China's humanoid robot factory & Universal Robots Gen 7

Samantha: Welcome to The State of Tech, The European Edition, Tuesday September fifteen, 2026. I'm Samantha Lawrence.

Bob: And I'm Bob Russell. Today: a Chinese factory that spits out a humanoid robot every ten minutes, Universal Robots launches its seventh-generation cobot platform built for AI, Fujitsu unveils a diamond-based quantum computing prototype, an Australian wealth platform buys an AI advice startup, and to close out two things you can actually try today: the read-later app Readeck and the offline puzzle game Simon Tatham's Puzzles. Let's start with the humanoids.

China's UBTECH opens a humanoid robot factory producing one unit every ten minutes.

Samantha: We're starting with what may be the clearest sign yet that humanoid robots are moving from demo videos to actual mass production. Chinese robotics company UBTECH has reportedly opened a new factory that can roll a humanoid robot off the line every ten minutes. That's a potential annual output of ten thousand units. To put that in context: most humanoid robot companies worldwide are still counting their deployments in dozens, not thousands. This is China trying to industrialise a whole new category of machine, and doing it at a scale nobody in the West is currently matching.

Bob: And that scale is the whole story. Because as long as humanoids were bespoke prototypes, they stayed niche and expensive. Once you build a proper assembly line, unit costs drop, iteration speeds up, and suddenly you can start selling into logistics, retail, elderly care, all the places where labour is scarce. UBTECH is signalling it wants to be the humanoid equivalent of what BYD became for electric cars: a domestic champion that then floods export markets. Whether the robots are actually good enough to do useful work at scale is a separate question, but the manufacturing capability is now real.

Samantha: Which brings us straight to the European angle, and it's an uncomfortable one. Europe has serious robotics research, from ABB in Switzerland to KUKA in Germany, but nobody on the continent is building humanoid platforms at anything close to this volume. If UBTECH can produce ten thousand units a year at a competitive price, European industrial buyers will simply order from China, the way they already do for industrial drones and increasingly for electric vehicles. And that creates the same strategic dependency conversation we keep having about batteries and chips.

Bob: There's also a jobs and standards question. If low-cost humanoids arrive in European warehouses and shops within two or three years, unions and regulators will be scrambling to keep up. Safety certification, liability when a robot injures someone, data protection when these machines have cameras and microphones running all day. The EU AI Act has some hooks here, but nothing specifically designed for a mass-produced humanoid in a supermarket aisle. Expect Brussels to start paying attention very quickly if these units actually start shipping.

Samantha: The bottom line: mass production of humanoids has quietly begun in China. Europe now has a narrow window to decide whether it wants to compete, regulate, or simply buy.

Universal Robots launches Gen 7, its first out-of-the-box AI-ready cobot platform.

Bob: Staying with robotics, but in a very different corner of the market. Universal Robots, the Danish company that basically invented the collaborative robot category, launched its seventh-generation platform at the IMTS trade show. It's called Gen 7, and the pitch is that it's out-of-the-box AI-ready. That means modern software tools, better connectivity, and a completely reengineered control box, teach pendant, and tool flange, plus a new g-Series line of robot arms. All aimed at making it much easier to deploy AI-powered applications directly on the factory floor.

Samantha: The context here matters. Universal Robots has shipped more than a hundred thousand cobots over two decades, so this isn't a startup making promises. What's changing is that until now, adding AI to a cobot meant a lot of custom integration work with third-party vision systems and separate compute boxes. Gen 7 is trying to collapse that into a platform where an integrator can plug in a modern AI vision model or a language interface without rewiring the whole workflow. In practice, that could mean a small factory in Brescia or Eindhoven deploying a robot in days instead of months.

Bob: And that's genuinely useful for European manufacturing, which is dominated by small and medium enterprises that can't afford a full robotics integration team. Universal Robots is Danish, part of the American Teradyne group, but its engineering and much of its supply chain sit in Europe. So this is one of the rare cases where the European tech industry is actually setting the pace globally in a fast-growing category. The contrast with the UBTECH story is striking: China goes for humanoid volume, Europe stays focused on flexible industrial arms with tighter AI integration.

Samantha: What to watch next is pricing and the software ecosystem. A platform is only as strong as the applications built on it, and Universal Robots will need third-party developers to actually ship those AI-powered skills. If they can pull that off, Gen 7 could become the standard cobot for European industry for the next five years.

Samantha: Quick interruption. If you listen to The State of Tech regularly, hit that like button and subscribe, that way you'll never miss an episode. Okay, moving on.

Fujitsu unveils its first quantum computing prototype using electron spins in synthetic diamond.

Bob: Onto quantum, and a genuinely novel approach. Fujitsu has announced its first prototype quantum computing module that uses electron spins inside synthetic diamond as qubits. This was developed with Delft University of Technology in the Netherlands, and it focuses on what are called tin-vacancy centres sitting on photonic circuits. The key claim: this design is built for scalability, because you can optically connect multiple modules together, and it operates at warmer temperatures than the superconducting quantum computers from IBM or Google, which need to be chilled to near absolute zero.

Samantha: And that temperature point is a bigger deal than it sounds. Superconducting quantum systems need enormous cryogenic refrigerators that consume serious power and take up significant floor space. Diamond-spin qubits operating at warmer temperatures could, in principle, mean smaller machines that are cheaper to run and easier to network together. Fujitsu is essentially betting that the winning quantum architecture won't be one giant machine but many modules stitched together optically. Whether that bet pays off is far from certain, but it's a credible alternative path.

