Monzo has been given the green light to expand across Europe, after winning a banking licence. The UK digital bank announced today that it has secured a European banking licence.The winning of the licence comes amid reports that well-known Monzo shareholders are agitating to reinstate the Monzo CEO and push out the Monzo chair. Monzo outgoing CEO TS Anil was asked to step down by the fintech’s board amid concerns over international expansion and his post-IPO commitment, according to reporting from the Financial Times. Anil's exit was announced by Monzo in October this year.It has now been reported that venture capital firms Accel and Iconiq are among the investors who have hired lawyers as they look to bring back Anil as CEO, who is set to be replaced by ex-Google executive Diana Layfield early next year. The VCs also want greater representation on the board and for Monzo chair Gary Hoffman, who has been in the role since 2019, to step aside.The revolt is said to have the backing of investors holding more than 40 per cent of shares, the report says.Separately, Monzo has won a European banking licence from the European Central Bank (ECB) and the Central Bank of Ireland (CBI), as it looks to expand across the EU. Monzo, which has over 13m customers, already has a UK banking licence. The challenger bank has been vocal about its EU expansion plans. In its 2024 annual report, Monzo said: "We've set our sights on Ireland as the destination for our EU base as we start laying the foundations for expansion across Europe."The licence, which can be passported across the EU, means Monzo can begin offering its banking products across the EU, kicking off in Ireland, which is Monzo’s EU headquarters. The licence will allow Monzo to hold customer deposits, opening the door to income streams such as loans and mortgages.In the coming month, Monzo’s Irish customers will be able to apply for retail and business Monzo accounts, Monzo said.Michael Carney, EU CEO at Monzo, said: “The approval from European regulators means we can now take our much-loved products and services to millions more personal and business customers. “Monzo has already proven that by combining the trust of a regulated bank with cutting-edge technology, we can truly transform people’s relationship with money. Today marks a significant step forward in our global mission to make money work for everyone."Elaine Deehan, country manager, Ireland, Monzo, said: “We’re excited to be launching an Irish digital bank serving customers and businesses."

As digital ordering reshapes how people buy food, one of the biggest operational bottlenecks in restaurants isn’t the kitchen — it’s the pickup counter.   Off-premises orders now dominate in many markets, yet the moment where digital orders meet the physical world remains largely manual, inefficient, and data-blind.  Startup Pickpad was founded in 2024 with a hardware-first system designed to bring automation and real-time data to restaurant pickup operations. I spoke ot co-founder and CEO Yaro Tsyhanenko at Slush this year to learn all about it.  Tsyhanenko has been building technology for the restaurant and food-delivery sector for more than 15 years. He began with a simple city guide — a better way to help people discover where to eat and how to spend their time. That product gradually evolved into a marketplace model similar to today’s delivery platforms. Across Europe, the company scaled the marketplace to 15 cities in Ukraine, facilitating millions of orders along the way. “We started with a city guide, just helping people understand where they could go and what they could do,” Tsyhanenko explained. “That naturally evolved into a marketplace, and in Ukraine, we scaled it across 15 cities and delivered millions of orders.” Building on that foundation, the team launched a ghost kitchen designed exclusively for delivery. The experience of running both the marketplace and the kitchen revealed the strength of the underlying infrastructure they had developed. “Once we started operating ghost kitchens, we realised we’d built really strong internal technology,” he said.  "At that point, it made sense to turn it into a product.” The company then began offering its platform as a service to third parties, including delivery providers and fast-food chains, effectively transitioning from an operator to a technology supplier for the broader restaurant ecosystem. “We started providing our technology as a service — first for delivery providers, and later for fast-food chains — because we knew it worked at scale,” Tsyhanenko added. The pickup counter Is restaurants’ new front line Then Russia invaded Ukraine, and he had to move his family abroad. Once in the US, he began rediscovering the restaurant space — not as an operator, but as an observer: “I was sitting in different restaurants — fast food, fast casual, coffee shops — with my stopwatch and my notebook, making notes. Just to better understand how flows are changing. Because tech is changing how millions of people order food and drinks, and how businesses operate.” He observed a shift in demand, with many restaurants seeing pickup and off-premises orders overtake in-store dining. “Statistics say that three out of four orders in the US leave the building. So it’s a complete change in consumption patterns. We all like ordering — grabbing and going, or delivery.” (In the future, it will be robots like with Estonia’s Starship). He discovered that every store has a designated place where all the digital orders are placed — the pick up area. The data gap at the pickup counter A new set of challenges is emerging at the intersection of digital ordering and physical pickup — and they all stem from the same issue: a lack of real-world data. The first is labour efficiency, he shared.   “Staff spend hundreds of hours of manual work every month just managing interactions with drivers and customers. Every interaction takes time. It creates an operational bottleneck.” There’s also a huge problem with order accuracy. People forget drinks, mix up orders — it’s very common. That leads to food waste. And there’s a data problem.  “No one knows when an order was placed on the shelf or when it was taken. That means they can’t control the full order cycle — customer experience or delivery driver performance.” In other words, restaurants don’t have the data to effectively monitor orders. Pickpad provides that missing layer of data points.  How Pickpad connects digital orders to the physical world Pickpad is a smart order-pickup system that connects the digital order world with the physical pickup counter in a seamless and automated way.  Its proprietary hardware offers a differentiation from other order management services . As Tsyhanenko shared, “We can source those unique data points and make operations better for restaurants and delivery providers.” At the core of Pickpad is a modular system of smart pads placed in a restaurant’s pickup area, each equipped with sensors and machine-learning capabilities that automate the handoff between digital orders and physical pickup. Each pad combines four sensors, a microcontroller, and a display into a simple, proprietary, and cost-efficient unit designed to work seamlessly at scale. It helps restaurants measure their pickup operations: Was the order ready on time?  Was it accurate?  What was the prep time — from receiving the order to placing it on the shelf? What was the waiting time? Was there a delivery driver delay? Where exactly was the problem when the customer complained? Pickpad sources those missing data points. “We create a new data layer using sensors. A lot happens in the real world, and no one wants to build physical hardware because it’s harder. But someone needs to do it.” The system integrates directly with existing point-of-sale (POS) software, pulling order data and updating statuses in real time as orders move through the pickup workflow. When an order is ready, the corresponding pad lights up with a customer name or identifier, prompting staff to place the prepared order on the pad. Using its sensors, Pickpad verifies order completeness by matching the physical order against POS data, helping catch missing items before pickup. Once the customer or courier collects the order, the system automatically updates the order status — eliminating manual clicks and reducing friction for both staff and customers. The user experience is intentionally minimal. “The idea is to help people without bothering them with extra tech,” Tsyhanenko said. He illustrated the flow with a simple example: “You order your coffee. The order goes into the system. The restaurant has shelves or a countertop — that’s the pickup area — and they upgrade it with Pickpads." And here the order flow works like this: “For example, Starbucks uses stickers. We just print the pad number on it. And they know where that order goes. No RFID, no expensive tags — nothing special about the cup.” The system assigns the order to a pad, which can show a number or a customer name. If something’s missing — like a doughnut — the pad shows a red light. When it’s complete, it turns green. “At that moment, we record a timestamp automatically. Staff don’t press buttons — no one does that in reality. Customers and delivery drivers get notified. The order is ready, you grab it, and it’s done.” Pickpad has gained traction for its frictionless tech. Restaurants are small — especially in cities like Berlin, New York, and London. Drivers are milling around, space is tight, and if you add complexity, it breaks. “So this is about synchronisation — timing, statuses, readiness. Not fancy displays. Data and protocols,” shared Tsyhanenko. Making hardware viable for price-sensitive hospitality Pickpad’s business model is straightforward. Restaurants receive a box containing 10, 15, or 20 smart pads, which connect to each other and require minimal setup. “It’s very simple,” said Tsyhanenko. “We ship a box with 10, 15, or 20 pads. They connect to each other, and you just plug one into a socket.” Installation is entirely self-serve, with no need for specialist technicians or complex configuration. “The setup is like IKEA,” he explained. “You unpack it, arrange the pads based on your layout, and plug them in.” Pickpad operates on a subscription model, with pricing starting at $9 per month per pad, including the hardware itself. “You pay a subscription — there’s no capital investment,” Tsyhanenko added. “Hospitality is very price-sensitive. Price can’t be the blocker.” So far, the product has gained the most traction in the UK and US, especially in fast food, fast casual, and coffee chains, where up to 80 per cent of orders leave the building. “We want inexpensive hardware for reliability and scale. We’re building physical infrastructure. Both our hardware and software are patent-pending.” Why timing beats temperature Of course I wondered about temperature monitoring as a potential data asset. Tsyhanenko admits that this was his first idea:  “I tested mug heaters, disassembled them, and talked to Starbucks managers. But temperature isn’t the problem. Many cups contain plastic. Some foods shouldn’t be reheated. It’s a timing and data problem. If you know how long an order has been waiting, you can remake it quickly and make the customer happier.” Industry validation for Pickpad’s pickup tech Pickpad’s proprietary tech has gained significant interest from tech enthusiasts and industry insiders:  CES 2025 Innovation Award honoree in the Artificial Intelligence category. Best of Tech Podcast Network recognition during CES 2025. Winner in Fast Company’s 2025 Innovation by Design Awards, which celebrate excellence in design and innovation across industries. And in terms of sector support, Pickpad won at MURTEC Startup Alley 2025 — a competition for emerging tech solutions in restaurant and hospitality operations.  50 locations, seven countries — and this is just the beginning Today, Pickpad have LOI in over 650 locations across seven countries as the startup is still in the pilot phase, rather than a full commercial rollout. Tsyhanenko is building the hardware in-house, with early versions produced using three 3D printers in his lab. “The first pilots were really about proving the concept,” he said. “Now the second wave is focused on measuring everything, building strong case studies, and getting ready for mass production.” Pickpad has built 50 new pads for the next pilot wave, which will include one large Ukrainian brand with hundreds of locations and a  US brand with about 100 locations. “The pads are approved by cybersecurity and IT teams. Now we measure everything, then scale production and sales. That’s why I’m raising a pre-seed round,” shared Tsyhanenko.

