Best of LinkedIn: Defense Tech CW 48/ 49

00:00:00: Provided by Thomas Allgaier and Frennus, based on the most relevant posts on LinkedIn about defense tech in CW-Forty-Eight and Forty-Nine, Frennus is a B-to-B market research company that equips product and strategy teams with market and competitive intelligence across the defense industry.

00:00:17: Welcome back to the deep dive.

00:00:19: Today, we're doing a rapid scan of the top defense tech trends we've seen bubbling up on LinkedIn over the last two weeks.

00:00:25: That's calendar weeks, forty-eight and forty-nine.

00:00:28: Exactly.

00:00:29: We've tried to cut through the noise to find the real signal and what we're seeing is this steady critical momentum in a few key areas.

00:00:37: For

00:00:37: sure.

00:00:37: I mean, we're talking autonomy, sensors, command and control, and AI.

00:00:41: And especially how these things are actually being debated and, you know, implemented within the European defense ecosystem, which is fascinating right now.

00:00:47: It

00:00:47: is.

00:00:47: And this isn't about white papers.

00:00:49: We're looking at what people who are actually in the industry are sharing.

00:00:52: And we saw what four major themes kind of crystallize.

00:00:56: You've got the race towards truly autonomous systems, the absolute need for layered counter UAS stacks, and then this sort of sudden arrival of generative AI and strategic planning.

00:01:08: And the last one, which might be the most important.

00:01:10: Oh, absolutely.

00:01:10: The growing urgency to fix how we buy and field all this new tech.

00:01:15: That structural piece is huge.

00:01:16: But, okay, let's start with the hardware and the code.

00:01:19: Our first theme.

00:01:20: Uncrewed systems and the autonomy horizon.

00:01:23: Right.

00:01:24: The big story here isn't just that drones exist.

00:01:27: It's how little space is left between a human supervising the operation and... Well, full autonomous decision-making.

00:01:35: And this is where it gets really serious.

00:01:37: What stood out, and this is noted by Aviv Barzohar, is how close we are to crossing a major threshold with international humanitarian law.

00:01:44: What do you mean?

00:01:45: Well, frontline drones are already getting AI-generated strike recommendations.

00:01:49: The operator is there, but they're just giving the final OK.

00:01:52: So it's less of a decision and more of a final confirmation, a checkbox.

00:01:56: A checkbox.

00:01:57: Exactly.

00:01:58: It's a tiny but absolutely critical step away from lethal autonomous weapons making that final call all by themselves.

00:02:05: The

00:02:05: legal and ethical side of that is just enormous.

00:02:08: It's huge.

00:02:09: And as these systems get more powerful, the question of accountability, if an AI gets it wrong, I mean, fratricide, civilian casualties, that gets harder to answer, not easier.

00:02:20: Okay, that's the ethical complexity.

00:02:22: Let's... pivot from that risk to the reality of what's being fielded.

00:02:26: We saw some really successful experiments linking unmanned systems across the entire kill chain.

00:02:31: You

00:02:31: must be thinking of the Harakas Storm exercise in Kenya.

00:02:33: That's the one.

00:02:34: Yeah, a textbook case of integrating tech quickly.

00:02:37: Joseph Trulove and George Huckle detailed how it all worked.

00:02:40: You had the ARX Robotics Churian UGV.

00:02:43: The ground vehicle.

00:02:44: Right,

00:02:44: the UGV.

00:02:45: It detected and classified a target, an armored vehicle.

00:02:48: That data moves super fast HQ.

00:02:50: and then an HX-II delivered the effect.

00:02:52: So what does a demo like that actually prove strategically?

00:02:55: Is it just a cool test?

00:02:56: No, it proves that really short decision cycles are achievable.

00:03:00: I mean, the time from spotting something to hitting it used to be minutes, sometimes hours.

00:03:05: Now we're compressing that.

00:03:07: And that speed is the new measure of combat effectiveness.

