Prague-based energy scaleup Delta Green has secured €2 million in fresh funding from Credo Ventures, Tilia Impact Ventures, and Purple Ventures to accelerate its pan-European expansion.  Delta Green aims to turn ordinary European homes into a powerful force on the energy market. The company's platform connects ordinary homes into a virtual power battery, enabling households to shift consumption, discharge batteries, and export rooftop solar at times of peak demand. Every kilowatt activated at the right moment lowers bills, supports renewables, and reduces fossil fuel reliance. Delta Green is co-owned by four engineers and developers, David Brozik, Prokop Cech, Lukas Benes and Jan Hicl. I spoke to Delta Green co-founder Jan Hicl to learn more. Delta Green was originally called Nano Green and has been around for about 13 years.  Why the energy industry's conservatism is creating opportunity According to Hicl, "We acquired and rebranded it three years ago to focus on what we do now." "Our background is in building software startups, so we've turned Delta Green into a software-first company for the energy sector — which is traditionally quite conservative and slow to change." According to Hicl, a good example of that conservatism is the growing number of negative price hours on the spot market.  "Everyone knows how to fix it — by shifting when electricity is produced and consumed — but very few are acting fast enough.  The fact that those hours are still increasing shows just how rigid the industry remains." Hicl shared that this is why "flexibility' has become such an energy buzzword: everyone recognises it's critical for the transition, but real-world implementation is still limited." We've already seen small blackouts in the Czech Republic, Spain, and Portugal. However, without flexibility to stabilise the network, especially at the low-voltage level — the local networks closest to households.  Hicl contends that most investment goes into high-voltage transmission networks and utility-scale batteries, "but the real fragility is in local distribution systems." "These networks were built decades ago for completely different consumption patterns. If five EVs on one street start charging simultaneously, you can overload the grid.  That's where residential flexibility — our speciality — becomes essential. Our platform can coordinate devices, telling some to wait and others to charge, so the network doesn't collapse." How ending subsidies is actually helping Delta Green grow Greentech is a tough sector for many startups at the whim of European policy priorities, especially in areas like EV-subsidies and ESG. But interestingly, Hicl asserts that the fact that many countries are ending net-metering schemes—where households could feed excess solar power into the grid in summer and draw it back in winter, is helping the company.  He admits, "That was effectively a form of subsidy, and now it's disappearing." However, the end of subsidies is actually prompting both consumers and retailers to consider energy management more seriously.  "As a result, consumers are starting to ask new questions: When should I consume energy? How can I use what I produce more intelligently?  Retailers, too, have lost many of their government incentives, so they're looking for new ways to stay competitive. Our technology gives them that edge. While first movers benefit most from the shift, it's more difficult in markets where there aren't enough EVs, heat pumps, or residential batteries—" assets we can control. But overall, the shift away from subsidies is creating stronger demand for what we do." Further, household flexibility is no longer in a pilot phase — "our business growth proves it works across thousands of households," says Hicl. "We're now accelerating our plans as market demand is clear, with major retailers across Europe approaching us to deploy our solution." From test lab to pan-European platform In the Czech Republic, Delta Green is a small energy supplier, serving about 7,000 customers.  "That's our test lab, where we can develop and refine our technology before scaling it," shared Hicl. "Then we license our platform to large utilities. They use it to manage flexibility across their customer base." Delta Green runs a lean ship, which Hicl admits is partly out of necessity.: "Margins in the energy sector are extremely thin — around 1 per cent. When we entered the industry, we were shocked to find that no one accepted credit card payments. Then we discovered that card processing fees are about 1.5 per cent, which would instantly wipe out profits! We've managed to stay profitable because we utilise our proprietary technology to operate efficiently.  We make money from our customer portfolio and from selling the platform to other suppliers. We're sustainable, even if we're not the fastest-growing company out there — and that's okay. A lot of hyper-growth players in our space have already burned out." Hicl asserts that for decades, we've been accustomed to having electricity available whenever we want it, at a constant price. That era is ending. "In the future, prices will vary hour by hour depending on supply and demand. Sometimes electricity will even be free, or have negative prices. For households, that means comfort will come at a cost. If you insist on consuming energy during peak times, you'll pay more. But if you're flexible — if you shift your consumption, like charging your EV or heating water when solar or wind generation is high — you'll save money and help stabilise the grid." E.ON partnership sparks snowball effect across Europe By the end of this year, Delta Green platform will manage tens of thousands of households, putting it on track to become Europe's largest household-based virtual power plant.  "We just closed first deals in Romania, Italy, Hungary and Slovakia, and we are in advanced discussions with several major international energy suppliers about adopting our technology," shared Hicl.  Since its last funding update in 2024, Delta Green has made significant strides. It has become the first company in CEE to successfully involve households directly in grid balancing.  In the Czech Republic, one of country's top three energy providers E.ON has integrated Delta Green's flexibility service into its operations and plans to scale it to thousands of customers within the next 18 months. Thanks to this collaboration, E.ON customers can earn up to €200 per year by monetising their household energy flexibility.  Hicl admits that the momentum has created a snowball effect — "everyone realised they needed something similar to stay competitive. That's when we knew we were ready to expand internationally." "Now we're active in Austria, Romania, and Hungary, and preparing to launch in other European countries." From here on, Delta Green is expanding the types of assets it can control. "Right now, we mainly manage home batteries, but we're adding heat pumps and EVs to the platform." "We are really happy to see that Delta Green's timing for their flexibility aggregation solution was perfect. Just as the market realised the importance of this kind of solution for energy transformation towards greater sustainability, the company already had a real solution in place that the energy companies can instantly use, because it's been tested on Delta Green's own retail customers," says Tilia Impact Ventures partner, Pavel Petřek.  Delta Green's mission is clear — to place households at the very centre of Europe's energy transition and build the continent's largest network of flexible homes.

London-based Orbiri has raised £320,000 in an oversubscribed angel round with 14 individual investors, following an initial £45,000 in seed funding. The company is preparing to launch a community-wide solution that aligns children, parents, and schools around shared digital limits (shifting peer pressure from a barrier to a support) instead of relying on parents to police screen time individually or enforce blanket smartphone bans. Orbiri provides a community-powered platform with preset, considered screen-time boundaries for children. Its collective-action approach aligns children, parents, and schools around shared boundaries to eliminate daily device battles and protect the conditions necessary for healthy childhood development. Rather than leaving schools and families to act alone, it enables coordinated limits within a single framework, aiming to make healthy digital habits the norm. Amid growing concern about children’s unrestricted smartphone use, with schools in England adopting phone bans and ministers considering Australia-style limits on under-16s’ social media, Orbiri contends that durable progress depends on bottom-up community coordination rather than top-down restrictions. Set to roll out next year, Orbiri will complete product development, secure compliance certifications, and run trials with early-adopter schools while expanding its core and product teams. The goal is to demonstrate that shared community boundaries can replace daily device conflicts with a more natural shift toward healthier technology use, addressing shortcomings of school bans and individual parental controls. The funding will support upcoming trial phases ahead of a broader launch, aligning with a growing view that approaches such as school phone bans and individual parental controls are insufficient and difficult to implement at scale.

