Authors: Elena Vrabie and Bogdan Iordache

“That was when I lost my sleep, and I decided that someone needed to do something about it,” shares Dimitrios Kottas, CEO of Delian Alliance Industries, about the birth of an emerging defence prime. The European defence-tech startup is building advanced autonomous systems to protect Europe in the era of AI‑driven warfare.

Formerly known as Lambda Automata, Delian was co-founded in 2021 by Dimitrios, along with Georgios Kontogiannis and Ioannis Souriadakis. Their goal was to deploy multiple systems in Greece for wildfire detection and coastal surveillance. Fast forward 4 years and €7M raised later, the defence tech startup expanded into autonomous strike systems, indicating readiness to scale operations and serve other high-risk areas. 

Think of the Delian tech stack like an octopus, the architect of the oceans. Jericho and Strikeweb form its brain, coordinating autonomous operations. The Interceptigon series acts as its arms, executing precise swarm attacks from hidden positions. The LAST tower serves as its eyes, providing continuous, high-accuracy surveillance. OSIRIS enables GPS-denied navigation, and M4K15 protects through electronic warfare. Together, they deliver stealth, adaptability, and distributed autonomy in contested environments.

In this interview for Underline Ventures, Dimitrios Kottas offers a powerful, multifaceted perspective on technology, geopolitics, ethics, and the evolving nature of defense. His journey from academia to entrepreneurship laid the foundation for Delian’s mission: transforming data into actionable intelligence to identify and overcome some of Europe’s bottlenecks. More than a strategic vision, it reads like a founder’s field report from the frontlines of the future of war, where “the processes right now in Europe are designed for peace”.

Underline Ventures: From academia to Apple — how did your research background lead you into big tech?

Dimitrios Kottas: I grew up in Greece. I studied electrical engineering, and in 2010, I moved to the United States to do academic research on GPS-denied navigation for mobile robots. The US Air Force funded some of the projects we had at the University of Minnesota in Minneapolis that made numerous significant contributions to GPS denied navigation. 

Around 2014, the advent of augmented reality and self-driving cars sparked the interest of tech companies in Silicon Valley, such as Google, Facebook, and Apple, and many of the technologies we were developing. That’s when I saw how different it is to work on a project where the partner is a traditional defense company, versus the same project with similar technology, where the partner is Google. So, in 2016, I dropped out of school, moved to California, and joined Apple.

I started a family young, so I wanted to have financial freedom. A lot of the top U.S. & EU academics in robotics left their universities and joined tech companies for self-driving cars and augmented reality. 

What kind of work did you do at Apple, and how did it shape your thinking?

DK: I was at Apple for about six years, at the Special Projects Group. I was fortunate to work with some top professionals tackling problems rarely accessible for testing on the scale that Apple can.

We were working on a large, mobile robot with many sensors, cameras, lidars (ed. note: Light Detection and Ranging), and radars. The project I worked on was how the robot understands where it is on the planet, with high accuracy. Later, we worked on how the robot identifies objects around it, including their shape, size, velocity, and category.

What drove you to leave Apple and launch a defence tech company in Europe?

DK: If someone back then had told me that I would start a defence company, I would have called them crazy. I had zero entrepreneurial aspirations. I did have acquaintances with people working at Anduril. They claimed the defence industry was falling behind, and we needed a more dynamic defence company resembling a startup.

I had ethical questions, not understanding the societal value of building a defence company at the beginning of the 2020s. But with geopolitical events happening, with knowledge of mobile robotics, and hearing the story of Anduril and the war in Armenia, it was clear. 

The Armenian army, modern compared to Soviet standards but with equipment from the 80s, faced Azeris who had incorporated drone technology from the Israelis and the Turkish defence industry. That scared me because many Western armies had similar dogmatic problems Armenia faced, and they hadn’t realized it yet. That was when I lost my sleep, and I decided that someone needed to do something about it.

It’s a broader mission, we live in this time in history, where we have two historical forces that act. You have powers that want to change their borders because they are not happy with how they were set after the Cold War, and they are going to renegotiate them through violence; and then, powers number two, modern technologies, which can provide them with capabilities that used to be reserved only for the superpower.

In the 80s and 90s, if you wanted to have strategic weapons or invade a country and change your territories, you needed to have fighter jets. Nowadays, you can do that using emerging technologies like drones, many of which are commercially available. Some have no moral dilemma in incorporating all those new learnings quickly and building lethal drones at an industrial scale, while others have philosophical discussions about whether we can make a weaponized drone and if it’s ethically correct.

So, in the summer of 2021, I departed Apple, I moved with my family back to Athens, and I started what was back then called Lambda Automata company because no one was doing that.

Why is integrating sensors and autonomy so critical to modern defence systems?