IP patent platform Ankar has raised a $20 million Series A round led by Atomico, with Index Ventures doubling down and Norrsken VC and Daphni participating.  This Series A brings Ankar’s total funding to $24 million. AI is fundamentally changing how inventions become protected assets. It enables patent teams to analyse novelty, draft claims and evaluate prior art with unprecedented depth and speed.  Yet fragmented tools, gaps in AI trust, and slow enterprise adoption have held organisations back. Ankar solves these challenges with an operating system that unifies the entire patent lifecycle into one secure workflow for the first time.  I spoke to co-founder Wiem Gharbi to learn more. Why Ankar’s founders saw an untapped problem For Ankar’s founders, these challenges were not theoretical — they had seen them firsthand. Alongside co-founder Tamar Gomez, Gharbi previously built mission-critical software at Palantir in Europe and the US. Gharbi helped adapt Palantir’s platforms to real-world business needs, supporting organisations in extracting value from complex data environments. Beyond customer deployments, she was also involved in strategic initiatives, including efforts to make Palantir’s software more accessible to startups, and worked closely with executive teams on high-level collaboration and product strategy. Toward the end of her time at Palantir, Gharbi began to notice a recurring pattern: “While exploring new ideas with other alumni, we began looking at intellectual property as a way to create long-term value.” Gharbi said. “Coming from a company that builds software for governments and large organisations, I was struck by how little innovation had actually happened in IP tooling itself.” At the same time, her co-founder, Gomez, was working in defencetech as a product manager, responsible for managing IT lifecycles. She was encountering the same issues — innovation data being tracked through tickets, spreadsheets, and fragmented systems.   Those shared experiences really became the spark for Ankar. The broken reality of modern patent workflows In practice, the traditional patent process is stubbornly manual. Most teams manage innovation and IP work largely by hand. To understand the state of the art in areas such as additive manufacturing, companies often relied on patent engineers or PhDs to read hundreds — sometimes thousands — of patent filings and scientific papers, manually extracting insights, assessing relevance, and tracking everything in spreadsheets. The process was is, fragmented, and highly inefficient. Patents can take up to 24 months to secure and still rely on scattered Word documents, spreadsheets, emails and outdated IP systems. This inefficiency slows teams down and discourages inventors from filing altogether. In response, Ankar is building an operating system for innovation. Turning data into defensible assets Ankar's platform helps patent teams go from idea to granted patent in a fraction of the time, delivering an average 40 per cent boost in productivity and hundreds of hours shifted to high-value work. That efficiency matters as IP increasingly underpins long-term competitiveness. “Today, around 90 per cent of company value is already intangible — patents, know-how, R&D output — and that proportion will only grow in the age of AI,” shared Gharbi. At the same time, enormous volumes of innovation data already exist: patents, research publications, and technical documentation. But they’re extremely hard for teams to digest and act on. “What we do is provide AI-powered assistance that accelerates that process dramatically — extracting, structuring, and surfacing relevant insights so teams can move much faster and more confidently,” explained Gharbi. Ankar aims to make that data usable. It builds AI tools that sit at the intersection of IT teams and R&D teams, helping them analyse existing innovation, build on top of it, and turn it into structured, defensible assets that actually contribute to enterprise value. One core use case is analysis and synthesis.  “Imagine a pharmaceutical company developing a new compound,” shared Gharbi.  “To understand what already exists, teams need to analyse millions of patents and research publications — for example, to identify relevant excipients or prior formulations.”  Ankar’s system uses LLMs to aggregate, analyse, and synthesise that information so teams can make sense of it quickly and make informed decisions. Once a team has created something new — say, a novel product feature or technical solution — they often need to formalise it as a patent. Writing a patent application is time-consuming, highly structured, and can run to dozens of pages.  “Our tools assist with structuring and drafting those applications, helping teams move faster while maintaining quality,” shared Gharbi. Ankar replaces the fragmented patent process with a unified platform that orchestrates the entire patent lifecycle, helping teams turn ideas into defensible global IP in hours rather than weeks.  Acting as an assistant rather than an autopilot, it combines instant novelty and prior-art analysis across 150+ million patent applications and 250+ million scientific publications with strategic drafting, prosecution support, and a single consolidated view of examiner responses—delivering stronger claims, clearer framing, and more defensible portfolios as competition for IP intensifies. However, for enterprise adoption, speed and intelligence are only part of the equation. Designing patent AI for high-sensitivity R&D environments Patent workflows require the highest confidentiality in the enterprise world, and Ankar has been built for this.  This matters more than ever: generative AI is making it easier for competitors to replicate designs, architectures and experimental approaches. As a result, C-suites, especially in automotive, electronics and other R&D-heavy sectors, are prioritising deeper, more defensible patent portfolios to secure future revenue. Ankar ensures that customer data is not fed back into shared models, implementing strict access controls and designing strong checks and balances throughout the system.  Gharbi asserts:  “Innovation data is extremely sensitive, so security and confidentiality were not afterthoughts for us — they were foundational design principles. We’ve spent significant time ensuring that data stays isolated, secure, and confidential within each customer environment.” “We also pursued ISO 27001 compliance very early, which many startups only tackle much later." For us, security isn’t a feature — it’s a prerequisite,” shared Gharbi. Further, because accuracy is critical when dealing with chemical formulas or engineering specifications, Ankar ensures its models are grounded, meaning they operate only on verified data sources such as patent databases and research publications rather than relying on free-form model memory. “On top of that, we’ve added verification layers in the product itself, so users can trace outputs back to source documents and validate results before acting on them,” explained Gharbi. According to Tamar Gomez, co-founder of Ankar AI:  “Invention is how we solve humanity’s biggest challenges, yet the systems that protect those ideas are decades out of date. AI will redefine how global organisations innovate over the next five years, turning IP from a cost centre into a growth driver. The companies that adopt Ankar now will shape the future of innovation." Andreas Helbig, partner at Atomico, said: “Tamar and Wiem bring exceptional technical depth and first-hand experience of how broken the patent process is. Their momentum with Fortune 500 companies shows they’re building the right product at exactly the right moment. We’re proud to support their mission as they build the foundational infrastructure for how the world’s most important ideas are protected and commercialised.” Ankar’s long-term ambition is to become the software layer that orchestrates how ideas become globally defensible patents, serving as the core infrastructure for innovation in the AI era.  The capital will be used to double the company’s 20-person team, expand engineering, product and design, and grow Ankar’s go-to-market organisation to support rising demand across Europe and the US. 