SLNG.ai, the speech infrastructure startup building the first radically global platform for voice AI, today announced it has raised €3.3 million in Pre-Seed funding to challenge the US-centric voice technology ecosystem that has left billions of users worldwide in digital silence. The round, led by Earlybird VC, will accelerate SLNG's mission to restore the universal power of voice by providing developers, startups, and enterprises with a unified platform that integrates multiple speech models, supports true global deployment with regional compliance, and eliminates the barriers that have kept voice technology concentrated in a handful of languages and markets. Spain-based SLNG was founded by Ismael Ordaz and Luke Miller after spotting the same pattern globally: companies struggling with 1) lack of regional compute, 2) missing industry-specific models, 3) compliance barriers, and 4) latency issues. SLNG solves all four.  According to Luke Miller, CEO and co-founder of SLNG: “We're not just building another voice AI platform, we are rethinking the entire ecosystem to be developer first, truly global and compliant." SLNG aims to solve fundamental limitations of the current voice AI infrastructure. Most platforms optimise primarily for premium languages like English, Spanish, and French, leaving other languages/dialects as afterthoughts. High latency and data residency issues entangle deployment in many non-US-centric regions. Developers are forced into single-provider ecosystems with limited flexibility in terms of cost and model testing. Lastly, regulated industries can't deploy voice AI products due to data sovereignty requirements that existing platforms simply can't meet. Unlike traditional speech AI providers, SLNG offers a model-agnostic platform where developers integrate any speech-to-text, text-to-speech, or voice cloning model through a single API. This fuels a truly global deployment through regional infrastructure and ensures low latency and compliance worldwide. The company's open ecosystem approach lets developers combine open-source and proprietary models without vendor lock-in, while developer-first design delivers SDKs and APIs built for rapid integration. This avoids months-long implementations, giving developers unprecedented flexibility in building voice-enabled applications. The platform already supports deployment across over 20 regions, enabling companies in regulated industries, like healthcare, finance, and government, to finally deploy voice AI while meeting strict data residency requirements. “What excites us about SLNG is the audacity of their mission: to make voice AI universal, compliant, and developer-first from day one. Luke and the team are not just building a product - they’re building the missing infrastructure for a truly global voice ecosystem. We believe SLNG can define the next era of speech technology, “ added Akash Bajwa, Principal at Earlybird VC.

Christopher Savoie, one of the original inventors behind the AI system that led to Apple’s Siri, has raised over €2M for SiC Systems, a startup combining agentic AI and quantum sensing technologies for use in critical industrial and defence applications. The full amount raised has not been disclosed. The company is a spinout from the Technical University of Denmark (DTU). The round was led by Dutch quantum investment fund QDNL Participations, with additional backing from Propagator Ventures, Plug and Play, and Wavepeak Ventures. SiC Systems aims to develop intelligent, adaptive AI agents capable of orchestrating complex physical systems in real time. Its proprietary platform integrates classical and quantum sensors with agent-based AI and generative models to support decision-making in high-stakes sectors such as biomanufacturing, industrial automation, and defence. Savoie, who co-founded the company with DTU professor Seyed Soheil Monsouri, said: "From the early multiagent natural language interfaces that brought Siri to life to today's agentic AI for the physical worlds of industrial and defense applications, my personal AI journey has always been about making intelligence actionable." QDNL Participations Investment Director Kris Kaczmarek said: "SiC's fusion of agentic AI and quantum innovation represents a leap forward for mission-critical applications." Founded across Copenhagen and Nashville, SiC Systems builds on DTU research in model-based tools for dynamic, multi-scale systems. The company’s quantum-inspired virtual sensors aim to reduce reliance on costly hardware by predicting physical states with high precision, even in harsh conditions. The startup plans to use the funding to scale its platform, form strategic partnerships, and pilot deployments across the chemical, biological manufacturing, energy, and defence sectors.

Oura, the Finnish health tech startup known for its popular smart rings, has raised over $900m in a new funding round, valuing the startup at “approximately $11bn”, it said. The funding round was led by Fidelity Management & Research Company with participation from new investor ICONIQ and investment from Whale Rock and Atreides. Oura's valuation has now more than doubled since its $200m last funding round in December last year, valuing it at $5.2bn.Oura says it will funnel the new funding to speed up AI and product innovation, expand global distribution and invest in the development of new health features.Oura says it has sold more than 5.5 million Oura rings since they were launched in 2015, with more than half of those sales in the past year.The startup says it recorded revenues of more than $500m in 2024, and says it’s on track to top $1bn in sales for 2025.“This new funding is a testament to the strength of Oura’s business and the trust millions of members place in us every day,” said Tom Hale, chief executive officer, Oura.“We’re proud to be building not just a product, but a global movement toward proactive health—helping people understand their bodies, make better lifestyle decisions, and connect more effectively with their healthcare providers. "Today, our technology supports consumers, employers, insurers, and clinicians working together to advance preventive health at scale. With this investment, we will accelerate innovation, expand our global reach, and set a new standard for what wearables can achieve in advancing preventive health.”David Shuman, Oura board chair, said: "Oura’s growth and impact over the past year have been truly remarkable.“The company has doubled revenue, expanded its global reach, and delivered transformative solutions that are reshaping the wearables industry."

Strawberry, an agentic browser with built-in, personalised AI companions, has raised $6 million led by General Catalyst and EQT Ventures, with participation from founders of Lovable, Supabase, and Hugging Face. While enterprises often deploy AI via large budgets and multi-year programs, freelancers, startups, and small businesses need tools that work immediately. Many existing solutions require technical expertise, disrupt workflows, or sit in standalone interfaces outside the browser. Strawberry embeds AI directly into browsing to create a no-code, adaptive workspace that automates routine digital work so people can focus on higher-value tasks. Founded in Stockholm by a technical team focused on bringing AI beyond developers, Strawberry streamlines tab switching, data movement across tools, spreadsheet population, and message drafting, while adapting to each user’s context and workflows. Early users have built custom assistants for competitive intelligence and sales prospecting, earning Product Hunt “Product of the Day” and “Product of the Week.” The platform signals a shift from generic enterprise AI toward personalised companions that make everyday work more collaborative and usable. Charles Maddock, CEO and co-founder of Strawberry, commented:  The browser is a tool most people use daily and know intimately. By giving Strawberry powerful AI features out of the box, we're making AI automation accessible to everyone. This funding helps us continue our mission to help modern workers love what they do and skip their busywork. The financing will fuel Strawberry's expansion across engineering and design to support rapid iteration with its growing beta community. Alongside the financing, the company is rolling out an upgraded version with new features, offering early access to select customers and expanding onboarding monthly.

Stockholm-based Epiminds, which develops multi-agent AI systems to run marketing end-to-end, emerged from stealth with $6.6 million in funding. The funding was led by Lightspeed Venture Partners with participation from EWOR, Entourage, and high-profile angels, including the former CMO of Booking.com. Agencies face pressure from both clients and operations as clients expect greater transparency, faster reporting, and measurable ROI on tighter budgets, while internally, fragmented data slows decisions, and AI adoption is uncertain. Traditional responses like hiring more specialists, adding dashboards, or reacting after problems arise raise costs and complexity without fixing core inefficiencies or preparing for the future. Epiminds addresses these issues. Founded in 2025 by Google and Spotify alums, Elias Malm and Mo Elkhidir, respectively, the company builds multi-agent AI systems that agencies can train and evolve. Its core product is Lucy, an AI marketing manager coordinating over 20 specialised agents across reporting, optimisation, budget pacing, bidding, and creative. Agencies can onboard a client in under 30 seconds and immediately deploy an AI team to run campaigns end-to-end. Lucy and team not only surface insights but execute them, learn each agency’s playbooks, and proactively monitor accounts to flag risks before performance declines. According to Mo Elkhidir, marketers are increasingly expected to achieve more with fewer resources: Lucy and her team take on the busywork so that marketing talent can do their best work. This is not about replacing creativity; it’s about giving it room to flourish. Agencies using Epiminds report faster onboarding, improved performance, reduced wasted spend, and more time for creative and strategic work. The multi-agent system manages routine tasks such as reporting and pacing, as well as audits, creative analysis, competitive insights, and strategic planning. By linking insights to execution across platforms, Lucy can increase output without additional headcount. The product targets a gap in the market. Legacy dashboards and optimisers are siloed and manual, while point AI tools address narrow problems without coordination. Epiminds’ multi-agent approach provides an integrated, adaptive system that learns and improves over time. Elias Malm added: Our vision is simple. We're building Epiminds because we see where marketing is headed. The future is about dynamic, agentic teams that analyse, plan, execute, and improve in real time. Every marketer should have access to a 24/7 AI workforce that frees up their time for creativity and strategy. Our goal is to give them that future, today. Looking ahead, Epiminds plans to expand Lucy’s capabilities across more integrations, increase the level of autonomy, and self-improving capabilities. Each new feature strengthens the entire system, creating a network effect where every agency benefits from smarter, more capable AI.

