous integration pipelines it merely introduces unpredictable risk vectors.

00:12:00: And, you know that software requires specific hardware to function locally.

00:12:04: ArrayCanNli raised this critical issue of edge compute sovereignty.

00:12:08: This is huge.

00:12:09: Yeah if a European defense platform's onboard AI relies entirely on microchips and computing architecture from single trillion dollar overseas vendor Is that platform truly sovereign?

00:12:19: No

00:12:19: Right, we focus heavily on the cyber models but locally developed AI accelerators.

00:12:23: The physical hardware that processes data at the edge without needing to ping a distant cloud server those are massive vulnerabilities in the supply chain

00:12:31: which forces the industry to confront how they actually validate these integrated systems.

00:12:36: because localized compute sensor fusion and AI models or so insanely complex to align The entire defense sector is shifting toward collaborative combat platforms that can be validated under intense operational stress.

00:12:50: Yeah, Ismail Alvarez captured the shift perfectly he noted.

00:12:54: as we move through twenty-twenty five defining metric of success.

00:12:57: no longer who has most impressive technology functioning in a sterile lab environment?

00:13:02: Lab tests don't win contracts anymore

00:13:04: Exactly!

00:13:04: The Metric is Who Can Validate, Iterate and Adopt These Systems Fastest Under Real World Constraints.

00:13:11: R&D funding generates press releases, but operational validation generates adoption.

00:13:15: And we're seeing unprecedented levels of that validation right now specifically in the realm man-on-man teaming or MUMT?

00:13:22: Oh!

00:13:23: Voon and Grian shared.

00:13:24: an update perfectly demonstrates this.

00:13:26: Just four months after its initial flight The YFQ-IVA autonomous fighter jet managed to swap between two completely different mission autonomy stacks while it was midflight.

00:13:35: Yes We are talking about the aircraft dynamically switching from Enduro's lattice system to SHIELD AI's hive mind system in air.

00:13:45: To understand the magnitude of that, we really have look at mechanism.

00:13:49: they achieve this using a modular open systems architecture.

00:13:53: Right, switching from Endural to SHIELD AI mid-flight isn't just like swapping from a Mac operating system to Windows on your laptop.

00:14:00: it's like swopping out the entire central nervous system of a predator while its actively hunting without missing a step.

00:14:07: That

00:14:07: is great way to look at it.

00:14:08: It proves that a Modular Open Systems Architecture Isn't Just A Buzzword and Procurement Document Anymore.

00:14:15: Its Finally Functioning Code.

00:14:17: It means a military can swap out the artificial intelligence driving its jets, depending on specific mission parameters without having to rebuild entire aircraft.

00:14:25: And this modular approach is global directive.

00:14:29: Luca Leon reported that Airbus is currently integrating their European Mars Mission System which actually includes an advanced AI layer known as Mindshare directly onto US-built Kratos Valkyrie drones.

00:14:40: Oh so crossing borders with tech

00:14:42: Exactly.

00:14:42: By ensuring European software can pilot American built, a tradable hardware they are building the interoperability required to deploy these collaborative combat drones alongside manned Eurofighters by twenty-twenty nine.

00:14:55: and we're also seeing this validation happen in some of the most unforgiving environments.

00:14:59: on And Elliott Pence discussed Canada's Dominion dynamics, which just secured fifty million dollars to develop a sovereign autonomous loyal wingman aircraft.

00:15:11: Right the Arctic one?

00:15:12: Yeah they aren't just copying an existing US or European drone.

00:15:16: They have specifically ruggedized the platform To operate in negative sixty degrees Celsius arctic conditions

00:15:22: That is insanely cold.

00:15:23: At those temperatures standard battery chemistry fail hydraulic actuators freeze and sensor optics warp their engineering autonomy for zero infrastructure, extreme weather environments.

00:15:33: That environmental adaptability is so crucial.

00:15:36: and this push for autonomy is equally urgent.

00:15:38: on the ground The validation of uncrewed ground vehicles or UGVs accelerating rapidly.