DK: One of the core technologies for any future weapon system is going to be autonomy and sensor fusion. If you have a lot of sensors in the battlefield — cameras and radars that look at the world, and you need to interpret the data from them to make decisions faster, then you need to share that data.

This is a growing trend. For example, in the Vietnam War, you had one or two cameras per reconnaissance. Then, during Operation Desert Storm (ed. note: part of the Gulf War) in the early 90s, you had one camera per Apache helicopter and one camera per military human. Now, in the Ukraine war, you have two cameras per soldier — their phone or their drone.

You have an exponential growth of the number of sensors in the battlefield, and those sensors, which are now cameras, will gradually also become radars. And we don’t utilize them properly; there’s a large gap in how this abundance of pixels in the battlefield leads to an abundance of intelligence and to quick decisions.

What are your current systems doing to solve these challenges?

DK: The first set of offerings we have, where we are mature, is a surveillance network that we provide to two clients, both to the civil protection in Greece, which deals with natural disasters like wildfires, and to the Greek armed forces. The surveillance network is tasked with detecting the start of wildfires as well as with monitoring vessels.

Our goal is to have all the cameras for civil protection, and all the cameras and the radars for the case of coastal surveillance, connected into one brain effectively, so that they can quickly identify targets. Up until recently, that was a manual process, where an operator was behind the camera. That operator looks at something, analyzes it, decides to communicate the coordinates to correlate that data with another sensor till that information reaches an operational center.

Now, that’s our core product, one for Civil Protection and one for the Greek Army. This comes with various components that enable it, one is the surveillance towers, which are designed to be deployed at remote locations, with no power or communication infrastructure, no road, no construction, nothing. And you can have a full-blown observation system with radar for vessels, radar for aircraft and drones, a camera, and a computer to intelligently analyze the data and stream it, stream properly to the command center, along with communications and power, and all the required rapidity to survive the remote environment. We go and we deploy an observation system from scratch.

Another scenario is we go to existing observation posts that have personnel behind the sensor to analyze data, decide on targets, and communicate them, and we try to automate their workflows using the computer we have on the tower. Here we have Jericho, which is a portable version of that with communications, where sensors that soldiers or the fire service may have in the field — a small commercial drone they use for patrolling, a small camera on a tripod they use for patrolling, a camera on a boat that they use for Patrol Unit. We offer solutions to particular use cases for our clients, a surveillance network by fusing hundreds of sensors.

What about the offensive or protective capabilities you’re developing?

DK: The second thing we want to serve is countries that are in this unstable geopolitical environment. They want to defend and protect the current status quo — the borders, the territorial waters, or the current exclusive economic zone.

To do that, you need to be able to, if requested, define what’s called in military terms “force an area denial,” deny entrance to someone in water or land. For this, you need surveillance. You have to detect quickly any potential threat that has entered that space and be able to counter those potential threats.

We now have products that are undergoing trials, and one of them is what’s called in the market: soft neutralization scan, or soft kill systems, the M4K15. It’s a system that can jump the communication protocol of a drone. Then, we have some systems on a suicide vessel and a suicide drone that are designed for neutralizing large marine targets.

The new products we’re building are heavily inspired by the learnings we see from the modern battlefield. For example, in Ukraine, we see many small Chinese anti-drone or timing systems. And there is no Western equivalent for them, so we can create low-cost, acceptable systems, incorporating many of the learnings from recent conflicts.

What can you tell me about production in Europe? 

DK: Europeans haven’t realized how far behind they have fallen compared to China in terms of manufacturing and industrial capabilities. If you try to buy something comparable to a DJI drone or a DJI selfie camera that is American or European, you are going to find something that is 10 times more expensive, and it’s not going to be the same user experience for you. 

I don’t think we should stop using any product that’s not European. We would cripple our R&D capability. Having said that, I believe that over the next 10 years, we’re going to see a wave in Eastern Europe, a growing competitive advantage. Europe is realizing that it has lost some of its core manufacturing capabilities, which need to be homegrown. It’s not something that will happen overnight. 

When you first started working on autonomy in GPS-denied environments, did you have any misconceptions, and what did you learn once you engaged with real-world defence partners?

DK: Being naive enough, coming from the commercial sector, I thought that was a technological problem that was solved in defence as well. That the leaflets or the advertisements on social media posts were enough. When we started discussing with Ukrainian companies, we realized that this is not true. 

Many of the large Western defence companies overclaim their capabilities on that problem. So, I realized that this is a capability I have worked on extensively in the past, and it was a motivation to start testing it. This, in defense, is the secret sauce, and it’s complicated; you must identify the right partners to work closely with.