London-based Clove, a new challenger in personal finance and wealth management, has emerged from stealth with a $14 million pre-seed round led by Accel. The round also saw participation from Kindred Capital, Air Street Capital, and angels, including Barney Hussey-Yeo (Cleo), Patrick Pichette (via Inovia Capital), Erez Mathan (GoCardless), and Gideon Valkin (ex-Monzo, ClearScore). The FCA estimates that people who receive financial advice may be up to 10 per cent wealthier in subsequent years than those who do not. To narrow the advice gap and support market growth, it has introduced regulatory changes to encourage investment and innovation. In the UK, 13 million mass-affluent individuals hold £3.8 trillion in investable assets, including more than 3.7 million open to professional advice with over £50,000 to invest; a further 7 million adults with £10,000+ in cash savings may be missing out on long-term investing benefits. Clove is building a financial institution for how people live and work today, aiming to make money management and advice accessible, affordable, and personal. The company combines human advisers with AI to scale high-quality, personalised guidance and automate administrative work, helping to close the advice gap. Co-founder and CPO Alex Loizou said everyone should have access to high-quality financial advice. He added that Clove aims to reshape the economics of advice by combining human expertise with AI to make planning more accessible, affordable, and effective for young professionals, entrepreneurs, growing families, and those beginning to plan for retirement. Clove will use the pre-seed funding to expand its team and platform ahead of a planned full launch in 2026, subject to FCA authorisation. 

Sometimes the most compelling startup ideas are born from simple, everyday problems that demand a tech-first fix. Take two UK founders who bootstrapped through COVID and are now the global leader in the highly competitive space of luggage storage. UK startup Stasher. During COVID, they built systems that now onboard hundreds of locations monthly without meetings.  They've expanded from 1,000 to over 8,000 locations across more than 1,000 cities and 80 countries, and hit profitability during the summer and 100 per cent YoY growth, while their US competitors (like Bounce, backed by a16z) burn through much more cash. I I spoke with founders Jacob Wedderburn-Day and Anthony Collias to learn how they did it, and ultimately how UK founders can still build category leaders without following the Silicon Valley playbook.  From student flat to global platform The business started when students kept knocking on the co-founder's King's Cross door, asking to store bags for the summer holidays.  According to Collias, after uni, he was living between Euston and King’s Cross, and Wedderburn-Day was still at UCL, which had a startup programme.  They’d both realised they didn’t really want to go into corporate graduate schemes — Wedderburn-Day had done a stint with the government and had a job offer from the Bank of England.  Collias recalled:  “I’d done an investment-banking internship and hated it — long hours, smart people doing boring work. At the time, I was living with my brother, who was at City University.  Between us and our friends, people were always asking, ‘Can I leave my stuff at yours for a couple of weeks?’ You know, getting kicked out of student halls, going travelling, that sort of thing. Jake would often leave his bag with me, too.”  The duo decided to put the idea in the UCL’s Hatchery startup incubator program.  Refining the model: from homes to hotels The duo took some time refining their idea. Wedderburn-Day shared: “At the beginning, it was kind of 'Airbnb for storage.’ You could store your stuff in someone’s house for a day, a week, or a month. But we soon realised that was a bad idea — too much friction for what you could charge. We also realised everyone was using it as left luggage, not self-storage. They just wanted to leave their bags for a few hours or a day, not months. We didn’t know how to price one day versus one month — they were completely different things.” So they narrowed it down to same-day storage, and saw it made more sense to work with local businesses.  According to Wedderburn-Day, “they were already open, had spare space, and liked having people come in.  “That’s how the business model evolved — from 'store anything anywhere' to a clear left-luggage service run through shops and hotels.” Image: Stasher storage location. Photo: uncredited. Scaling through marketing and digital infrastructure Left-luggage offerings are nothing new - walk through any tourist precinct and you’ll see stores offering left luggage services. But according to Collias, “a lot of those standalone locker stores — especially in places like Barcelona or Madrid — are becoming oversaturated." "Ten or twenty years ago, if you opened the first few, you made a killing. But now, in some cities, there are 40 or 50 of them, and the economics don’t add up. Most of those operators aren’t experts in online marketing, and it’s not worth running ads for just one or two locations." Stasher has developed a platform that connects travellers with trusted storage spots — including hotels, shops, and lockers — across cities worldwide, allowing them to drop off their luggage for a few hours or several days. One thing they discovered is that 90 per cent of luggage storage is booked on the same day, making it on-demand urban logistics rather than planned infrastructure. With Stasher, users can easily search, book, and pay online for a convenient storage location for specific dates and times through the website. Once booked, travellers simply “stash” their bags at the chosen location and collect them later.  In addition to short-term storage, Stasher also offers a luggage shipping service, enabling users to send their bags directly to their next destination. Collias contendst: “We can solve a lot of the problems those small storage shops face. They have the physical presence; we have the digital infrastructure. Combine the two and you get real synergy.” Further, he asserts that platforms like Slasher have the advantage because we can spread the marketing costs across hundreds of sites. For customers, that means more choice — they can compare all the nearby options instead of relying on chance.” Winning in a crowded market Wedderburn-Day shared that to win in a crowded market, “You have to nail three things: product, price, and distribution.” “Product is tricky because you’re working with fragmented partners — shops, hotels, different setups — but smart lockers are helping with consistency. We also like hotels because they’re used to handling guests’ belongings professionally. Our pricing strategy aims to be the best-value option. Lockers have to maximise revenue per square metre, but we can offer lower prices and still profit thanks to scale.” Data-driven expansion Stasher is now active in about a thousand towns across cities globally. Early on, it picked places based on search volumes, tourism, and common sense — “like, obviously, downtown New York will have demand.” But another metric is Airbnb density. According to Wedderburn-Day, numerous short-term rentals result in a large number of bags. With its own data, the team can now base expansion on where people are searching. He shared: “We get thousands of bookings and tens of thousands of search queries a day, so we can see demand patterns in real time. If a location has lots of searches but few conversions, we know we need more or better supply there. It gives us a really rich picture of global demand, and helps us grow intelligently rather than guessing.” In terms of customer acquisition, Stasher’s biggest channels are partnerships and search. According to Collias, partnerships make perfect sense because travel is time-specific. “Think Airbnb hosts whose guests need storage before or after check-in. They already have the customers who need us.” Search – a mix of organic and paid — is the other big one. And when people search “luggage storage,” It’s a location-based query, so Maps is crucial too. In terms of revenue, hosts charge a fixed fee per bag per day. It varies by geography and partnership type, but for most major partners, it’s roughly a 50/50 revenue share. Wedderburn-Day  asserts that the economics are solid:  “We’ve been profitable for the last few months. We do have seasonal fluctuations — summer is strong, winter is slower — but the team's global reach means it's always summer somewhere, and overall, the business is safe and cash-flow positive. The unit economics are very healthy. Growth just depends on how aggressively we choose to reinvest.” Flywheel style growth In terms of growth, Wedderburn-Day muses that “People assume you just pour money into ads, but that’s not how it works." Our growth is more like a flywheel: A bit more demand lets us add more supply; more supply attracts more demand. Hosts generate visibility — more pins on Google Maps, more signage, more reviews — which boosts traffic. That improves our economics, so we can invest more in marketing. It’s a conventional marketplace loop: demand → supply → demand. The bigger you get, the better your product and conversion, so you can scale without burning cash.” From startup to scaleup Now Stasher is at a turning point — shifting from startup to established company.  Collias asserts, “When we talk about raising money now, it’s not about survival; it’s about how we can use capital to grow intelligently. There’s still huge room for expansion.” Once you know where it makes sense to invest, you can shift from a variable-cost to a fixed-cost model in key areas, which is a powerful way to scale a marketplace sustainably. Significantly, Stasher achieved its goals with less than £5 million in funding — partly by learning how to do "more with less". The duo cold-emailed their way from there to their first investor (Big Yellow Storage CEO), securing their first game-changing partnership with Premier Inn shortly after. “Too many founders build to exit” Ultimately, Collias believes that Europe needs to stop selling itself short.  “Too many founders build to exit. We need bigger swings and the patience to stick around.” Wedderburn-Day asserts that US companies have easier access to capital and higher valuations, this makes it easier for them to acquire European startups. “Take Airbnb. One of the teams they acquired in Europe had incredibly capable founders. One of them stayed on as Head of Growth for years. Clearly, he could’ve built something great independently. But the ecosystem didn’t make it worthwhile to stay independent.” Collias believes: “We need to make it make sense for founders to build long-term European champions, not sell early because of the system." In many ways, Stasher’s journey offers an example to many startups. With just £2.5 million in total funding, the company built a profitable, global business in a fiercely competitive market — not by outspending rivals, but by outthinking them. It’s proof that capital efficiency can outperform capital abundance, and make it possible to turn a simple student idea into a worldwide category leader. Lead image: Stasher co-founders Anthony Collias and Jacob Wedderburn-Day. Photo: uncredited.