00:15:44: Yeah, Doug Livermore highlighted some intense operational footage from Ukraine on this.

00:15:49: The UGV platoon ruptor operating within the Third Assault Brigade is deploying ground robots to carry explosive charges directly into heavily fortified adversarial positions.

00:15:59: Wow they're utilizing uncrewed systems for direct kinetic attacks.

00:16:03: by sending a machine to breach of bunker They are entirely avoiding the devastating human cost of a frontal infantry assault.

00:16:11: To scale that kind of ground autonomy beyond bespoke tactical units you really need modular hardware solutions.

00:16:17: Florian Seibel recently introduced the mosaic ground autonomy kit.

00:16:22: which does that do?

00:16:23: This system is designed to transform existing manned Deanler transport trucks into unmanned logistics vehicles

00:16:30: And the here's what matters for this audience, right?

00:16:33: They aren't rebuilding.

00:16:33: The truck from scratch.

00:16:34: they were applying a drive by wire autonomy kit that physically actuates the pedals and steering column.

00:16:40: exactly runs on a decentralized edge compute module so the truck can navigate hostile unmapped terrain without requiring a continuous easily jammed satellite uplink.

00:16:50: precisely it allows For high-risk logistics runs Without risking human drivers.

00:16:56: yeah similarly Paul Bauer and Cecilia Peruba highlighted the engineering behind AIREX Robotics Hector Platform.

00:17:04: Oh, The Hector platform right?

00:17:06: Yeah this system is intentionally designed so that human crews can switch between manned and unmanned operations in the middle of a mission.

00:17:13: If the operational environment degrades or if vehicle enters high threat zone then crew can dismount and operate it remotely.

00:17:20: It's engineered for graceful degradation under pressure.

00:17:23: But every single one these breakthroughs You know, the modular jet brains that autonomous logistics trucks.

00:17:30: The directed energy weapons all hits a wall if the industrial base cannot actually manufacture them at scale or one

00:17:36: hundred percent

00:17:36: which leads us to the fundamental restructuring of the European Defense Industrial Base.

00:17:41: it is arguably the most significant industrial pivot we have witnessed in decades.

00:17:45: .The overarching strategy has shifted away from designing bespoke theoretical capabilities towards scaling mass production and reducing supply chain bottlenecks.

00:17:54: Jonas Singer provided a stark breakdown of five concrete improvements in Europe since twenty-twenty two.

00:18:00: First, ammunition production is aggressively scaling moving from a baseline three hundred thousand artillery shells annually to a hard target if two million

00:18:10: massive jump.

00:18:11: second missile and air defense manufacturing capacity as physically expanding across the continent.

00:18:17: third Procurement cycles that used to take ten-to fifteen years are being compressed into months, specifically for atortable drones and EW or electronic warfare systems.

00:18:28: To add context to that third point, you know procuring an electronic warfare system in months instead of decades means a military can actually field jammers That respond to the specific frequencies and adversaries using today rather than deploying a system designed for The threat landscape at twenty fourteen

00:18:43: exactly.

00:18:44: his fourth Point is that drone warfare?

00:18:46: Is finally being institutionalized into standard doctrine not just treated as a special operations experiment.

00:18:52: And fifth The defense startup ecosystem has broken through legacy procurement barriers, bringing venture back speed to military engineering.

00:18:58: And this industrial momentum is really enabling the creation of fully sovereign integrated networks.

00:19:05: Harold Cup had made an incredibly sharp observation regarding us.

00:19:08: what did he notice?

00:19:09: Well while much of that political tension is focused on massive next generation fighter programs like SCAS and GCP which are realistically fifteen years away from deployment Europe is quietly assembling a functioning autonomous combat network today.

00:19:25: Yeah,

00:19:25: he detailed how they're layering Helsing's Centaur AI operating system Hensseltys MDO core sensor fusion architecture and Schwartz digits sovereign defense cloud.

00:19:35: All of this runs through the CA-I Europa Autonomous Combat Node.