00:03:10: It is, and they demonstrated it in the field.

00:03:12: Speaking of fielding advanced stuff, Enduro and EdG Group launched the Omen VTOL AAV.

00:03:20: Torst & Stobb flagged this and it feels like a big shift in platform size and capability.

00:03:26: The Omen is a really interesting aircraft.

00:03:28: It's a hybrid electric tail-sitting platform.

00:03:30: So vertical takeoff and landing, no runway needed.

00:03:33: Exactly.

00:03:34: But the key thing here is that it has the payload and the range you'd normally see on a much bigger group four or five drone.

00:03:41: But with the flexibility of a smaller platform.

00:03:43: That's the game changer.

00:03:44: It shifts you away from relying on big, vulnerable air bases.

00:03:47: It's all about that operational flexibility.

00:03:49: And Enduro seems to be betting everything on just accumulating massive amounts of operational data.

00:03:55: Jared Wilson pointed out something that's just mind-boggling.

00:03:58: Their autonomous undersea vehicles have clocked over forty thousand one hundred kilometers of mission time.

00:04:03: That's

00:04:03: more than the circumference of the earth.

00:04:05: It is.

00:04:06: And if you're in this industry, that fact should define your perspective.

00:04:09: You cannot simulate that kind of real world complexity.

00:04:13: The live data from actual ocean currents, salinity, noise, it's just.

00:04:19: So much better for training your AI

00:04:21: models.

00:04:21: That kind of operational experience becomes a competitive moat that legacy firms just can't match.

00:04:26: Yep,

00:04:27: and that leads us perfectly into our second and honestly deeply frustrating theme.

00:04:32: Procurement, iteration, and the industrial base.

00:04:35: Because all this amazing tech means absolutely nothing if the procurement system just stops it cold.

00:04:40: You said it.

00:04:41: You can have the best AI on the planet, but if your acquisition process was built for you are failing your war fighters.

00:04:47: And the evidence for this is just dramatic.

00:04:49: Artem Rose cited it.

00:04:51: The U.S.

00:04:51: Army Secretary's statement.

00:04:53: Yeah.

00:04:54: Saying contractors have, and I'm quoting, conned the U.S.

00:04:56: military into these insanely expensive buys.

00:04:59: The example was that helicopter knob.

00:05:03: For a

00:05:03: part you could buy for fifteen.

00:05:04: It's almost hard to believe.

00:05:06: It is, but it keeps happening.

00:05:07: And like Morose said, that's not a supply chain issue.

00:05:10: It's a failure of the procurement architecture itself.

00:05:13: It rewards complexity and vendor lock-in, not performance.

00:05:18: Exactly.

00:05:19: Which is why reform isn't just a nice idea anymore.

00:05:21: It's an operational necessity.

00:05:23: So if that's the broken model, what does a functional one look like?

00:05:27: It

00:05:27: looks market-driven.

00:05:28: It looks performance-based.

00:05:30: Look at what's happening with Ukraine's Bravy One Market.

00:05:33: It's

00:05:33: been called an Amazon for the military.

00:05:35: That's a great way to put it.

00:05:36: It forces providers to compete on pure performance, capabilities, pricing, reviews.

00:05:42: It's all visible.

00:05:43: The soldiers on the front line choose what actually works.

00:05:46: And the system rewards battlefield results, not who has the best lobbyists.

00:05:50: It completely flips the incentive structure.

00:05:52: You have to.

00:05:52: optimized for speed and capability, not just margin on a slow contract.

00:05:57: But that requires a whole cultural shift, especially in how we view development and testing.

00:06:01: Oh, absolutely.

00:06:02: Because when a company like Andorrol runs rapid tests, the media often jumps on any failures.

00:06:08: I've seen that.

00:06:09: And it totally misses the point of iteration.

00:06:12: It does.

00:06:13: Madeline Hart and Matt Beb both stress this.