Under the motto “Connecting Europe”, Bits & Pretzels once again underscored its ambition to be the central platform for founders, investors, and decision-makers in Europe. Coinciding with Oktoberfest, Munich became a hotspot for innovation, networking, and entrepreneurial spirit – with over 7,500 participants from more than 70 countries, top-tier speakers from around the world, and a clear message: Europe must come closer together, become bolder, and harness its innovative power with greater determination. Europe needs more ambition, risk culture, and open markets International voices shaped the debate on the future of the European startup ecosystem. “Connecting Europe was more than a motto for us – it was a promise. This year’s Bits & Pretzels has shown what’s possible when Europe’s founders work together on a shared vision,” explained Andy Bruckschlögl, co-founder of Bits & Pretzels. Skype founder and Atomico CEO Niklas Zennström called the founders the gamechangers of Europe and urged for more courage, independence, and assertiveness. Raycho Raychev, founder and CEO of space startup EnduroSat, advocated strengthening democracy through bold technologies rather than angry tweets. He also called for open market structure: “We still do not operate in a truly open market in Europe. Instead of focusing on transparent competition, many players remain stuck in outdated management structures that slow down innovation and progress.” The time for founders and bold technologies is now Despite ongoing economic uncertainties and global tensions, one thing was evident this year: Europe is growing closer together. Germany’s Federal Minister for Economic Affairs and Energy, Katherina Reiche, emphasized that crises offer great opportunities for startups, as the pressure for change on established companies increases and innovation becomes the key to sustainable success. “Bits & Pretzels is, more than ever, the platform for European innovation. The feedback from the community confirms it: Europe wants to – and, even more importantly – Europe can! And we will do everything we can as a platform to create even more space for collective progress,” said Bernd Storm van’s Gravesande, co-founder of Bits & Pretzels. Matchmaking with impact: How Bits & Pretzels connects founders and investors Formats such as the Startup Expo, CIO AI Summit and Investor Summit made European collaboration tangible at Bits & Pretzels. At the Startup Expo, more than 200 startups from across Europe showcased their innovative ideas and solutions. In the Matchmaking Area, founders and investors met in curated one-on-one sessions, supported by experienced networkers who acted as personal guides and fostered valuable connections. “Our mission is to build bridges – between countries, between capital and ideas, between startups, investors, and corporates,” said Felix Haas, co-founder of Bits & Pretzels. “These bridges were visible everywhere at Bits & Pretzels 2025.” European Pitch Contest: Sitegeist wins with a smart construction platform A major highlight was the final round of the European Pitch Contest. The first prize went to Munich-based robotics startup Sitegeist, which has developed an intelligent platform that automates repetitive construction site tasks – improving safety and efficiency. Founder Lena Pätzmann celebrated the win: “Our mission is clear: we want to save infrastructure. Now we have the attention, and we’re ready to get things done.” For more information, visit the official website: www.bitsandpretzels.com

The Serbia-born CEO of Nvidia-backed UK AI voice startup PolyAI has a nice way with words. The US is the “new Rome” which “demands loyalty”, says Nikola Mrkšić; meeting his wife during a spell interning at Credit Suisse generated “very high ROI activity”; while PolyAI’s client base is “heterogeneous”. Mrkšić is talking to Tech.eu in PolyAI’s commodious and new-looking basement offices (albeit sparsely occupied as it’s Friday and WFH Friday mania persists in the UK) in London. PolyAI, which was founded in 2017 and valued at nearly $500m, is often cited in top ten lists of UK AI startups, a triumph which was given a boost last month when Nvidia’s man of the moment CEO Jensen Huang named it as one of its eight star startups (along with the likes of Wayve and Revolut), it would be investing in. Serbian heritage A namecheck by arguably the most famous businessman on the planet might have given the 30-something co-founder and CEO a moment to reflect on how his life had changed since his 1990s childhood growing up in troubled Serbia, in the aftermath of the breakup of Yugoslavia, its parent nation. Mrkšić, who is slight in stature and softly spoken, says: “Serbia has had to find its own way after stumbling for quite a long time, which meant for my parents’ generation, it was a very like jaded feeling. “And I think there was a reflex to push your kids into hardcore education so that they can rely on themselves and maybe be less dependent on the system, which had crumbled in front of their eyes.” Growing up in Serbia, Mrkšić attended a Serbian maths high school (the equivalent of a UK grammar school) and was one of seven out of eight applicants who received a full scholarship to Cambridge University. “We’re not designed to crack Cambridge interviews the way that many elite schools in Britain are”, he wryly points out. His links with Serbia, where PolyAI has an office, are still strong, returning three or four times a year to visit friends. Cambridge University Serbia might be his home country, but London is where PolyAI was founded, seven years after Mrkšić came to the UK to study computer science and maths at Cambridge University. He could have quite easily pursued a banking career, interning at Credit Suisse, where he met his wife during his undergraduate years. But an encounter with Blaise Thomson, the founder of speech-related AI startup VocalIQ, a Cambridge University spinout, convinced him to join the startup as its first employee. His time at VocalIQ, which went on to be acquired by Apple 18 months later, also lit the fuse for him wanting to launch his own startup, at a time when deep learning was gathering momentum. During this time, Mrkšić also met his Taiwanese co-founders: Tsung-Hsien Wen (ex-Google), PolyAI’s CTO, and Pei-Hao Su (ex-Facebook), PolyAI’s SVP, engineering, at a Cambridge University group focused on spoken dialogue systems. He says: “We were in high demand by the research labs. What frustrated us is that we didn’t really see automated voice improving.” What PolyAI does? PolyAI has developed AI voice assistants for call centres which guide customers through enquiries, handling millions of calls, which can, some say, sound indistinguishable from human voices. PolyAI has worked with linguists to build voice assistants that reflect human speech patterns and the voices can be tailored by accent, tone and vocabulary. The tech can complete many tasks a customer service rep can, including taking payment information as well as names, addresses and account numbers. PolyAI uses its own proprietary large language models as well as frontier model companies like OpenAI and DeepSeek. The early years of PolyAI, pre ChatGPT, were tough, says Mrkšić, given they were three researchers-cum-callow-entrepreneurs trying to commercialise a product that few in the business world knew about. Back then, the scrappy startup, looking for commercial traction, speculatively approached London pub owners to experiment with its tech. Competition and challenges While PolyAI is making headway, automated call handling remains a turn-off for some people, partially due to historic bad experiences with rudimentary voice tech. Furthermore, complex requests often require empathy and judgment that only humans can provide, experts say. A McKinsey survey found that 71 per cent of Gen Z respondents (rising to 94 per cent for baby boomers) believe live, human calls are the quickest and easiest way to reach customer care and explain their issues. Mrkšić has previously spoken about these challenges in the Tech.eu podcast. Meanwhile, earlier this year Klarna CEO Sebastian Siemiatkowski said its AI voice approach led to “lower quality” and Klarna flipped from its AI-first policy to ensure customers always had a human to talk to. Despite these challenges, it is easy to see why tech giants such as Google, Amazon and Microsoft have tried to crack automated call handling, as the call centre software market is estimated to be worth tens of billions of dollars. Current startup competitors include the likes of Germany’s Parloa and US startup Rasa. ChatGPT moment The ChatGPT moment was a big boon for PolyAI, says Mrkšić, quadrupling inbound leads. Now, PolyAI works with major businesses, amid growing enterprise 24/7 demand for call centre support, which suffers from high attrition rates. These include Las Vegas casinos, Hilton and Marriott hotel chains, US delivery service FedEx, and the financial institution Unicredit. Given the co-founders' stellar CVs, funding has not been a problem, says Mrkšić. Last year, PolyAI raised £50m in funding, valuing it at close to $500m. The Series C was led by investors Hedesophia and NVentures, the VC arm of Nvidia, with participation from existing investors including Khosla Ventures and Point72 Ventures. It has previously raised $66m from investors. It is understood that PolyAI will announce a fresh funding round later this year, likely to include fresh funds from Hedesophia and NVentures. Relationship with co-founders “I am the talker”, says Mrkšić, when asked how the founders divvied up the roles. He says he loves all the key functions of being CEO: fundraising, hiring and selling. That said, he admits his management skills are a work in progress. The three co-founders, he says, are still happily professionally married. “We are really close,” he says, but points out that there are clear demarcation lines. For example, he says it was decided from day one that Mrkšić, a tech aficionado, would not write a line of code or voice any strong technical opinions. He says: “I am deeply technical. I still very much enjoy the moments where I have time with product and tech teams to talk about things.” In between running PolyAI, he also hosts an AI-infused podcast, discussing news and AI breakthroughs. US footprint Mrkšić has adopted a kind of second identity, as an American, as PolyAI’s revenues have grown in the US, which now accounts for roughly 80 per cent of its revenues, compared to 20 per cent in Europe. He calls the US the “new Rome” which “demands loyalty”. He says: “If you want to be successful in the US, you are going to use American spelling and work American hours.” For example, when in London, he says he works from 10am to 1am UK time to fit in with the US. He now spends a big chunk of his time in the US, where PolyAI has New York and San Francisco offices, meeting existing and new clients, meeting investors, and attending conferences. The future As AI becomes more commonplace and the frontier model becomes more effective, the tech, in theory, should improve. PolyAI, which employs around 270 people, reported revenues of £11.9m in the year ending January 31 2025, up 75 per cent on the year, according to Companies House figures. Revenues were boosted by snapping up new customers and expanding existing contracts. And what about PolyAI’s plans for the rest of the year? Mrkšić says: “We have been growing the team like mad. So everything from accelerating on that and platform changes and announcements which are going to come towards the end of the year, opening up the ecosystem a bit more.”