What continues to fascinate me about Ukrainian startups and scaleups is their ability to solve problems for other countries — problems many of those countries are still struggling to fix. From edtech and lifelong learning to mental health, cybersecurity, and HR, Ukrainian founders are building globally relevant solutions shaped by constraint, resilience, and deep technical skill. Take Brighterly (edtech), Headway Inc  (lifelong learning), Pleso Therapy (mental health support), Stackbob.ai (enterprise IAM) and PeopleForce  (HR). All of these companies are addressing systemic challenges at scale. Kyiv-founded health product marketplace Liki24 fits squarely into this pattern. Its ambition is unapologetically European: to build the continent’s leading health marketplace across fragmented pharmacy markets. But that ambition was tested earlier — and harder — than most startups ever face. Founded in Kyiv, the company has navigated COVID, war, and investor flight, emerging with 70 per cent of its revenue now coming from the EU and €19 million in funding since its founding in 2017. I sat down with Anton Avrynskyi, CEO and co-founder at IT Arena in Lviv, earlier this year to learn all about it. How Liki24 lets customers trade speed for price in healthcare Liki24 is a marketplace for health products. It connects sellers with buyers of physical health products — medicines, supplements, vitamins, medical cosmetics, and OTC drugs. Liki24 operates as a marketplace that aggregates products from multiple pharmacies and optimises for price, speed, and availability. A customer, for example, may choose from: Delivery in one hour from a local pharmacy, Next-day delivery from within the country, or Delivery from another EU country at a much lower prices. Besides Ukraine, its present in 10 European countries and the UK. And, the model was shaped by a very personal frustration Avrynskyi had experienced years earlier. A mid-career leap: how Anton Avrynskyi walked away from stability to fix a broken pharmacy market Anton Avrynskyi is from Kyiv, Ukraine. He studied computer science at Kyiv Polytechnic University and, from his second year, began working at a tech company that developed and implemented ERP systems for large enterprises and corporations. He went from developer to project manager, then head of the project department, and eventually became a shareholder and the company's CEO. After 15 years, he started thinking that he wanted to build a big international interest. He walked away from the company and sold his shares without fully knowing what he would do next, a leap that highlights how much harder risk-taking becomes for founders with families and long-built careers. Avrynskyi admits, “It was a big risk. At the time, I had two kids — one was four years old, the other just one — and my wife. She asked me, “What are you doing?” But she supported me, which was incredibly important. I understood that if I didn’t leave, I would stay there for the rest of my life. My biggest internal fear was losing motivation and internal drive.” He spent a few months in the US with his family, attended many conferences, and saw how fast digital health was growing. At that time, Avrynskyi remembered the problem he faced in Ukraine when his children were born: There were huge price differences for the same medicines between pharmacies — sometimes two or three times; lack of availability — you couldn’t buy everything you needed in one place; and inconvenient or impossible delivery of everything in one order. He realised that the US hadn’t cracked this problem, so he decided to create a solution in Ukraine. He asked former colleagues from his previous company to join him and started the company in mid-2017, about half a year after exiting his previous business. Things moved fast. He recalled: “We built an MVP very quickly, saw good traction within a few months, and then I was invited to Austria to present the company. We won a startup competition there. After that, we raised our first funding round of €1 million. Two years later, COVID hit — and our growth accelerated dramatically, as demand for remote access surged. During that period, we raised a €5 million round led by Horizon Capital. That’s when we began executing the company’s full-scale vision.” From investor pause to European expansion By the end of 2021, just months before the full-scale invasion, Liki24 was experiencing strong growth and had already begun fundraising. But as geopolitical tensions escalated, that momentum stalled. By mid-December, investor conversations abruptly paused. “We had great numbers and a clear direction, but Ukraine was suddenly seen as too high risk,” he says. “Everyone was talking about the possibility of war.” When the team re-engaged investors in April 2022, the message was consistent — and blunt. “They told us, ‘Let’s talk only when more than half of your business is outside Ukraine.’ Some said it directly, others said it more carefully, but the reason was obvious.” At the time, around 97 per cent of the company’s business was still inside Ukraine, with international operations not yet financially sustainable. “From an investor perspective, it was absolutely fair. The risk was simply too high.” How Liki24 mobilised medicines for a country at war On 24 February 2022, Avrynskyi was in Kyiv with his family. He recalls: “We woke up to bombs. During the first two weeks of the war, more than 80 per cent of pharmacies were closed — not because products weren’t available, but because pharmacists had fled." Just a day later, Liki24 created a free map of open pharmacies across Ukraine. Almost 2 million Ukrainians used it in the first two weeks. Then the team expanded it so that people could find pharmacies that actually had the medicines they needed. However, Liki24 also observed that when supply chains were disrupted, people began facing shortages of essential products. Fortunately, during that time, Avrynskyi received many messages from international friends saying, “We’ll send you everything you need — just tell us what’s missing.” Friends from Greece, Turkey, and France sent entire trucks — big trucks — filled with medicines, an incredible range of much-needed medications. Avrynskyi explained: “We created three logistics centres and sorted everything. It was extremely hard because product names differed across languages, and dosages were sometimes different. We worked with more than 50 volunteer pharmacists and built a platform that allows Ukrainians to order exactly the medicines they need for free. This was crucial — medicine requires precision. In total, we distributed around $5 million worth of medication and helped more than half a million Ukrainians.” However, the project was a victim of its own success. Healthcare organisations told them that free humanitarian medicine was hurting local pharmacies' finances. So they changed the model, asking international partners to send money instead of medicines, created an NGO, and upgraded the platform so Ukrainians in need could upload documents and order what they needed — donations covered around half the cost. “This way, we supported local pharmacies and helped around one million Ukrainians. Donors could see exactly where every euro went via an online dashboard.” Business survival vs humanitarian responsibility However, what made the situation especially complex was that the business was operating on two fronts at once: running a company while simultaneously responding to a humanitarian crisis. “On the outside, it was a 24/7 charity effort. Inside the company, there was still a business that had to survive — with a team to support and salaries to pay.” But while still running humanitarian operations, Liki24 focused on European expansion — Romania first, then Italy, Germany, France, and Spain. While many companies focus on their US for growth, according to Avrynskyi: “Europe is underestimated. Over 500 million people, fragmented markets, and around $400 billion in pharmacy sales — larger than the US. Fragmentation is our advantage. Our goal is to become Europe’s leading health marketplace.” By 2023, 35 to 40 per cent of revenue came from the EU. Today, it’s 70 per cent. Revenue in Ukraine has still grown twofold since the invasion, and Liki24 raised €9 million this year, bringing its total funding to €19 million. AI under the hood: product matching, logistics, and support at scale AI was a key enabler of making European scale possible and plays a major role across three key areas at Liki24: Product mapping. Avrynskyi explained: “There are thousands of sellers, each with thousands of products, often the same products with different names. We use machine learning to match them accurately.” So the company used AI for three specific needs: Product mapping: Thousands of sellers, each with thousands of products, often the same products with different names. Likl24 uses machine learning to accurately match them. Logistics: Traditional carriers failed for cross-border delivery—slow and expensive. “We broke logistics into local segments and built an AI delivery agent that chooses the optimal chain. This cut delivery times sevenfold and costs fivefold.” Customer support: Liki24 reduced its support team from 80 to about 40 people while expanding to more countries, using AI-driven multilingual support available 24/7. The company also uses AI for recommendations, marketing, and B2B pharma advertising—without sharing personal data. But this is not the end of innovation for Liki24. The company launched a lab test marketplace and a health coaching platform focused on prevention. “People do regular blood tests, track trends, and receive personalised health plans. It’s especially valuable for expats navigating unfamiliar healthcare systems. Eventually, everything will connect: diagnostics, coaching, doctor access, and product delivery via Liki24,” shared Avrynskyi. Avrynskyi stresses that readers should support Ukraine — not only through aid but also by supporting Ukrainian businesses. “Ukraine should not be seen only as a country in need, but as a country that creates value and innovation. Our story shows that even during war, it’s possible to build and scale international companies."

AI-powered unmanned systems company Quantum Systems today announced the acquisition of FERNRIDE.   With the acquisition of FERNRIDE, Quantum Systems is expanding its leadership position from air and software intelligence to autonomous ground mobility,  creating a connected, cross-domain offering for intelligent unmanned systems.  FERNRIDE offers an industry-proven software platform for ground autonomy in the areas of container terminals, defence logistics, yard operations and autonomous driving. The technology is already in use by well-known customers.  In 2025, FERNRIDE became the first company to obtain TÜV approval for autonomous trucks in Europe, and has already used them in initial tests with the  German Armed Forces and expanded its portfolio to include defence logistics.  Check out our earlier interview with Hendrik Kramer, CEO and co-founder of FERNRIDE. “‘FERNRIDE has developed one of the most advanced and scalable autonomous ground platforms,’ said Florian Seibel, Co-CEO and Co-Founder of Quantum Systems. "By integrating their technology into MOSAIC UXS, we are consistently  implementing our vision of creating a connected ecosystem in which unmanned  systems think, move and act as a single entity across different dimensions." "Europe urgently needs sovereign autonomy solutions. By joining forces with  Quantum Systems, we can take our technology to a new level,"  said Hendrik Kramer, CEO and co-founder of FERNRIDE. "Together with Quantum Systems, we are accelerating the deployment of our platform in the European defence sector, which is currently the most urgent environment globally for scaling autonomous ground systems. In the future, this experience will also be applied to civilian logistics applications, making our society safer and more resilient." Earlier this week, Quantum Systems and Frontline Robotics announced the creation of “Quantum Frontline Industries” (QFI), a German-Ukrainian joint venture that will establish Europe’s first fully automated, industrial-scale foreign production line for drones for the Ukrainian Armed Forces. Quantum Systems' operational experience in Ukraine has demonstrated the relevance of the interaction between air and ground robotics. The integration of FERNRIDE solutions into MOSAIC UXS, Quantum Systems' autonomous mission software, is intended to enable multi-domain operations that improve situational awareness and decision-making. The financial deals have not been disclosed.