How do you view the role of dual-use technologies, like your smart towers and drones, in both civilian and defence contexts, especially given the evolving nature of threats like climate crises and drone attacks?

DK: I think dual-use is becoming a blunt expression that kind of loses its meaning. There are many types of dual-use right now, and they need to be more specific as the nature of war is changing. There is the recent Ukrainian operation in Russia, called “Spiderweb”, and you can see that a lot of the targets were on critical infrastructure or sabotage operations, and those attacks usually involve drones. 

A common dual-use case for us is protecting critical infrastructure from drone attacks, which is going to be an entire market in the future, a market where there is no silver bullet, there’s no perfect solution right now. It may be that the enemy is not necessarily another state actor; the enemy may be the climate crisis.

For us, it was useful when we started the company to begin with a civilian use case, because it had less regulation, and it was a much more open network for us to enter. If you have something that works well in one sector, why not get it in another sector if the changes required are small, like for fires, where we are selling a surveillance system to the government?

What’s your footprint in Ukraine today, and what has the conflict taught you about the future of autonomous systems?

DK: We don’t have a permanent footprint there; however, just from visiting, the learnings are extremely valuable. The picture of the capability of the industry as a whole that you get from Twitter or the newspapers, versus what you get from people on the ground, is a different thing. It’s valuable for any professional in defence: if you are building a technology that does not work there, you have a problem that you need to solve. 

However, the Ukrainian cause is a peculiar case, and it’s different from most European governments. When they procure weapon systems or when they produce surveillance systems, they procure infrastructure. They procure a system that they want to use 10, 20, or 30 years from now, so they need to go through a complex and transparent process. In Ukraine, due to the nature of the cause, times are much faster.

What’s your experience with the current NATO certification and procurement processes in Europe, especially given the slow pace of these regulations?

DK: Procurement is designed to be as transparent as possible. Compliance is designed for extreme transparency and security. There are examples from World War II. A giant aircraft, like an American bomber, in the 1940s, took from concept, to prototype, to test flight, to mass manufacturing, and flying to Germany, less time than it may take us now to, I don’t know, do a software upgrade at a Ministry because of the whole procurement (ed. note: approx. 5–8 years in the ‘40s vs. 12-20 years now).

These processes right now in Europe are designed for peace. But we are going through a much more unstable era. And historically, democracies have demonstrated the ability to adapt in crisis. I believe we are going to see a shift where, instead of overly optimizing for transparency or security over efficiency, we are going to see things properly balanced.

Some say Southern Europe lacks the ambition or talent density for ‘real tech.’ Is that changing? Will the rise of dual-use and defence innovation be what nudges Europe forward?

DK: If you draw a line from Finland down to Greece and all the countries in between, many democracies think that conflict is a real scenario. It’s not sci-fi. If you are to build or work in a defence company, you’d rather work somewhere where it serves a societal purpose. If you combine that with a huge diaspora of engineers in many of these countries – Poland, Romania, Bulgaria, Greece, I think that is quite a unique combination for the region.

You’ve previously stated that “inaction is the least ethical choice” in the face of emerging threats. What changed the most in Delian from 2021 to now, and what does the next 12 months look like as you scale into an emerging defence prime?

DK: We have been extremely methodical and pragmatic so far. We have focused on a single, small European market, but at a challenging location, with large armed forces, and part of the EU and NATO. We concentrated on very particular use cases, trying to build the company in the right way, not just in terms of technology, but from business development and marketing, to sales. 

We are going to announce a round soon. We are growing into an international team; we’re 35 people now in London and Athens, with many who previously worked in markets like the U.S. and U.K. in big tech companies.

Our goal is to expand internationally, and we are getting a lot of interest from markets we never imagined would be getting interest from. That is what you’re going to see a lot of in the next year. It’s time to recreate the success we created in one country abroad.

Going to sales, some defence tech co-founders seem to be focusing solely on marketing their products. In that context, what’s your perspective when it comes to engaging stakeholders like governments or large institutions that play a critical role in shaping perception and opening doors for defence tech companies?

DK: You are a company. You sell. So you need to identify the decision maker, the end user, and the owner. The decision maker is the politician or a high-ranking general, the user is a middle-management military officer, and the owner is the taxpayer. You have to do marketing right, or more like reinvent marketing for this sector. 

Marketing is a core pillar of a defence company, but many European founders are going to tell you: “Oh, I don’t do marketing. Lockheed is not doing marketing”. But Lockheed is doing marketing. Lockheed Martin has the whole of Hollywood thinking their products are the best; they have their logo in Top Gun: Maverick. 

At the end of the day, the owners of our systems are going to be people, and you have to go for an alignment between the decision makers and the end users to sell your products. You should have a good network, and as a founder in defence, you have to find your way around that.