00:19:39: This isn't a future concept, it's a fully European highly classified quantum resistant architecture that operates independently any American hyperscaler cloud providers.

00:19:48: It uses the manned Eurofighter as command node The CA- I as loyal wingman swarm And the Centaur AI processing mind.

00:19:56: They are

00:19:56: actively stitching together software and hardware for sovereign kill chain.

00:20:00: But we have to look objectively at the friction points still throttling progress.

00:20:05: Absolutely.

00:20:06: If you're in the strategy boardroom listening to this, Jean-Braise Dumont's recent point at Bedeck is likely your daily headache.

00:20:13: He argued that the industry still being held back by a culture of zero risk procurement.

00:20:18: Yeah The pilot program track

00:20:19: Exactly.

00:20:20: You simply cannot scale a manufacturing production line or hire hundreds specialized engineers based on a twelve month Zero Risk Pilot Program.

00:20:30: If governments want true industrial readiness and surge capacity, they have to signal continuous multi-year demand through scalable contracts.

00:20:38: Industrial capacity requires financial predictability.

00:20:41: It's that simple.

00:20:43: Furthermore regulatory bottlenecks remain a severe threat to innovation.

00:20:47: Thomas Kineshny provided us stark warning regarding inefficient export controls specifically within Germany.

00:20:53: The timeline gap is staggering there.

00:20:56: If it takes a German defense startup five months to secure an export license for dual-use AI technology, while a competitor in a neighboring EU state can secure the same license in three weeks that German company burns through its venture runway just waiting on paperwork.

00:21:08: Exactly.

00:21:09: Regulatory inefficiency directly costs nations the AI race

00:21:13: and this administrative friction heavily impacts how defense consortiums operate.

00:21:17: Mike Richardson recently published a highly rigorous scoring framework that evaluated the actual value and output of European Defense Fund, or EDF consortiums.

00:21:26: He analyzed friction return on investment in national defense baselines And the data heavily disrupted conventional industry assumptions.

00:21:35: Yeah Poland completely dominated this scoring.

00:21:38: they did.

00:21:38: Poland ranked at the absolute top of the framework in terms of efficiency and output.

00:21:43: Meanwhile, the traditional heavyweights of European defense Germany and France scored primarily as credibility anchors.

00:21:50: Credibility anchors right?

00:21:52: Yeah their presence into consortium lends political weight but they are not functioning.

00:21:56: The data clearly shows that institutional prestige and political convenience do not automatically translate into funded, rapidly deployable capabilities.

00:22:07: Execution speed is the new industrial currency.

00:22:09: It's a completely reorganized landscape And understanding that speed in integration leads us to a final underlying reality That everyone operating in this space has to confront.

00:22:24: We recently saw Simon Proud highlight that a purely civilian weather satellite, Meteosat-II captured the thermal explosion of Russian LNG carrier in the Mediterranean.

00:22:34: High fidelity space observation is now inherently dual use whether it belongs to military or meteorological institute And Glenn

00:22:41: McCartan pointed out a recent incident where commercial AWS data centers in the Middle East were actively targeted by drone strikes.

00:22:48: Because these commercial server farms process vast amounts of economic and logistical data, they sit at the precise intersection of political and economic pressure.

00:22:57: They are no longer just civilian tech infrastructure—they're primary military targets!

00:23:01: Which leaves us with fundamental provocation for anyone designing technology today... If a civilian weather satellite provides actionable military intelligence and a commercial cloud data center is treated as high-value kinetic target, the operational boundaries have vanished.

00:23:17: As a defense professional, how do you architect networks design physical security and write governance for a world where the boundary between civilian infrastructure in military target simply no longer exists?

00:23:28: That is the complex reality that industry has to build.

00:23:31: If you enjoyed this episode new episodes drop every two weeks.

00:23:34: also check out our other editions on ICT and tech insights health tech cloud, digital products and services artificial intelligence in sustainability.

00:23:42: And green ICT.

00:23:44: thank you for joining us today.

00:23:46: stay curious make sure your subscribed.

00:23:47: we will catch on the next one.