London-based deeptech AI company geoSurge has launched out of stealth, focusing on helping brands understand and improve how they are represented within large language models. As AI increasingly serves as a starting point for everyday decisions, visibility within generative AI systems is becoming more commercially significant. Recent reports find that ChatGPT prompt volume rose by nearly 70 per cent in the first half of 2025, while shopping-related queries doubled over the same period. In this context, appearing in LLM-generated answers can influence consideration, while brands that are not surfaced may be less likely to be evaluated. geoSurge says its data indicates that even well-known brands can lose visibility between GPT releases. Its platform helps organisations assess how their brands appear in LLM-generated responses, identify weaknesses in representation, anticipate how upcoming releases may affect them, and take steps intended to improve and maintain visibility as models change. geoSurge’s approach, which it describes as “corpus engineering,” is designed to strengthen how large language models learn, store, and surface brand information, with the aim of making generative visibility more stable across model updates. Francisco Vigo, co-founder and CEO of geoSurge, explained: LLMs don’t pull from a live index. They generate answers from compressed memory that shifts with every update. That means a brand can go from ‘high visibility’ to ‘completely gone’ overnight - and most organisations won’t even know it’s happened. Founded in London, geoSurge works with companies across retail, travel, fintech, and consumer services, supporting visibility in AI-driven discovery as LLMs evolve.

Wodan AI, a startup specialised in running artificial intelligence models directly on fully encrypted data, has closed a €2 million pre-seed round led by Spanish funds JME Ventures, Swanlaab and Adara Ventures, with additional participation from ScaleFund in Belgium. Wodan AI is a European company developing homomorphic encryption technology for use in artificial intelligence. Its platform is designed to enable ML, computer vision, and LLM models to run on encrypted data without decryption, reducing exposure of sensitive information. Its platform ensures compliance with the EU’s most stringent requirements on privacy, regulation and digital sovereignty, especially relevant in sectors such as finance, defence, healthcare and critical infrastructure. Commenting on the investment, Bob Dubois, CEO and co-founder of Wodan AI, said the funding will support the company’s plans to expand its technology and strengthen its position in the European secure AI landscape. Recent advances in homomorphic encryption allow, for the first time, large organisations to use artificial intelligence on sensitive data without compromising security. The company plans to make Spain a central part of its European strategy by relocating its global headquarters to Madrid and consolidating its R&D team there, with a focus on advanced cryptography, machine learning, and privacy technologies. It also develops a pilot projects with a Spanish financial institution, which it describes as an early validation of encrypted AI for critical-sector use cases. Spain has the technical talent required to develop critical sovereign AI technologies. With our headquarters, R&D centre and team growth in Madrid, we aim to consolidate a European hub of reference in private, secure AI designed for the most demanding sectors, adds Manuel Pérez Yllan, CTO and co-founder of Wodan AI. The funding will support R&D hiring, accelerate the technical roadmap, and extend the platform to more advanced encrypted computer vision and LLM capabilities, with an aim to deploy these technologies at scale. It will also be used to strengthen commercial capacity and pursue contracts with European organisations in strategic sectors.

Global connectivity provider Roamless has closed a $12 million Series A funding round led by Rasmal Ventures, with participation from Shorooq, Revo Capital, Finberg, and JIMCO. The new capital brings Roamless' total funding to $18 million, following a $6 million seed round in 2024. Founded in Türkiye and headquartered in the US, Roamless was established in 2023 by Emre Demirel, Ali Gazioglu, Asim Alp, Selim Aykut, Cagdas Yalti, and Cengiz Oztelcan. It offers a borderless mobile connectivity platform built around a single eSIM intended to work across multiple countries. Roamless provides a “Single Global eSIM” with pay-as-you-go and trip-based plans across 200+ countries, designed to keep travellers connected without changing SIM cards. The company says it operates on its own telecom infrastructure and plans to add features such as local numbers, voice, SMS, and partner APIs. Roamless reports serving more than one million travelers and working with partners in travel, aviation, and financial services. Roamless telecom stack is a cloud-based, carrier-grade infrastructure layer that manages connectivity across hundreds of networks through a single service. This supports cross-border coverage and is intended to enable services beyond data-only offerings, while giving it more control over the customer experience than a resale model. The eSIM market has been growing quickly, supported by wider adoption from smartphone manufacturers. While fewer than 20 per cent of global smartphone connections used eSIM in 2024, the GSMA projects penetration could reach up to 88 per cent by 2030. Roamless also expects adoption to continue rising as more devices move away from physical SIM trays. Emre Demirel, co-founder and CEO of Roamless, said the company expects the market to grow significantly and believes long-term winners will combine a strong product with reliable infrastructure and carrier-grade technology. We are rolling our local market, b2b and AI-enabled features in our next phase - all to ensure we continue to lead on innovation, and give global mobile users what they need to maximise their experience. In the next phase, the company will expand network coverage, scale its global go-to-market and customer support efforts, and deepen supplier and corporate partnerships. It plans to launch Roamless Numbers, which would allow users to obtain local numbers in more than 20 countries and make calls or send and receive SMS within the app. The company will also invest in APIs and B2B solutions so airlines, airports, online travel agencies, financial institutions, and superapps can embed connectivity into their customer journeys. In addition, it expects to roll out AI-driven features aimed at improving network quality, lowering costs, and surfacing relevant partner offers, while growing its team across regional offices in major travel hubs.