Italy-based GEVI, a company specialising in vertical micro-wind turbines, has closed a €2.7 million seed round led by 360 Capital and CDP Venture Capital (through the Acceleratori Fund, MiSE Co-Investment Fund, and ToscanaNext Fund), with participation from NextSTEP One. GEVI develops a modular, distributed, low-noise micro wind turbine with a smart vertical axis that adapts to wind conditions in real time via AI-driven blade control. Designed to overcome limitations of conventional systems, the blades adjust every hundredth of a second as an AI system analyses wind data, solves fluid-dynamics models, and optimises output. The architecture targets higher efficiency (reported up to +60 per cent annual energy versus leading VAWTs), improved reliability and safety (active blade control reducing loads by up to 80 per cent in strong winds), and broad applicability thanks to a compact form factor (rotor height 3m, diameter 5.4m). Intended for household use and local micro-grids, it operates from wind speeds as low as 2.5 m/s with a nominal output of 5 kW, and is designed to work alongside solar in urban, industrial, and rural settings. According to Emanuele Luzzati, founder, CEO, and head of engineering, the funding enables GEVI to bring a distributed, modular wind solution to locations where photovoltaics and conventional wind are less effective, accelerating the transition from prototype to scalable industrial product. The support we have received allows us to transform our technology into a scalable, industrial product, accelerating GEVI’s contribution to the energy transition. Thanks to the trust of our investors, we will be able to give strength to a radical idea: turbines that learn from the wind and interact with nature, transforming into energy what was previously dispersed. Founded in 2022 by Emanuele Luzzati (CEO & Head of Engineering), Edoardo Simonelli (Head of Products), and Soufiane Essakhi (Head of Operations), GEVI operates a ten-person team across an R&D office in Pisa and a commercial/operations office in Rome. Alongside the round’s close, Giuseppe Imburgia was appointed General Manager to lead the scale-up with the founders, driving industrialisation and international growth. The funding will be used to industrialise the product, including starting serial turbine production, and to advance technology by enhancing the AI control system and introducing new turbine versions.

Retail iPaaS provider Patchworks has closed a £5 million round to support its next phase of growth. The round was led by Gresham House Ventures, with growth lending from Palatine Growth Credit. Retailers often operate fragmented systems that don’t communicate, leading to overselling, delayed or lost orders, peak-period bottlenecks, and heavy manual data entry across e-commerce, warehouses, and finance. These issues waste staff time and frustrate customers. Patchworks addresses this by connecting e-commerce, ERP, POS, PIM, CRM, CDP, and fulfilment, so stock, orders, and customer data stay in sync. Prebuilt connectors and real-time flows help retailers adapt to market shifts, reduce peak failures, and resolve issues faster. Teams can pinpoint bottlenecks (e.g., where an order is stuck) and trigger actions without code, while a broad partner network across ERPs, warehouses, and couriers supports reliable delivery. Growth is powered by a partner-first model, often described as its “flywheel”, that incentivises technology and agency partners to refer and deliver at scale, providing global coverage and ensuring the product evolves with current retail tech needs. As enterprises move off rigid legacy systems, Patchworks is becoming a preferred option for connecting commerce and supporting composable, MACH-based stacks. The new funding will advance the platform to help retailers respond more quickly to changing consumer behaviour and market conditions. Ongoing AI investment will enable smarter automation and allow LLMs to query Patchworks data, reducing manual work and freeing teams to focus on growth. Patchworks also plans a North American expansion, which will strengthen support for global brands operating across regions. 

French quantum hardware startup Isentroniq closed a €7.5 million funding round led by Heartcore with participation from OVNI Capital, Kima Ventures, iXcore, Better Angle, Epsilon VC, as well as the support from Bpifrance and the French National Research Agency (ANR) under the France 2030 program. Quantum computing could transform sectors such as medicine, energy, and logistics, but meaningful progress requires millions of qubits due to error-correction overhead. Superconducting qubits are promising, yet scaling is constrained by cryogenic wiring that adds heat and bulk; beyond roughly 100 qubits, systems hit thermal and spatial limits. Building to 1 million qubits with current approaches would demand massive facilities and tens of billions of euros. Isentroniq targets heat, cost, and space bottlenecks to enable up to 1,000× more qubits in existing dilution refrigerators, with a long-term goal of reducing the cost of a 1 million-qubit system to about €50 million. The company operates a fabless model—designing its architecture and outsourcing fabrication to specialised partners—to accelerate time to market, limit capex, and ensure industrial-grade quality. Today, wiring is the #1 bottleneck to scale superconducting quantum computers. Our mission is to turn it into an accelerator, said Paul Magnard, co-founder and CEO. Isentroniq was co-founded by Magnard, a superconducting-qubits expert with a PhD from ETH Zurich and former lead architect at Alice & Bob, and ThéodoreAmar, a second-time founder with prior roles at Bain & Company and Hilti. The announcement comes as major players, including Google, IBM, Amazon, IQM, Alice & Bob, and Rigetti, pursue roadmaps toward 100,000 to 1 million qubits. The new funding will support the development of Isentroniq’s wiring technology, team build-out, and partnerships to deliver a plug-and-play solution for scaling quantum systems.