Bob: The European angle here is Delft. Delft University of Technology has been a global centre for diamond-based quantum research for years, and this Fujitsu prototype is a direct product of that expertise. It shows how European fundamental research still translates into serious industrial partnerships, even when the commercial capital ends up in Japan or the United States. For the Dutch quantum ecosystem, and for spin-off companies in and around Delft, this validates the diamond-spin approach as more than an academic curiosity.

Samantha: The wider point is that quantum computing is quietly diversifying. Superconducting, trapped ions, neutral atoms, photonics, and now diamond spins. Nobody knows yet which one wins for practical fault-tolerant computing, so having a serious diamond prototype from a major vendor keeps the field honest. Don't expect this to break encryption tomorrow, but it's another step towards machines that could eventually tackle problems classical computers simply cannot.

Australian wealth platform Netwealth buys AI advice startup Paradino for fourteen million dollars.

Bob: Our fourth news story takes us to financial services, where AI is quietly reshaping how advice actually gets delivered. Australian wealth platform Netwealth has announced it's acquiring Paradino, an AI-enabled startup focused on adviser workflow and advice automation, for around twenty million Australian dollars upfront, or roughly fourteen million US dollars. Paradino's core product is called Athena, and it automates the tedious parts of financial advice: preparing advice documents, taking meeting notes, drafting client communications. Netwealth is also committing another ten million Australian dollars over two years to accelerate the product roadmap.

Samantha: The interesting thing is why a wealth platform is buying an AI startup outright rather than just licensing the software. Financial advice in Australia, and increasingly across Europe, has a productivity problem. Advisers spend a huge share of their time on compliance paperwork rather than actually advising clients. That limits how many people they can serve, which drives up the cost of advice and leaves many households without access to it at all. If AI can compress the paperwork from hours to minutes, an adviser can suddenly handle far more clients profitably.

Bob: For Europe, this signals a trend that's already visible in the UK and the Netherlands. Wealth managers and independent financial advisers are quietly rolling out AI copilots for meeting summaries and suitability reports. The regulatory sensitivity is high, because financial advice is heavily supervised, and any document a client relies on has to be defensible. So the tools that win in Europe won't just be the smartest, they'll be the ones with the tightest audit trails and the clearest human sign-off steps.

Samantha: The bigger pattern to watch is consolidation. When a platform buys an AI vendor rather than partnering with them, it usually means they want to lock the capability into their own ecosystem and stop competitors accessing it. Expect similar moves from European wealth platforms over the next year, either through acquisitions or exclusive partnerships. For advisers on those platforms, it becomes both an opportunity and a lock-in risk.

The read-later app Readeck saves articles for calm, offline reading without tracking.

Bob: To close out, two things you can actually use today. First up is Readeck, a read-it-later app that's been getting quiet attention among people fed up with algorithmic feeds and cluttered browsers. The idea is simple: you install a small browser extension, click a button on any article, and Readeck saves a clean text-only version to your library for later reading. No ads, no cookie banners, no pop-ups asking you to subscribe. Just the article, on any device, including offline.

Samantha: What makes Readeck stand out is that you can self-host it, meaning your reading list stays entirely on your own machine or a small server you run at home, rather than in someone else's cloud. For people who care about not leaking their reading habits to a US-based service, that's genuinely useful. If self-hosting sounds intimidating, there are also managed options where a small European provider runs it for you. The app works on desktop browsers and on Android and iOS through the web interface, and users report it handles long-form journalism, PDFs, and even ebooks reasonably well.

Bob: It's free and open-source, which means no subscription and no risk of the service disappearing overnight. The typical workflow is: during the day you bookmark articles you don't have time to read, and in the evening or on a train journey you open Readeck and work through them. Reviewers particularly praise the tagging system for organising research and the clean typography, which makes long reads much easier on the eyes than a busy news website.

Simon Tatham's Puzzles is a free collection of forty classic logic games that works entirely offline.

Samantha: Second thing to try today, and this one has nothing to do with AI. It's called Simon Tatham's Puzzles, a free collection of around forty classic logic and puzzle games. Think number grids, path-drawing games, minesweeper variants, tile-sliding challenges. All the kinds of things that keep your brain busy for ten minutes on a train without draining your battery or asking you to watch an advert every thirty seconds.

Bob: The reason to mention it now is that Simon Tatham's Puzzles has been quietly ported to just about every platform imaginable. There's a desktop version for Windows, macOS and Linux, an Android app, an iOS app, and even a version that runs in your browser. Everything is completely offline once installed, there are no accounts, no in-app purchases, no ads, and no tracking. It's the polar opposite of the mobile gaming market's current business model, and that's precisely why it's worth trying.

Samantha: Each puzzle can be generated at different difficulty levels, so you can dip in for a quick easy round or spend half an hour on something genuinely challenging. Long-time users say the pattern-matching game Loopy and the path puzzle Signpost are particularly addictive. It's the kind of thing to have on your phone for waiting rooms and commutes, without any of the guilt that comes with modern free-to-play games.

Bob: Free, offline, and no strings attached. Rare combination in 2026.

Samantha: Today we covered: China's UBTECH opening a humanoid robot mega-factory, Universal Robots launching its AI-ready Gen 7 cobot platform, Fujitsu unveiling a diamond-based quantum prototype, Australia's Netwealth buying AI advice startup Paradino, and two things to try yourself: the read-later app Readeck and the offline puzzle collection Simon Tatham's Puzzles.

Bob: Want to know more or react? Visit stateoftech.eu or email us at info@doorzetters.net.

Bob: State of Tech, the tech world in 15 minutes.