A David vs. Goliath shift is emerging in AI as a small Nordic team is challenging Big Tech’s cloud-LLM dominance by bringing Small Language Models (SLM) directly to users’ devices. Copenhagen-based open-source startup NobodyWho has raised €2 million in pre-seed funding. The company aims to strengthen Europe’s position in global AI by championing Small Language Models as a cost-efficient, data-secure, and climate-aligned alternative to today’s massive cloud-based LLMs. I spoke to founder to CEO, Cecilie Waagner Falkenstrøm, to learn all about it. NobodyWho is founded by award-winning entrepreneur and artist Cecilie Waagner Falkenstrøm, whose pioneering work with interactive AI dates back to 2016. Together with co-founder and CTO, Asbjørn Olling, and a team of software engineers, she has spent nearly a decade advancing local-AI technologies — from UN-commissioned projects to a 2021 edge-AI experiment aboard NASA’s International Space Station.  Today’s cloud-based LLMs are controlled by a handful of non-European tech giants and require massive computational resources, constant internet access, and the transfer of vast amounts of data to third-party servers. This creates high costs, lock-in, and a structural loss of European data security.  NobodyWho takes a fundamentally different approach. Its engine enables Small Language Models (SLMs) to run locally on laptops and mobile phones, so organisations and individuals keep full control over their data. With device-first architecture, no data needs to leave the device enabling true data sovereignty and privacy by design.  Waagner Falkenstrøm explains: “These models are still large by most standards, but they’re much smaller than systems like ChatGPT. They’re comparable to earlier generations of large modelsn— and they’re more than capable for many real-world use cases.” Local inference as a security and privacy advantage Running models locally has immediate privacy and security implications. From a security standpoint, this also creates a more resilient architecture. Rather than relying on a single, centralised cloud server that can be targeted, computation is distributed across thousands—or even millions—of devices. Further, local inference shifts the cost burden away from cloud infrastructure entirely. Users bring their own hardware, meaning the system can scale without increasing inference costs. Whether an application serves ten users or ten million, there is no escalating cloud bill.  This makes advanced AI accessible to organisations that would otherwise be priced out, including NGOs, public-sector bodies, and early-stage startups. A 500x reduction in AI’s carbon footprint NobodyWho’s local-first SLM architecture dramatically reduces this footprint by shrinking the models and moving them close to where they are used. Early benchmarks show up to 100x lower training footprint and up to 500x lower inference footprint. So this approach isn’t just cheaper and faster — it’s also dramatically more sustainable. Open source by default Everything core to NobodyWho is open source — from its inference engine and inference libraries to its developer integrations — and that will remain the case. Rather than building proprietary language models, NobodyWho focuses on the infrastructure layer that makes existing open-source models usable in real-world products. Its engine enables more than 10,000 open-source language models to run efficiently across devices and operating systems. “Our belief is simple: the models already exist,” says founder and CEO Cecilie Waagner Falkenstrøm. “The real bottleneck is making them practical to deploy — especially for developers who don’t have machine-learning expertise.” Most developers, she explains, aren’t ML specialists. NobodyWho’s goal is to make running a local language model as straightforward as integrating any other software dependency. “An app developer should be able to run a local model with two lines of code,” she says. “You shouldn’t need a PhD in machine learning to ship AI.” To achieve this, NobodyWho integrates directly with major developer frameworks. The company recently launched Python support and is expanding into additional ecosystems, allowing developers to drop NobodyWho into existing projects without deep ML knowledge or custom infrastructure work. The company operates on an open-core business model. While all core components remain open source, NobodyWho monetises fine-tuning services — an area where compute requirements quickly become expensive and operationally complex for teams to manage alone. “Companies could fine-tune models on-prem,” Waagner Falkenstrøm explains. “But that means servers, engineering time, and ongoing maintenance. Instead, they can fine-tune models using our engine, pay for the compute, and we take a small cut. It’s still significantly cheaper and simpler than doing it themselves.” Once a model is fine-tuned, it can be deployed to millions of end users with no additional inference cost, a key advantage of running models locally rather than in the cloud. Small language models, running where the data lives I was curious what the trade-off is from using SLMs instead of LLMs. “Historically, making an API call to a cloud-based model was easy, and running models locally was hard. "We've solved that problem,” shared Waagner Falkenstrøm. “With NobodyWho, the traditional complexity gap of local inference is effectively removed.” There are still some use cases where very large models are necessary — extremely broad or complex reasoning tasks. Those models won’t disappear. But for most real-world business applications, such as chatbots, HR assistants, customer support, and domain-specific tools,  Small Language Models are more than sufficient, especially when fine-tuned. “Fine-tuning smaller models is also easier, explained Waagner Falkenstrøm., “You need less data and less compute, and you get more controllable behaviour. Most companies operate within specific contexts, and small models excel there.” NobodyWho uses the European Public License (EUPL) 1.2, which explicitly allows both individuals and companies to build commercial products on top of its code — a deliberate choice aimed at driving real-world adoption with cross-platform support across mainstream operating systems and development frameworks included. “If you want genuine uptake, commercial use has to be allowed,” says Waagner Falkenstrøm. “Otherwise you don’t get an ecosystem — you get a demo.” Over 5,000 devs building The ecosystem is already forming. NobodyWho now has more than 5,000 developers building with the platform via GitHub, alongside an active Discord community where contributors discuss use cases, share feedback, and help shape the roadmap. “The open-source aspect is critical,” Waagner Falkenstrøm adds. “It’s what allows a real community to emerge — not just users, but contributors.” With the platform well beyond MVP, the company’s focus has shifted firmly to scale. “We’re past the experimentation phase,” she says. “Now it’s about expanding framework support and enabling more developers to build production-grade applications.” She believes the next leap forward in AI will come from making models smaller, more local, and more human-centric.  “The future of AI won’t be won by size, but by decentralised models that anyone can run on their own devices.” Investors view the rise of local, energy-efficient AI as a major strategic opportunity for Europe, especially as demand grows for privacy-compliant and cost-effective alternatives to cloud models. The round is backed by PSV Tech and The Footprint Firm, and Norrsken Evolve. “I’ve known Cecilie for nearly a decade and have seen first-hand how she consistently turns bold ideas into real, working technology,” says Christel Piron, co-founder and General Partner at PSV Tech: “Backing NobodyWho was a no-brainer for us: this is an exceptional team building critical European AI infrastructure that is privacy-protecting, energy-efficient, and accessible to developers and companies everywhere”. Sofie Käll, CIO at The Footprint Firm, shared: “NobodyWho is pioneering the infrastructure that makes these ultra-efficient models truly plug-and-play for developers.  This is a transformative climate-tech opportunity in one of the fastest-growing emissions categories, and we’re excited to support a team capable of moving the industry toward more responsible AI."  Waagner Falkenstrøm contends that no matter what we do, Europe will not outcompete the US or China in the “bigger is better” game.  “The compute, capital, and hyperscale infrastructure simply aren’t comparable. “But from our experience, we knew that smaller models are genuinely powerful in many fields. That creates an opportunity for Europe to compete differently.” At the same time, there’s a strong values-based dimension to what NobodyWho is doing.  She asserts: “Coming from the EU and the Nordics, we care deeply about data security, GDPR compliance, sustainability, and data sovereignty. NobodyWho is designed to reflect those values.” “Technology is power” Waagner Falkenstrøm asserts that Europe needs to believe in itself. We have some of the best education systems, software engineers, and research institutions in the world. “We don’t need to copy the US — we need to build AI that reflects European strengths and values. Further, technology is power. If we care about privacy, sovereignty, sustainability, and democratic control, those values must be embedded in the technology itself. That’s what we’re trying to do with NobodyWho: decentralised, open, privacy-preserving AI that anyone can build on.” Waagner Falkenstrøm sees NobodyWho as part of the first real wave of companies building infrastructure specifically for Small Language Models. “A year ago, SLMs weren’t widely discussed. Today, developers and investors understand the category. The models are improving rapidly, and the tooling has matured. Big tech companies will enter this space — but they’ll optimise for their own ecosystems. Apple will build for Apple. Microsoft will build for Microsoft. We’re platform-agnostic. That creates a meaningful opportunity.”

Oxford-based materials science startup A&B Smart Materials has closed a £1.5 million pre-seed funding round. The round was led by existing investor Sake Bosch, with new strategic investors Caesar and Living Hope VC, alongside participation from Archipelago Ventures, Triple Impact Ventures, Cranfield University Seed Fund, Oxford Seed Fund, and several business angels from Cambridge Capital Group and Oxford Innovation Finance. The superabsorbent polymers (SAP) market is sizeable and growing, driven mainly by demand from absorbent hygiene products such as nappies and menstrual pads, with additional use in agriculture and smaller applications in areas like medical products, construction, consumer goods, and water treatment. Most SAPs in use today are synthetic and fossil-based, and because they are not designed to biodegrade, they can persist in the environment and contribute to microplastic pollution, including through high-volume single-use products such as disposable nappies. A&B Smart Materials is developing fully biodegradable alternatives to fossil-based superabsorbent polymers, using polymer science and natural feedstocks to create absorbent materials intended to match existing performance requirements and fit within established manufacturing processes. Its approach is based on modified biopolymers derived from widely available, lower-cost natural materials. A&B Smart Materials co-founder and CTO Dr. Benjamin White said the persistence of synthetic superabsorbent polymers contributes to long-lasting pollution, including microplastics in the environment. We intend to completely replace these products with biocompatible and biodegradable materials, without compromising on product performance or affordability. The funding will primarily be used to accelerate R&D to refine A&B’s sustainable superabsorbent polymer formulations and target a combination of strong performance, competitive cost, and industrial-scale validation for hygiene and agricultural applications. The company’s longer-term objective is to replace synthetic SAPs in a market projected to reach about $17 billion by 2035.

Resale platform Bought, launched in March 2025, has announced its third acquisition of the year, Robes Rental, to strengthen its position in Europe and explore peer-to-peer renting. Earlier this year, Bought acquired the Finnish secondhand platform Zadaa as well as the bankruptcy estate of Vähänkäytetty.fi. Bought was the first to launch peer-to-peer local logistics in partnership with Wolt earlier this year. Check out our earlier interview with Erik Kymäläinen, co-founder and CEO of Bought. Robes Rental is a Finnish peer-to-peer rental platform focused on luxury fashion. Featured in Vogue and loved by celebrities and politicians alike, Robes’ technology, industry knowledge, and dedicated community allow for Bought to strengthen its position as the up-and-coming circular platform arising from the Nordics. “Robes and Bought have been aligned in many ways from the start. Both were founded in Helsinki by young and ambitious teams with the goal of changing the way we buy, sell, and own. Both believe that technology plays a key role in enabling this shift in circularity. This collaboration gives us the tools to realise our vision even more efficiently,” says Erik Kymäläinen, co-founder and CEO of Bought. “Two years ago, we set out to build a new kind of wardrobe experience, one that made luxury fashion accessible and circular. By combining Bought’s automated resale and digital closet technology with Robes’ expertise in rental and circular wardrobes, we can offer a more complete circular fashion experience and create greater impact from one place. This is an exciting new chapter for our community and for the entire circular fashion ecosystem in Europe,” says Anna Sillanpää, co-founder and CEO of Robes Rental. Moving forward, Robes’ technology will be transferred to Bought to support the development of advanced circular wardrobe capabilities on Bought’s platform. This means that the standalone Robes app will sunset at the end of the year, making way for something even more impactful through Bought.