Afori, a new AI platform purpose-built for insurance brokers, announced its official launch out of stealth alongside the closing of a €4 million pre-seed funding round. The round was led by General Catalyst, with participation from Yellow and Booom. Individual backers include industry leaders Christopher Lohmann (founder of Mulberry Ventures, former CEO of HDI and Gothaer), Chris Leifeld (founder of Thinksurance), and Garrett Koehn (CIO, CRC), as well as AI pioneers Mustafa Suleyman (CEO, Microsoft AI), Alex Rinke (co-founder and co-CEO, Celonis), and Mehdi Ghissassi (formerly of Google DeepMind). Founded in 2025 by Fabian Wesemann (wefox co-founder) and Sergi Banos (wefox’s first employee and former CTO), Afori is an AI platform for insurance brokers that automates back-office work starting in the email inbox. Integrated with Outlook, it understands insurance processes, turns unstructured communications into structured workflows, and enables brokers to focus on client relationships and growth. Addressing a key industry pain point, administrative work that takes over 60 per cent of broker time, Afori’s initial product, an AI sidebar, turns emails and documents into cases with structured tasks and delegates partial automations to insurance-specific AI agents within the inbox, saving users about an hour per day on average. According to Fabian Wesemann (CEO), Afori was developed in close collaboration with brokers to reflect their terminology and workflows, ultimately giving them more time to focus on client advice. By combining Agentic AI with deep industry expertise, we’re creating a product that integrates seamlessly into brokers’ daily workflows and delivers real, tangible value from the very first day, added Sergi Banos, CTO. The pre-seed round will accelerate product development, expand integrations, and drive German market adoption, with the platform debuting at DKM 2025 in Dortmund this month.

Italy-based Maia Ventures, an early-stage agrifood tech fund, has launched its first fund valued at €55 million to back 20–25 companies, with initial tickets ranging from €0.5 million to €1.5 million. Limited partners include institutional investors such as the European Investment Fund (EIF) and CDPVenture Capital Sgr, alongside private investors including leading Italian food corporates and their family offices (e.g., Teseo Capital sicav-sif, Cereal Docks via Grey Silo Ventures, Andriani). Maia, an Article 8 SFDR fund, targets both financial returns and impact, aiming to build a healthier, more efficient, and more resilient food system. The firm positions itself as a value-add partner, linking Italy’s established food industry with high-potential startups from top accelerators, universities, and the broader agrifood tech ecosystem. The team comprises former founders, investors, industry operators, and scientists, supported by a technical advisory group with representatives from leading agrifood universities and corporates. Maia views the current agrifood tech funding downturn as an opportunity to back solutions aligned with structural shifts and pressing industry needs, with a focus on the convergence of food, health, and sustainability. The fund is managed by Praesidium S.A., a Luxembourg-based AIFM, and advised by Maia Advisors SRL. It is fully operational, has completed six investments to date, and is targeting a final close in the coming months.

Arbio, a pioneer in the service-heavy property rental sector, is announcing a $36 million Series A funding round,  bringing the company's total funding to over $45 million.  Arbio is a full-stack AI-first services company in the property management and hospitality industry, which sits right at the intersection of AI-enabled services and roll-up plays, with the ambition to become the go-to brand for short-term rental in Europe.  I spoke to Constantin Schröder, founder and CEO of Arbio, to learn more.  From the hockey field to proptech  Schröder was born and raised in Cologne, a smaller German city. For most of his life, he was a professional field hockey player: “I played in the German Bundesliga and had aspirations to make the national team. That was everything I did back then—through high school and into university.” He studied business at WHU Otto Beisheim School of Management, a university renowned for producing founders behind companies such as Zalando, HelloFresh, and JOKR. “It wasn’t one of the big schools,” he says, “but a lot of founders have come out of there. It gave me a solid grounding in business and entrepreneurship.” A solution for a personal pain point After graduating, he reconnected with his oldest friend, Paul Bäumler who’d previously founded two companies: a maths-learning platform in high school that he sold when he was just eighteen, and a corporate social responsibility startup letsact, which he launched during university and sold after graduation. They decided to take a break and travel. “We started in Europe, then went to Mexico and South America,” he says. It was during that trip—in Mexico—that the idea for Arbio was born: “We stayed in maybe ten or twelve Airbnbs,” he remembers. “Some were beautiful, but the overall experience was often disappointing: unreliable hosts, inconsistent standards, no 24/7 check-in or support. We realised that most small hosts couldn’t be on-site or manage round-the-clock communication. There was a clear pain point—but also a big opportunity, because Airbnbs still offered more flexibility and value than hotels.” Modernising an outdated industry There are 6.5+ million alternative accommodation properties in Europe,but managing them is stuck in the 20th century.  Owners face unpredictable returns, endless hassle, and poor reviews, while guests experience increasing inconsistency.  Traditional property managers claim they can help owners self-manage their properties by relying on legacy software, creating a patched-up, inefficient solution.  Turning a passion project into a dynamic, scalable business Back in Berlin, the founders of Arbio started digging into the local Airbnb market.  “At first, it was just a passion project,” he says. “We took over a small hosting business from someone who owned a beautiful apartment in central Berlin with great reviews. We did everything ourselves—cleaning, guest communication, even setting up baby beds for families. I remember scrubbing bathrooms with toothbrushes to get between the tiles!” Their hands-on approach paid off.  “Within a few months, we’d significantly increased that property’s revenue,” he says. “We introduced technology like a dynamic pricing algorithm, which made a big difference. That’s when we realised there was huge potential to improve both the guest experience and host operations.” The team noticed that most small property managers charged the same rate year-round—maybe slightly higher at Christmas — and didn’t sync pricing across booking platforms. It therefore built an algorithm that collects vast market data — how others price similar properties — and correlates it with live search volume and demand on Booking.com, Google, etc. It updates prices daily to optimise occupancy and revenue. “That product immediately increased income for property owners while maintaining guest satisfaction. It showed us technology could genuinely uplift both sides.” Since then, Arbio has grown quickly, raised venture capital and venture debt, expanded into Germany, Austria, and the UK, and manage over a thousand units. AI as Arbio’s engine Arbio's AI-native operating system that acts as a full-stack property manager. Thanks to AI workflows, it automates distribution, accounting, operations, guest communications, and dynamic pricing, delivering higher revenues, lower costs, and peace of mind for owners. Guests benefit from personalised stays, faster responses, and standardised quality. AI is core to Arbio, as Schröder stresses:  “We’re a technology-first, AI-first company. Everything revolves around product development that simplifies owners’ and guests’ lives.” He explained that AI enables the company to build and iterate faster while empowering teams to develop their own internal tools. According to Schröder, AI underpins three key areas: pricing optimisation, using machine learning across vast datasets; operational orchestration, where large language models automate communication between guests, owners, and service providers; and internal efficiency, allowing every department to solve problems autonomously. “Long-term, I think 90–95 per cent of what’s now done manually in property management can be automated,” he added. “Of course, physical tasks like cleaning and maintenance will still need people for quite a while—robots aren’t climbing Berlin staircases yet!”  This technology-first approach has enabled 1,000+ units currently under management across DACH,  10x revenue growth in the last two years, and 30+ strategic acquisitions completed through its dedicated M&A team. Balancing tech in property management Property management has traditionally been a very people-facing industry—hotel front desks and so on. Now there’s a tech wave—digital check-ins, automated communication.  However, Schröder believes that software alone doesn’t solve hospitality.  “Many solutions are pure tech — they sell AI software to property managers or owners. We take a different approach: being both a property-management company and a software builder. You need human touch to deliver real quality. Software can’t comfort a guest stuck in the rain who can’t check in. So we use AI to enhance—not replace — people. We extend the capabilities of those who have deep hospitality knowledge, enabling them to manage more properties efficiently.” He imagines a future where a team of 30–40 people can manage tens of thousands of units — “because we’ve automated the right things while keeping local teams for hands-on needs.” Arbio believes in a vertically integrated model—not just matching homes to guests, but guaranteeing quality. Schröder believes that Airbnb can’t always do that; “they have too much inventory.  And with AI-driven search tools like ChatGPT or Google’s travel results, I think we’ll see new discovery patterns and more niche travel platforms emerging.” According to Schröder, last year they began heavily building out their product with AI to orchestrate all communication between guests, property owners, and service providers. He asserts: “Our mission is to build “the one home for holiday homes,” where property owners who struggle with operations—and guests who struggle with reliability—both find a solution. When guests tell us they had a wonderful honeymoon at one of our apartments, or an owner says we’ve made their life less stressful—that’s what drives me.” Competing with hotels, not residents Importantly, addressing concerns about gentrification and housing shortages, particularly in cities like Barcelona, where short-term rentals face criticism, Schröder clarified that Arbio operates under a completely different model.  “We don’t use residential properties at all,” he explained. “We only operate in spaces with commercial licenses, designed from the start for hospitality use—just like hotels” He emphasised that this represents “a tiny fraction of the total market” and doesn’t affect local housing availability. “Housing shortages come from financial speculation and poor city planning, not regulated holiday-home use,” he said. “We see ourselves competing with hotels, not residents.” With the Series A funding, Arbio plans high-speed expansion across Europe, targeting hundreds of thousands of property owners who would benefit from its AI-native management platform. The company's technology-driven approach to scaling operations positions it to handle this expansion while maintaining its lean operational model.  Eurazeo led the funding, with investors including Open Ocean and previous investors Atlantic Labs and leading angels Philipp Freise and Justin Reizes (KKR), Johannes Reck and TaoTao (GetYourGuide) and DinBisevac (Buena), amongst others.  Elise Stern, Investment Director at Eurazeo, said:“ Arbio is pioneering an AI-native model in one of Europe’s largest and least digitised service sectors. By combining technology, data, and operational excellence, they’re redefining what property owners and guests can expect. We believe Arbio will become the category leader in the multi-billion-euro holiday rental management space, and we are excited to support them in this ambition.”  The funding will accelerate Arbio's acquisition pipeline, enhance its AI capabilities, and support expansion into new European markets where vacation rental management remains fragmented and owners are underserved. 