00:06:16: Failing a lot isn't a bug.

00:06:18: It's a feature of iterative development.

00:06:20: You have to break things to make them better.

00:06:22: Of course.

00:06:23: That's how you get reliable capability.

00:06:25: You find the bugs, fix them, ship again, and repeat until the thing is solid.

00:06:29: It's

00:06:29: about compressing that time to relevance, right?

00:06:31: The difference between a ten-year cycle and a ten-month one.

00:06:34: Exactly.

00:06:35: So once you have fast tech and a better procurement system, what's left?

00:06:40: Matt McCran argued that once you clear a certain tech threshold, it all comes down to leadership.

00:06:44: Interesting.

00:06:45: His point is that the companies that really build momentum are the ones obsessed with the customer and aligned around the mission, not just the next contract.

00:06:53: Cool tech alone is not a sustainable advantage.

00:06:56: And that focus on execution is really gaining steam in Europe.

00:07:00: Florian Seibel highlighted the recent valuation uplift for quantum systems.

00:07:04: Which is a huge signal of investor confidence in European dual-use scaling.

00:07:09: The bet is on agile private companies.

00:07:12: And that feeds into a broader strategic vision for Mark A. Whitefeld.

00:07:16: He's arguing Europe has to connect its military innovation with its existing industrial base.

00:07:23: Yes.

00:07:24: Think about it.

00:07:25: How can an automotive supply chain adapt to produce drone parts faster than a traditional aerospace firm.

00:07:31: By using the scale and precision manufacturing they already

00:07:34: have.

00:07:35: That's the connection Europe needs to make for sustainable deterrence.

00:07:38: Okay, so let's try to connect all these pieces.

00:07:40: Theme three is C-II networks and layer defense.

00:07:45: This is where it all comes together.

00:07:47: And where you have to deal with a sheer number of threats like drones.

00:07:50: Kirill Jokimovich had a great take on this.

00:07:52: He argues modern war is a sensor war, not a drone

00:07:56: war.

00:07:56: Yeah, the old idea of a continuous front line is just gone.

00:07:59: It's replaced by these dense, interconnected kill zones.

00:08:02: So

00:08:02: the system of observation and coordination is more important than the specs of any one drone.

00:08:07: Infinitely

00:08:07: more important.

00:08:08: If it's a sensor war, then integrated air defense is the answer.

00:08:12: And MBDA seems to be having a lot of success here.

00:08:14: They secured an export contract for their Sky Wardens system.

00:08:17: And it won the Frontex Prize, twenty twenty five.

00:08:19: That's a big deal.

00:08:20: A

00:08:20: real mark of maturity.

00:08:22: Florentaloo and stiffenrb pointed out that its strength is how integrated it is.

00:08:26: It blends advanced sensors, AI-enabled C-II, and a bunch of different effectors.

00:08:31: Like

00:08:31: the helipede laser and the Mistral III missile.

00:08:34: Right, and that blend is the key innovation.

00:08:36: Why

00:08:36: is that?

00:08:37: Why not just one or the other?

00:08:38: Because you need layers.

00:08:40: The laser is great for short-range swarms.

00:08:43: High volume, low cost per shot.

00:08:45: But the missile gives you the range and kinetic punch for bigger or faster targets.

00:08:49: It's

00:08:49: about having the right tool for the right threat.

00:08:51: It's robust, flexible, and ultimately more cost-effective.

00:08:56: And we see that same layered idea on platforms themselves.

00:09:00: Robin Rhine was discussing Rhine Metal Strike Shield.

00:09:02: Strike

00:09:02: Shield is a perfect example.

00:09:03: It's a hybrid active passive system.

00:09:06: Its design is modular, so it minimizes weight.

00:09:09: But crucially, it also minimizes the electronic signature.

00:09:13: So you're not a giant blinking target on the battlefield?

00:09:15: Precisely.