Monzo is making its first-ever acquisition, buying UK digital mortgage broker Habito. The acquisition of Habito, which employs around 100 people, comes as the UK challenger bank looks to beef up its mortgage offering and diversify beyond its core digital banking products. Kunal Malani, Monzo chief banking officer, posted on LinkedIn: “Buying a home should be exciting - not overwhelming. Together, Monzo and Habito will make it easier for customers to find, compare and secure the right mortgage, all from the Monzo app." Monzo said the deal, for an undisclosed amount, means it’s the first UK bank to offer an end-to-end mortgage broking experience within its app. Monzo, which has more than 14m customers, currently offers users a Homeownership feature, which it says more than 450,000 users use to track their mortgage, home value and mortgage deals. The UK digital bank pointed to data showing 87 per cent of UK mortgage seekers used a mortgage broker. London-based Habito, founded in 2016, is a digital mortgage broker which provides tools, including mortgage switching and home-buying services. It is backed by investors including Augmentum Fintech, SBI Investment and Volution. Ying Tan, CEO of Habito, said: “At Habito, we’ve always believed mortgages should be easier, fairer, and simpler for everyone. "I’m incredibly proud of what our team has built, and I couldn’t be more excited for this next chapter with Monzo. Together, we’ll transform what the mortgage experience feels like - effortless, empowering, and truly built around people." Financial terms of the transaction are not disclosed, with the deal expected to close in the spring of 2026.

OpenAI for Startups is OpenAI’s programme designed to help early-stage companies build and scale products using AI. Rather than just offering model access, it focuses on removing practical barriers for founders by combining technical support, resources, and credits to accelerate product development. In practice, that means pairing access to OpenAI’s models with people, infrastructure, and hands-on support designed to help teams move faster from prototype to production. Startups in the programme can receive OpenAI API credits, higher rate limits, and access to hands-on guidance from OpenAI’s technical teams. Those backed by participating VC partners can unlock additional benefits, including enhanced support and invitations to founder-focused events. I spoke with Mark Manara, Head of Startups at OpenAI, and Romain Huet, Head of Developer Experience, at Slush in Helsinki, to learn more about their support for startups and the biggest trends. Mark Manara , who leads startups at OpenAI, heads a global team working closely with companies building on top of the OpenAI platform. VC partnerships are an extension of startup support. The team also collaborates extensively with venture capital funds through a dedicated VC partnerships function, focused on providing the right resources and hands-on support to the startups they back OpenAI also hosts events and VC summits in places like London, San Francisco, where it brings together its leadership, product teams, and startup teams to share what it's building, what patterns it's seeing, and to hear feedback directly. The focus, Manara says, is on giving founders practical leverage — from infrastructure and credits to direct time with OpenAI’s solutions and engineering teams.  "We focus on resourcing portfolio companies — credits, technical support, and access to our solutions and engineering teams.”  According to M, VCs are understandably most interested in roadmaps but also assurance that OpenAI is closely partnered with their scaling companies. “There’s a lot of interest in understanding what we’re seeing, what we’re building, and we try to share those insights with VCs so they can better support their companies,” shared Manara. “Increasingly, they’re interested in ChatGPT as a distribution channel. With hundreds of millions of weekly users, there’s real interest in embedding startup experiences directly into that ecosystem. Commerce is another emerging area of interest.” Overall, the company’s work with VCs and startups shapes how OpenAI evaluates startups. OpenAI’s litmus test for startups Despite OpenAI’s scale and resources, Manara is quick to stress that the startup-facing team itself is still small. He admits that while the company is big in terms of funding,“We still feel like babies.” “My team is about 45 people globally. I’m based in San Francisco. We started there, but now we’re in Europe and Asia as well.” For Manara, a great startup, from our perspective, is pushing the frontier of how they’re using OpenAI’s models to build product.  “We work with what we call AI-native companies — where there’s an LLM at the core of the product. If you took it out, the product wouldn’t work anymore. That’s the litmus test. Within that, we want companies that are really operating at the bleeding edge. We work across many verticals — coding, customer support, legal tech, healthcare — and also newer companies taking entirely different approaches.” Being at that edge provides a feedback loop. Startups help the team understand how models need to improve to support specific tasks such as legal workflows, live conversation, or sales automation.  “That feedback accelerates our own pace of development.” In return, model improvements are usually very closely tied to product-market fit for those companies.  “It becomes self-reinforcing,” shared Manara.  As more startups build on the same underlying models, Huet argues that defensibility — not just technical capability — has become the defining challenge for founders. The real moat in AI startups: deep problem understanding When it comes to defensibility and category leadership, Huet argues that access to powerful models is no longer enough. He contends that in this wave of startups, there are a lot of competitors emerging in the same categories. “ Legal tech is a good example — I could name a dozen companies off the top of my head.” So what differentiates them? Part of it is product.  “There’s real skill in designing a great product, even without AI. User experience matters enormously," explained Huet. “With AI specifically, how you use the models matters a lot. Some teams have a much deeper understanding of how models are built — how to prompt better, how to provide context, what’s in distribution and what isn’t. That AI engineering sophistication really shows.” Speed still matters. As Manara puts it: “There’s a joke that speed is the only moat in the application layer right now — and there’s some truth to it. Teams that ship fast, get in front of customers quickly, and react in real time have a genuine edge.” Romain Huet agrees, but argues that speed alone is no longer a differentiator. “Speed has almost become table stakes. Builders can now go from an idea to a working feature incredibly fast,” he says. "What really matters is an obsession with the problem you’re solving. Unless you’ve spent dozens — even hundreds — of hours deeply understanding a customer's pain point, it’s extremely hard to solve it well, even with AI.” The strongest founders, Huet adds, are those who combine sharp AI intuition with deep customer obsession — using speed not as a shortcut, but as a force multiplier. The importance of understanding LLMs In terms of the teams that break through, according to Manara, most of the teams OpenAI sees doing really well have very strong engineering backgrounds that sometimes border on research.  “They’re not doing foundation model research, but they understand how models work. There’s a new skill set here that’s different from building a web app. Some teams experiment with fine-tuning — not as a first step, but when it makes sense. That requires understanding data composition, overfitting, and evaluation. That’s a different discipline than hooking up a database to compute.” How startups shape OpenAI’s roadmap Startups play a critical role in OpenAI’s feedback loop. “Startups often provide reproducible examples that our research teams can investigate and build evaluations around,” says Manara. “There’s an old adage in programming languages: the ones people complain about are the ones people use. Models are similar. Even very successful companies can point to many things they want improved.” That feedback feeds directly into OpenAI’s research priorities, which span everything from highly technical issues — such as improving tool-calling accuracy for AI agents — to more visible product capabilities. One area of sustained investment is coding. “Coding isn’t one thing,” Manara explains. “It includes code review, generation, schema adherence, language specificity, and more. We’re constantly iterating across all of those dimensions.” Romain Huet notes how quickly the role of AI in software development has evolved.“Coding has changed dramatically in the last few months. Where models once helped with snippets or light tasks, they now function more like teammates — taking on large, complex work for hours and returning complete outputs,” he says. “That’s why we’re continuing to release models optimised specifically for coding.” According to Manara, “This cycle is different from previous tech waves. When we release something new, it can materially change a startup’s roadmap or an investor’s thesis. So those conversations are critical.” Why pivots have become easier  I’ve seen more startup pivots in the last 18 months than ever before. Why now? According to Huet, pivoting used to be extremely costly — six to twelve months of runway. Now, with AI tools, teams can test new directions in days or weeks: “Founders can explore new customer segments or problems very quickly. Accelerator teams are pivoting multiple times in a short period, which would have been unthinkable a few years ago.” Part of pivoting is open expanding to new markets. In terms of emerging areas, Huet asserts that multimodality and speech-to-speech are still underused.  “The quality is now there, and pricing has dropped enough for startups to build viable products. As AI moves into the physical world — robots, devices, hardware — voice will likely become the primary interface.” OpenAI is learning from Europe According to Manara, some of the most exciting startups we work with are coming out of Europe.  “The ecosystem is vibrant, funding is increasing, and many European companies are category leaders. Several European startups have taught us the most about how our models are used in production.” Huet revealed: “I built my first company in Paris 17 years ago, when the startup ecosystem barely existed. Today, Europe has talent, capital, and experience. What excites me isn’t comparing valuations to the US — it’s the trajectory. Five years ago, many of these companies couldn’t have been built here. Now they can.” Ultimately, Manara wants startups to know “We’re here to work with startups, and we want feedback. Benchmarks matter, but what matters more is how models perform in real products. That’s how we learn and improve.” Huet stresses, "The pace of change isn’t slowing down — it’s accelerating. Founders who stay curious and master the tools will have a real edge. And we’re happy to help.”