Every enterprise runs on B2B transactions - the purchase orders, shipping notices, invoices, payments, and claims that form the heartbeat of modern commerce. When these flows align, shelves are stocked, supply chains move, and customers are happy. When they don’t, the cost is both immediate and compounding. For decades, electronic data interchange (EDI) systems provided the backbone for these flows. Today, however, new layers of technology, from APIs and managed file transfer to event streaming and cloud-based ERP, have created a hybrid landscape that is faster but far more fragmented. Each system captures part of the truth, but none can see the whole story. The result is a hidden loss that drains enterprise value every year. The invisible cost of disconnection Many organisations believe their current tools provide full visibility. Yet, in practice, most still struggle to reconcile transactions end to end. Files move, APIs fire, and dashboards go green. But behind the scenes, purchase orders go missing, invoices don’t match, and payments stall without explanation. The result is what experts increasingly call the assurance gap. This is the space between knowing something happened and proving it happened correctly and completely. Across industries, that gap has measurable financial and operational consequences: Between one and five percent of EBITDA is lost each year through chargebacks, deductions, and disputes. Compliance teams spend weeks preparing audits because records cannot be reconciled across systems. Modernisation projects stall because leaders fear losing visibility during migrations. Partners and regulators lose trust when continuity cannot be demonstrated. How confident are most enterprises that they could prove every transaction completed as intended? Monitoring is not assurance Most enterprises assume that their combination of monitoring and observability tools covers these scenarios. The problem is that these platforms only provide technical insight - not business assurance. EDI gateways confirm delivery but purge records after 30 days. API platforms show success codes but not whether transactions match downstream. Observability tools highlight performance metrics but not whether shipments, invoices, and payments align. The gap between activity and truth remains. “In complex B2B ecosystems, visibility alone isn’t enough, ” writes Andrew Mallaband. “Enterprises now need proof of continuity - the ability to demonstrate that every transaction completes, reconciles, and complies. ” A new layer of assurance This challenge has given rise to a new capability known as B2B flow intelligence, a unifying layer that provides cross-system visibility, reconciliation, and auditability. It does not replace existing systems but overlays them to correlate, track, and prove the continuity of transactions across multiple technologies. At the data layer, B2B flow intelligence correlates transaction identifiers, timestamps, and acknowledgements across systems to reconstruct an unbroken lineage from order to settlement. At its core, B2B flow intelligence delivers: Unified lineage - linking all data from order to payment across EDI, APIs, ERPs, and partner portals. Real-time exception detection - spotting missing acknowledgements, duplicates, or mismatches instantly. Contextual routing - sending exceptions to the right resolver with full traceability. Immutable history - creating long-term, tamper-proof records for audits and regulators. This combination turns reconciliation from a reactive burden into proactive assurance. Industry snapshots The effects of the assurance gap vary by sector, but the underlying cause is the same, fragmented visibility. In retail, peak trading periods such as Black Friday reveal how fragile order-to-cash continuity has become. A single missing shipping notice can trigger millions in chargebacks and strained supplier relations. In pharmaceuticals, a missing chain-of-custody record can delay shipments or regulatory approvals, disrupting both revenue and patient care. In banking, complex payment modernisation programmes, such as the migration to ISO 20022, demand complete reconciliation across legacy and modern systems to avoid costly compliance breaches. These examples highlight how a purely technical view of system health no longer guarantees business continuity. What enterprises gain Organisations implementing B2B flow intelligence are realising quantifiable outcomes: Leaner operations - 50 to 70 percent reduction in reconciliation effort. Margin protection - 1 to 5 percent of EBITDA safeguarded. Modernisation with confidence - ERP and API migrations delivered on time, without data loss. Audit readiness - 60 percent faster audit preparation with immutable lineage. Stronger trust - partners, customers, and regulators share the same source of truth. CFOs see measurable protection of margin. CIOs gain confidence that transformation programmes will not disrupt business continuity. Compliance teams eliminate the manual scramble before every audit. The value is enterprise-wide. From visibility to proof The adoption of B2B flow intelligence signals a broader shift across industries, from monitoring systems to proving outcomes. Monitoring shows that something moved. Observability shows how it moved. Assurance proves it moved correctly. That assurance is fast becoming a board-level priority. It protects profit, strengthens compliance, and restores trust across the digital supply chain. A growing movement Several enterprise platforms are now developing capabilities in this space, bringing together data correlation, lineage, and governance. One example is meshIQ, which has extended its operational intelligence suite to deliver cross-system reconciliation and assurance for hybrid B2B environments. While approaches may differ, the direction of travel is clear. Enterprises no longer just need to observe. They need to prove.