00:09:16: And that principle is now the standard for modern CUAS concepts.

00:09:20: We're also seeing big wins in pure C-II.

00:09:23: Kyle S. Erickson noted that Andrewle was selected for the U.S.

00:09:27: Army's IBCSM program.

00:09:29: That's a massive step.

00:09:30: The integrated battle command system, it basically solidifies their lattice platform for counter-UAS fire control.

00:09:36: And in Europe?

00:09:37: They're making progress.

00:09:38: The EDOCC project, the European Defense Operational Collaborative Cloud, had its successful final presentation.

00:09:46: Harold Mannheim confirmed they showed off autonomous C-II systems using cloud tech.

00:09:50: A real leap forward for collective defense over there.

00:09:53: For sure.

00:09:53: Okay, let's hit our final theme.

00:09:55: Generative AI for strategy and wargaming.

00:09:58: This is where things move off the physical battlefield and into the planning room.

00:10:02: Yeah, this is all about getting to continuous foresight.

00:10:04: Laura Sampso detailed how Gen AI is changing strategy through AI-supported wargaming.

00:10:09: Like the new Gen War Lab at Johns Hopkins APL.

00:10:12: Exactly.

00:10:13: The goal is to move past these slow episodic planning cycles and toward continuous model-driven scenario exploration.

00:10:19: So instead of a two-week paper exercise, you can test thousands of what-if scenarios really, really fast.

00:10:24: Right.

00:10:25: What are the risks of using it in that environment?

00:10:27: Well, the model could be predictable.

00:10:30: Or worse, it could have some kind of hidden bias baked in.

00:10:33: Exactly.

00:10:34: If it hallucinates a scenario that seems real but is statistically unlikely, any strategy you build on that is going to be flawed from the start.

00:10:43: Which is why you need labs like Genwar to verify and validate these models.

00:10:49: And operationally, security is everything here.

00:10:53: Ismail Alvarez made a great point about the difference between open weight and open source models.

00:10:58: A distinction a lot of people miss.

00:11:00: And it's crucial for security teams.

00:11:02: Open source means everything is public code, data, all of it.

00:11:06: Open weight means the model's parameters are public, but the training data and code might not be.

00:11:11: So defense agencies can get the capability and flexibility of the model without exposing the core platform.

00:11:17: That's the balancing act.

00:11:18: And it's no surprise the primes are moving fast.

00:11:21: Jacob Sprague noted that the Lockheed Martin AI Center is partnering directly with Skunkworks engineers to enhance their autonomy products.

00:11:27: So it's not stuck in a lab, it's getting integrated right into the development stream.

00:11:31: As fast as possible.

00:11:32: Okay, so let's summarize this deep dive.

00:11:34: We've seen autonomy maturing which brings successes like Haraka Storm, but also these huge ethical warnings.

00:11:42: We've seen urgent calls for procurement reform.

00:11:45: The forty-seven thousand dollar knob problem.

00:11:47: Right, and the push for market-based models like BRAV-E-one.

00:11:51: And then finally, the systemic integration across C-II sensors and the industrial base, all now being supported by the first serious uses of Gen.A.I.

00:12:01: in wargaming.

00:12:02: And if you connect all those threads, there was one thing that was mentioned over and over, kind of as a silent assumption.

00:12:07: Resilient PNT.

00:12:09: Position, navigation, and timing.

00:12:11: Backed by space links.

00:12:12: Yes.

00:12:13: We just assume GPS and comms will be there.

00:12:15: But they aren't guarantees.

00:12:16: They're assumptions that have to be stress tested.

00:12:19: And that leaves a pretty provocative question for everyone listening, whether you're a strategist or a product leader.

00:12:23: What's that?

00:12:24: What steps are you taking today to make sure your systems are resilient and can actually operate in a contested PNT environment?

00:12:31: because that backbone we all assume will be there.

00:12:33: That could be the next single point of

00:12:57: failure.