Document automation company PandaDoc has raised €5 million from Índico Capital Partners. Founded in 2013 by Mikita Mikado (CEO) and Serge Barysiuk (CTO), PandaDoc began in response to inefficiencies in document processes, particularly for smaller businesses. It later shifted its focus to document automation software, expanded its operations, and grew internationally. In September 2021, PandaDoc reached unicorn status with a valuation of more than $1 billion. Today, PandaDoc provides tools for creating, sending, signing, and managing business documents digitally. Its platform supports workflows such as proposals, contracts, quotes, invoicing, and payments, and is used by teams including sales, operations, HR, legal, and revenue to streamline document-related processes. PandaDoc integrates with a range of CRM, payment, and productivity tools and is designed to replace manual, paper-based document handling with digital workflows. PandaDoc has an office in Lisbon that serves as a European hub for the company, making Portugal an important base for its operations. We're thrilled to partner with Índico Capital Partners as we enter our next chapter of growth. Their deep expertise in AI and track record of scaling technology companies makes them the ideal partner to help us accelerate our AI innovation, said Mikita Mikado, Co-founder and CEO of PandaDoc. The company said the investment will support its next phase of AI-focused product development and team growth in Lisbon, drawing on Índico Capital Partners’ experience in artificial intelligence and scaling companies.

German neobank N26 has appointed a British executive as its new CEO, on the same day it was hit with new sanctions by the German financial regulator relating to compliance issues. Banking executive Mike Dargan will take over as CEO of N26, one of Europe’s most valuable fintechs, in April next year, replacing co-founder and co-CEO Maximilian Tayenthal and interim co-CEO Marcus Mosen. The appointment of UBS executive Dargan effectively draws to a close the leadership of Tayenthal and Valentin Stalf, who founded the challenger bank in 2013. The pair served as its co-CEOs until Stalf stood down as co-CEO earlier this year, following a reported dispute with some of N26’s investors over the handling of regulatory issues by the founders. Dargan, who has held senior roles at Merrill Lynch and Standard Chartered, will stand down from his current role as group chief operations and technology officer at UBS at the end of this year. Dargan said: “This marks the beginning of something new for me – a new opportunity, a new bank and a refreshed business model. N26 has been a pioneer in digital banking with a strong foundation and a forward-looking strategy.” Tayenthal said: “I am confident to step back and put N26 in the very capable hands of Mike. Beyond his extensive experience combining banking, technology and digital transformation, he is also fully committed to the N26 vision.” The appointment of Dargan came on the same day that N26, which has over five million customers across Europe, was hit by new sanctions by BaFin. These included N26 being banned from lending new mortgages in the Netherlands and BaFin appointing a special monitor to oversee N26's compliance activities, after the regulator found compliance shortcomings. The intervention followed previous sanctions imposed by the German regulator on N26, including a €9.2m fine in 2024 relating to late filings of suspected money laundering. Tayenthal will leave the management board at the end of the year and Stalf has moved to be a member of the N26 supervisory board.

The butterfly effect is a well-known pillar in chaos theory and its premise that the simple flapping of a butterfly’s wings in China can cause hurricanes in Texas has had huge cultural influence. It is cited in any number of films including Back To The Future and, more poignantly for this article, AI’s favourite (and lazy) touchpoint Terminator. One of my lowlights this year was attending a UK House of Lords session where the presentation continually cited Arnold Schwarzenegger's film as why the UK was leading the way in AI. More like AI for Idiots. However, there are other places outside the UK Parliament where more enlightened people are setting the standards for the future of AI (and humanity for that matter) with frameworks that will resonate far from where they were announced… not least in Europe. Europe has spent the past decade wrestling with the contradictions of its tech ambitions: yearning to lead in innovation while regulating with an iron fist, chasing Silicon Valley while side-eyeing it and insisting on “strategic autonomy” while relying on US cloud providers and Asian hardware. So when a new global declaration on AI emerges, this time via Seoul in South Korea, it’s worth asking not just what it means for global governance, but what it means for Europe that desperately wants a seat at the top table of AI rule-making. The Seoul Statement, announced last week at the inaugural International (AI standards) Summit in South Korea, is the latest in a string of international overtures attempting to corral AI into something safer, fairer and more widely beneficial. In front of a personally invited audience of 300 people, it framed AI as ‘an opportunity to advance the well-being of humanity’ and emphasised an ‘inclusive, open, sustainable, fair, safe, and secure digital future for all’. Vimal Mahendru is the IEC Vice-President and Chair of the Standardization Management Board and was in Seoul to witness the announcement. “It is always about people, about making technology work for all humanity, and not the other way around. Amidst rapid technological developments, the Seoul Statement aims to safeguard our shared future by placing the aspirations of all humans at the centre of AI governance and standards development,” he said. Europe, meet the world (finally) For once, the global AI conversation sounded very… European. The Statement stressed socio-technical contexts, how AI behaves not in the lab, but out in the wild, interacting with people, institutions and societies. It’s as if the world has been reading EU white papers on long-haul flights and decided they make sense after all. But this isn’t just a pat on the back for Europe’s regulatory evangelists. The Seoul Statement also lands at a moment when the EU is staring down the reality that rules alone won’t build AI champions. Europe is still over-indexed on governance, under-indexed on compute, and chronically under-funded compared to the US and China. Aligning with global standards matters, but only if Europe can engage from a position of technological strength, not moral superiority alone. Standards: the Brussels effect meets the Seoul effect? The Statement gives international standards pride of place, arguing that they ‘build trust, facilitate digital cooperation, enable interoperability across borders’ and strengthen regulatory collaboration. What’s interesting is that Seoul now positions standards as a tool not just for safety or accountability, but for development and inclusion. As Mahendru continues: Building on the work of the Council of Europe's framework on Artificial Intelligence, Human Rights, Democracy and the Rule of Law, the statement highlights the role that international standards play as part of a holistic approach to AI governance. “By making the connection between technical standards and human rights in this way, we have taken an important and necessary decisive step towards ensuring that transformative technologies like AI work for the good of society. Europe should pay attention to Mahendru’s words. Because while it has historically exported rules outward, AI is exposing its own internal divides: between data-rich and data-poor industries. Between countries with powerful research ecosystems such as France and Germany and those still digitising basic services in parts of Southern and Eastern Europe, between startups leveraging frontier models and SMEs still figuring out cloud migration. If the EU wants to remain relevant, it must consider standards not as a cudgel, but as connective tissue, something that keeps Europe interoperable with the rest of the world, prevents digital isolation and ensures European AI remains compatible with global markets. The multistakeholder moment The Statement emphasises building a ‘dynamic multistakeholder community’ for AI standards, one that is ‘inclusive, collaborative and consensus-based’. That may sound obvious, but it’s a subtle rebuke to the more top-down approaches emerging elsewhere. Europe has always prided itself on ‘multistakeholderism’, even if it sometimes forgets to invite stakeholders who aren’t regulators. The Seoul framing presents an opportunity for the bloc to rebalance: to ensure industry, academia, civil society, and, crucially, small companies and scale-ups have real influence in shaping how AI is governed internationally. Because while European corporations are well represented in standard-setting bodies, its startups often are not, and yet it’s the startups that will feel the effects most acutely. The global stage is shifting. Europe can’t just spectate. The Seoul Statement is part of a broader geopolitical rebalancing in AI. Leadership is no longer purely a transatlantic affair. South Korea, Japan, Singapore, the UAE and others are increasingly setting the pace, not just in technology, but in how AI is governed. Europe cannot assume that its frameworks will automatically become global defaults. It must engage deliberately, diplomatically and with humility. It must show up to standards bodies early, not late. It must ensure its safety narratives are backed by technical expertise, not just legislative brilliance. And it must invest in the infrastructure and talent that make participation credible. The bottom line The Seoul Statement echoes Europe’s values, but it also exposes Europe’s vulnerabilities. Alignment on paper is easy; influence in practice is earned. If Europe wants to remain the moral conscience of AI while also becoming a technological heavyweight, it must treat international cooperation not as a validation of what it has already done, but as a challenge to do more… and to do it faster. Because the future of AI will be shaped by those who build it, those who govern it and those who set the standards that sit between building and governing. Seoul is inviting Europe to help lead that middle ground.