Most small and medium-sized businesses face a dangerous paradox: they need robust Identity and Access Management to protect their data and systems, but the traditional IAM platforms designed for enterprises — with their six-month implementations, mandatory consultant fees, and requirement to upgrade every software tool to expensive enterprise plans — aren't built for companies of their size. I came across Stackbob.ai during a recent visit to Lviv for the IT Arena conference. It's a Ukraine-founded startup that uses AI to help SMEs and mid-market companies overcome the cybersecurity poverty line, where IAM has traditionally been too complex, too slow, and too costly for anyone but large enterprises. The company won two prizes in the pitch competition at IT Arena 2025:  3rd prize in the Business Track and the most interesting startup of the IT Arena. It's built a cybersecurity platform that helps businesses manage employee access across all the software tools they use. I spoke to CEO Ole Shved to learn more.  From $100M exit team to starting over in wartime Ukraine According to Shved, everything really started ten years ago. He'd worked at several B2B startups and at his last company before StackBob, he met his current co-founder, Yarik Rozum, when he hired him as one of the lead engineers.  The duo worked together for about three years. That company—called Ad-Lib—was later acquired by Smartly.io for around $100 million.  He admits, "We were employees there, not founders, but it was still an incredible experience being part of an exit." At the end of 2021, Rozum went to work for a Seattle-based company, and while Shved (originally from Kyiv) was figuring out what to do next in Lviv, Russia's invasion of UK started. "When the war broke out, everything changed. We started helping people evacuate, driving them to the borders, and buying supplies for the army. For a while, I should go and fight, because everyone around me was so motivated. But there were long lines everywhere to join, and a friend of mine in Kyiv told me it was the same there. Eventually, when Kyiv became safer, I went back. When I got back to Kyiv, I met with Yarik again. He said, 'We've always wanted to do something together, and it turns out life is finite — you never know when you might die. Maybe we should do it now.' And that's how StackBob started." Breaking the "API and SSO tax" SStackbob's proprietary AI agents automate access and license management without SAML SSO or SCIM API integrations, cutting implementation from months to days and making enterprise-grade IAM affordable to companies that were previously locked out. tackbob's focus is on small and midsize businesses —" we call it breaking the cybersecurity poverty line, shared Shved.  "Those are the organisations that need protection and automation but can't afford or manage enterprise-level IAM systems." According to Shved, existing IAM platforms like Okta, Microsoft Entra ID, and others all share three significant issues because they're built on older generations of technology. "The first problem is what we call the API and SSO tax. If you're a midsize company and want to manage employee access centrally across all your tools, you have to upgrade every single one of those tools to its enterprise plan — because the connectivity features are locked there. So you pay for the most expensive plan just to get SSO access." "It's basically a massive vendor lock-in, he shared, asserting, "I sometimes call it a cartel." "It's even become standard advice for startups: 'Just add SSO and call it an enterprise plan, then charge two, three, or four times more. For midsize companies, it's ridiculous—they can't afford that. It's a huge burden." The second problem is the setup complexity. Implementing these systems takes months — sometimes six, nine, or even twelve months — and requires highly skilled engineers or consultants.  Shved shared that this is standard for many Series A or Series B companies. "And in the end, you still have parts of the stack that can't be integrated because there's no compatible protocol. So you end up managing some tools manually or using spreadsheets." So the team thought: How can we do this differently? How can we minimise our reliance on APIs and legacy protocols such as SAML or SCIM? From APIs to AI agents After several iterations, the idea emerged to utilise AI agents and browser automation for integrating and managing any web application.  "That means we don't need APIs at all, and the company doesn't need to upgrade to enterprise plans," Shved explained. "You simply invite our agent — just an email address—to your software tools. It acts as an AI administrator in the cloud and uses browser automation to perform actions like onboarding, , password resets, and permission changes, exactly as a human would do in a web interface." Closing the offboarding gap Critically, Stackbob also solves the problem of offboarding — I know I'm not the only one who has had access to an ex-employer's accounts months after leaving. Shved often finds that employees leave, but their accounts stay active across CRMs, analytics tools, or internal systems.  "It's a serious security gap, and most companies don't even realise it until there's an incident. Our agents automatically detect and revoke access across every integrated tool, without needing traditional IAM setup." The platform also addresses shadow IT — where employees use tools or apps without approval – by automatically discovering all the tools employees log into, "even the ones management doesn't know about. It gives full visibility." To achieve this, Stackbob built its own integration engine supported by a pretrained model that helps automate the process.  As a result, it can integrate all your applications — third-party, custom-built, or even legacy tools.  According to Shved, "as long as there's a browser interface for managing users, our agent can work with it. "If it runs in a browser, we can manage it" "Our marketing team likes to say we can integrate with more than 300,000 applications. It's probably true! Basically, if it runs in a browser, we can manage it." Stackbob started a little over three years ago and, in 2023,  joined the Techstars Seattle accelerator, which helped it expand into the US market and start signing US customers. Currently, the company is expanding into the UK and Nordic markets to strengthen its European presence, and with its growing customer base Ultimately, Stackbob is proving that there's significant demand for IAM solutions built for the 99 per cent of companies left behind by traditional platforms.

The Tech.eu Q3 2025 Report reveals that European tech companies raised €21 billion across over 900 deals during the third quarter of 2025. Approximately 14.1 per cent of these companies successfully closed seed rounds, amounting to over half a billion euros (€587.8 million). Today, we share the list of the 10 largest seed rounds among European tech companies that were completed in Q3 2025. Amount raised: €25M Donut Lab is a company that offers a full-stack EV hardware platform featuring modular, plug-and-play components including in-wheel motors, batteries, compute units, and software. Their in-wheel “Donut Motors” deliver high torque and power density, eliminating the need for traditional drivetrains. The platform is designed for electric mobility across land, air, marine, and robotics markets. Donut Lab’s modular approach reduces development complexity and accelerates time to market. In July, the company raised €25 million in seed funding, supporting the validation of Donut’s drivetrain-free architecture, engineered for flexible use across land, sea, and air applications. Amount raised: $27M Tulum Energy is a climate-tech company developing a methane pyrolysis platform to produce “turquoise” hydrogen and solid carbon from natural gas or biogas, without CO₂ emissions. The technology repurposes electric arc furnaces commonly used in steel production to crack methane molecules thermally, enabling a scalable, energy-efficient route to hydrogen. Tulum aims to bring hydrogen production costs down to parity with conventional grey hydrogen, positioning its technology as a viable, lower-emission alternative in heavy industries. In July 2025, the company secured a $27 million seed round to build a pilot plant in Pesquería, Mexico, within Ternium’s industrial complex. Amount raised: $26M Arago is a Paris-based AI and computer hardware company developing an energy-efficient photonic AI processor that uses light instead of electricity, paired with a full software stack for standard framework compatibility. The approach targets data centres and edge AI with high throughput and significantly lower power consumption. Arago raised $26 million in seed funding in July to accelerate the commercialisation of its photonic processor, codenamed “JEF”. Amount raised: $25M Maisa is an enterprise automation company building accountable AI agents that execute complex, decision-heavy workflows with full traceability. Its platform provides a “Chain of Work”, a step-by-step, auditable record of every action, decision, tool, and rule used, so teams can deploy digital workers that are explainable and compliant rather than black boxes. Maisa is model-agnostic and offers developer docs and a studio for building agents, alongside its research into the Vinci Knowledge Processing Unit (KPU) to improve reasoning and tool use at inference time. Following a $25 million seed round closed in August, the company launched Maisa Studio, a model-agnostic, self-serve platform that enables users to deploy digital workers trainable through natural language. Amount raised: €21M Paid is a business engine built for AI agents, managing pricing, subscriptions, margins, billing, and renewals with minimal integration. It allows agent companies to monetise automatically, track per-agent profitability, and implement outcome- or usage-based pricing models without building custom billing infrastructure. The company aims to help AI agent developers capture more value and scale revenue operations. The company closed a €21 million seed round in September to further develop its results-based billing infrastructure for AI agents and expand its offering to more enterprise customers. Amount raised: $20M Motor Ai is a Berlin-based startup working to accelerate the development of autonomous mobility systems. It offers an operating system for robots and vehicles that handles perception, planning, and control, enabling developers to build and deploy autonomous agents. Motor AI combines modular software architecture with real-world data to support robotic delivery, autonomous shuttles, last-mile logistics, and mobility applications. MOTOR Ai closed a $20 million seed round to deploy its certified, neuroscience-based autonomous driving technology, beginning with public road trials in Germany. Amount raised: $18M THEKER Robotics is a company specialising in AI-driven inspection robotics and infrastructure solutions. Its technology enables autonomous visual inspection of industrial facilities such as pipelines, bridges, and offshore equipment using drones and robotic platforms. THEKER Robotics integrates advanced perception, simulation, and automation to improve safety, reduce downtime, and lower inspection costs. In July, the company raised $18 million to advance AI-driven industrial automation. Amount raised: €12M Kongsberg Ferrotech provides robotic solutions for inspection, repair, and maintenance (IRM) of underwater infrastructure in the energy and maritime sectors. Using remotely operated robots, it offers all subsea IRM services in a single operation. Since 2021, it has been collaborating with Equinor, SINTEF, and Gassco to develop in situ 3D metal-to-metal printing for underwater repairs. In July, the company raised €12 million to advance its technology for improving the repair and maintenance of critical underwater infrastructure worldwide. Amount raised: €11M Brainr is a company that develops AI agents to review, verify, and summarise video content at scale. Its system uses multimodal reasoning across video, audio, and visual metadata to automatically detect events, anomalies, and quality issues. Clients in media, sports, and brand safety use Brainr’s agents to streamline monitoring workflows, reduce moderation costs, and enhance trust in content. With €11 million raised in September, BRAINR plans to support key growth initiatives, including international expansion to new regions and facilities, and to accelerate its research and development in artificial intelligence. Amount raised: $12M Conduct.ai is a startup developing autonomous agents for workplace automation. Its platform enables users to deploy AI agents that execute multi-step workflows, such as document review, data entry, or complex decision paths, and integrates with enterprise systems and business tools. Conduct.ai emphasises transparency, audit Trails, and user control, aiming to give organisations scalable, responsible automation. In September, Conduct closed a $12 million seed round and emerged from stealth with a mission to lead the largest transformation in enterprise IT by modernising legacy ERP systems