The European agrifood entrepreneurial ecosystem is rapidly expanding: innovative agrifood startups are becoming increasingly numerous, particularly in sectors such as agtech, functional foods, biotech, and alternative proteins.  The European Union’s AgriFood Industrial Ecosystem generates around €603 billion in value add and supports 16 million jobs. The food and beverage industry alone accounts for €227 billion and employs 4.6 million people, with more than 99 per cent of companies being SMEs. However, the ecosystem must transform to become more sustainable and competitive. This places pressure not only on local startups and scaleup but also those seeking to enter the European market. Europe’s agrifood transformation demands innovation — and startups need a clear path to scale The pressure to reduce environmental impact, adopt digital technologies, and respond to increasingly conscious consumers is pushing the entire chain—from farmers to distributors — towards innovation. But at the same time, production costs in Europe (labour, energy, regulatory compliance) are high, posing a significant barrier for new companies seeking to scale.  Moreover, climate change adds uncertainty around agricultural productivity and the reliability of raw materials, demanding more resilient and circular models.  Another major challenge is market fragmentation, as each European country has its own regulations, distribution channels, consumption habits, and food-specific rules. For a startup aiming to sell across multiple EU markets, adapting to each context can be costly and time-consuming.  To address these barriers, the European Union and organisations such as EIT Food have launched programmes to support startups in the agrifood sector.  There is strong institutional support to strengthen European competitiveness in food innovation. EU programs and projects targeting startups and scaleups — especially those led by EIT Food — are helping drive a more sustainable, resilient and tech-enabled agrifood value chain. As well as a strong foundation for local startups, Europe represents a highly attractive market for internationalisation, as its diverse economies,  consumers, and regulations offer numerous opportunities for startups to scale beyond their local markets.  For a startup, expanding into new European markets not only means increasing sales but also gaining access to strategic partners (distributors, retailers, and companies within the food value chain),  optimising the supply chain, and diversifying risks.  In response to the opportunity for internationalisation, EIT Food has developed a tailored program called Sales Booster to help agrifood startups that are established in their home country’s entrepreneurial ecosystem and want to expand into new markets within the European region. Inside Sales Booster: EIT Food’s pathway for agrifood startups to enter new markets Sales Booster offers selected startups personalised support tailored to their growth stage, current needs, and geographical focus. Participants gain not only technical support but also recognition and visibility within the European food innovation ecosystem. The program is run through collective activities, such as online workshops and meetings with key industry players to facilitate networking, as well as one-to-one sessions.  Each startup has a dedicated Growth Advisor, expert guidance, and resources to design its expansion plan and ultimately develop an internationalisation plan for its target country or countries.  There’s also access to EIT Food’s wider network. Startups can request meetings with industry and business experts from across the ecosystem, as well as draw on support from an experienced pool of coaches. At the end of the programme, startups are expected to present a concrete action plan for their expansion and have established connections with experts across the European agrifood ecosystem.  Startups powered by EIT Food’s Sales Booster Here are some of the standout startups that have taken part in the programme. Many of them are pushing the boundaries in agrifood innovation — from biotech and agritech to functional foods.  They illustrate not only the breadth of solutions supported by Sales Booster, but also how the programme helps founders turn those innovations into solid expansion plans for new European markets. By backing these companies, EIT Food is both accelerating innovation and strengthening a more connected, resilient and sustainable agrifood system across Europe: Agritrack SA: a platform to automate post-harvest value chains, reducing losses and improving traceability.  Cynomys: IoT solutions to monitor environments (such as farms), enabling more sustainable resource management and improved animal welfare.  EarthAutomations: autonomous robots for agricultural tasks, facilitating automation even on smaller farms.  • BuzzUp: a startup that has developed a programme that “translates” the sound of bees to help farmers understand their condition and improve pollination, directly impacting productivity. Agricolus: a digital platform for precision agriculture, with satellite maps, sensors, and predictive tools to optimise crops. Rebread transforms leftover bread into valuable ingredients, promoting a circular economy in the food industry.  Image: Rebread's happy belly soda. Who can apply for EIT Food’s Sales Booster? Sales Booster targets two groups of innovative startups: agrifood ventures aligned with EIT Food’s mission areas, and non-agrifood startups whose solutions can be applied within the agrifood sector.  To be eligible, companies must be registered in one of the designated RIS countries, offer an innovative or impactful solution relevant to the agrifood industry, and aim to expand within EIT Food’s geographic footprint.  Ideal applicants are those with meaningful early sales traction, a consolidated home market, and a dedicated sales or business development team ready to scale internationally.

Tallinn-based logistics startup MyDello has raised €3.1 million in a funding round led by Icelandic venture capital firm Frumtak Ventures, with participation from existing investor Finnish early-stage venture capital firm Superhero Capital. Frumtak Ventures general partner Andri Heiðar Kristinsson and Jevgeni Kabanov, President of urban mobility company Bolt and a previous investor in MyDello, have joined the board. Freight and shipping across all transport modes account for an estimated 10–12 per cent of the global economy, with maritime shipping handling most international trade by volume, yet the sector still relies heavily on manual workflows, paper documentation, fragmented communication, and limited real-time shipment visibility. Founded by experienced logistics professionals, MyDello aims to address these challenges by digitalising international freight processes and reducing inefficiencies in global supply chains through a B2B platform for manufacturing, wholesale, and e-commerce customers that provides instant door-to-door pricing and routing across freight modes from a single inquiry, supported by agreements with 400+ carriers and partnerships including DHL, Lufthansa, Maersk, Qatar Airways, and Finnair. Businesses enter shipment details (origin, destination, dimensions, and weight) to compare quotes, book transport, and track deliveries in real time with an AI-powered delivery countdown, with a focus on complex long-distance international freight such as routes in and out of the EU, China-linked trade, and lanes between the Americas and Europe. In our coverage earlier this year, MyDello co-founder Magnus Lepasalu said sustainability is becoming a bigger priority in logistics and that the company’s platform is intended to help customers make more informed, sustainable choices, an approach it now plans to scale following the new funding. Since launching in 2021, the company has facilitated thousands of shipments and established partnerships with hundreds of carriers and industry participants. When we launched MyDello in 2021 we knew there was a better way for our industry to operate and in just a few years we have facilitated thousands of shipments, built up a deep customer base, and struck partnerships and agreements with hundreds of key carriers and industry peers. However, this is just the start, and we’re overjoyed to welcome Frumtak Ventures, and we thank them and our returning investors for their support, adds MyDello co-founder and CEO Joel Timm. The company reports onboarding 12,500 businesses from 110 countries and currently operates across 12 countries in Europe and China, with an aim to expand across Europe by 2027. The investment will be used to accelerate international expansion, starting with the UK, where the platform is expected to be available to customers from December. MyDello also plans to further integrate AI into its systems, with the goal of automating most shipment operations by the end of 2026.

The European platform for deploying safe and governed agentic workflows in healthcare, Mindoo, has secured €5 million in seed financing from 6DC, Syndicate One and a group of strategic angel investors. Healthcare organisations are increasingly balancing limited staffing with growing demand for care, leaving some workflows unfinished or not recorded because teams do not have the capacity to manage them. Mindoo addresses this by providing configurable AI agents that hospital teams use to handle structured intake, documentation drafting, follow-up interactions and front-desk communication within a single platform. The company currently offers four core agents that hospitals can adapt to their own protocols, languages, and speciality workflows: a receptionist agent for routine patient communication and registration, a pre-visit agent for structured intake and medical history, a scribe agent for drafting notes, letters, and orders, and a follow-up agent for post-visit communication and care pathways. We learned very quickly that workflows in healthcare cannot be adapted to a product. The product has to adapt to existing workflows. That is why Mindoo lets hospital teams configure and run their own agents, so automation fits naturally into how they already work, explained Gauthier Willemse, CEO and co-founder of Mindoo. Mindoo is currently deployed in hospitals in Belgium and Germany and is designed to integrate with modern EHR systems. The company plans to expand into the Netherlands and France as additional reference sites become operational. The investment supports Mindoo’s plan to provide hospitals and practices with a scalable AI workforce layer that can take on routine tasks, ease pressure on clinical teams, and keep organisations in control of their workflows. The funding will be used to develop the platform further, bring its four core agents to production readiness across multiple specialities, and expand the team across engineering, clinical, and deployment functions.

Lucis, a French startup focused on expanding access to preventive health testing, has closed an $8.5 million seed round led by General Catalyst, with participation from Y Combinator, Kima Ventures, Motier Ventures, Circle.Co, and North South Ventures. Founded by Maxime Berthelot, Baptiste Debever, and Max Guerois, Lucis offers a platform that translates blood test results into a set of indicators tracked every six to 12 months. It provides a structured view of more than 180 biomarkers, covering areas such as cardiometabolic health, hormones, inflammation, liver and kidney function, and selected micronutrients, with the aim of helping users identify potential imbalances and early signals associated with chronic disease. The company positions Lucis as a preventive tool to support understanding and monitoring of health, and notes that it does not replace medical consultation, diagnosis, or prescribing. Our mission is to empower individuals to take ownership of their health by giving them a clear view of key indicators before symptoms even appear, and providing practical guidance to improve their well-being, without replacing physicians, said Maxime Berthelot, CEO and co-founder of Lucis. Members complete check-ups through certified medical biology laboratories across Europe. Results are reviewed by a multidisciplinary medical team, with AI used to help surface key signals and track changes over time. Instead of receiving a static report, users access a dashboard and a prioritised action plan organised around five areas: nutrition, supplements, physical activity, sleep and recovery, and mental health. Lucis positions this model as an extension of broader consumer health tracking, adding biological markers and clinical oversight to support a more structured, public health–aligned approach. The new funding will be used to accelerate rollout in France, the UK, Ireland, and Portugal and into additional markets, expand its network of partner laboratories and clinicians, and further develop its AI-enabled preventive analysis and support platform.