26 juli 2026 · 12:46
Japan Physical AI Strategy + Enterprise AI Governance Gap
Japan has unveiled a national physical AI and robotics strategy, backed by public computing infrastructure and the FRONTia foundation model project, with a target of capturing more than 30% of the global AI robotics market by 2040. For European manufacturers and policymakers, the announcement is a direct competitive signal. This episode also covers why a new Arctera report finds most organisations cannot prove their AI oversight controls actually work, with the EU AI Act already phasing in hard obligations.
Beluister deze aflevering:
Transcript
Samantha: Welcome to The State of Tech, The European Edition, Sunday July twenty-sixth, 2026. I'm Samantha Lawrence.
Bob: And I'm Bob Russell. Today: Japan bets big on physical AI and robotics, Fireworks AI lands one and a half billion dollars for enterprise infrastructure, a new report exposes gaping holes in AI oversight, Nubia unveils a smartphone built for AI agents, and to close out, two things you can actually use today: the lithium polymer battery innovations from Pknergy, and Block's new agentic workspace Buzz. Let's start in Tokyo.
Japan launches national strategy to seize thirty percent of the global AI robotics market by 2040.
Samantha: Japan is going all in on what it calls physical AI. That's the label for artificial intelligence built into machines that actually move: factory robots, warehouse arms, humanoids. The government just unveiled a national strategy at Nvidia's AI conference in Tokyo, backed by fresh public computing infrastructure and tight cooperation with domestic manufacturers. The goal is bold: capture more than thirty percent of the global AI robotics market by 2040. The economy ministry is funding something called the FRONTia Project, which will develop foundation models specifically for robotics.
Bob: And this is where Nvidia's Jensen Huang framed it neatly, saying Japan has a unique opening to reinvent manufacturing for an era of intelligent industries. Which makes sense when you think about it. Japan has an ageing workforce, a shrinking pool of factory workers, and decades of robotics expertise sitting inside companies like Fanuc and Yaskawa. What was missing was the AI brain. Now the state is trying to wire it all together at national scale.
Samantha: For Europe this is a wake-up call. Germany, Italy, and France all have serious industrial robotics traditions, but there's no equivalent coordinated push. Individual initiatives exist, national AI plans exist, but nothing that ties a foundation model programme directly to the factory floor with a market-share target attached. If Japan pulls this off, European carmakers and machine builders could find themselves buying their next generation of intelligent robots from Tokyo instead of Stuttgart.
Bob: The other angle worth flagging is the geopolitical layer. This is a US chipmaker, Nvidia, essentially co-launching a Japanese industrial strategy on Japanese soil. That's a very specific alliance in a world where the US, China, and Europe are all trying to lock down where the AI stack gets built. European industry groups have been arguing for years that the continent needs its own equivalent of this: a coordinated hardware, software, and manufacturing bet. The Japanese announcement makes that argument harder to ignore.
Fireworks AI raises one and a half billion dollars to power enterprise AI infrastructure.
Samantha: Fireworks AI has closed a Series D funding round of one and a half billion dollars. The company builds infrastructure that lets businesses run multiple AI models side by side, keep inference costs under control, and build what people are calling agentic applications, which is jargon for AI systems that can carry out multi-step tasks on their own. The money will go into expanding that platform. And the size of the round tells you something about where enterprise IT budgets are flowing right now.
Bob: It really does. This isn't a company building a chatbot or a shiny consumer product. It's plumbing for the AI economy: the layer that sits between companies and the big models like GPT, Claude, or open-source alternatives. What's shifted in the past year is that businesses no longer want to be locked into one model provider. They want to switch, compare, and pick whichever model does the job cheapest or fastest. Fireworks is betting that flexibility is the future.
Samantha: For European enterprises this matters in a specific way. A lot of the concern in Brussels and Berlin has been about digital sovereignty: whose infrastructure runs your AI, and where does the data sit? Fireworks is a US-based provider, and this funding round strengthens the American grip on the enterprise AI middle layer. That puts more pressure on European players like Mistral, Aleph Alpha, or newer infrastructure startups to offer credible alternatives, or risk seeing European banks and manufacturers default to US infrastructure by habit.
Bob: The bottom line: one and a half billion dollars in a single round shows enterprise AI spending is still accelerating, even as some analysts warn about an AI bubble. The money is going to the boring but essential middleware, not just the flashy models.
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.
New report finds fewer than one in five companies can prove their AI oversight actually works.
Bob: A new report called State of AI Governance 2026, published by Arctera, has landed with some genuinely uncomfortable numbers. Seventy-eight percent of organisations using AI expect communication-related risks to rise. And fewer than one in five companies can actually demonstrate that their AI oversight controls work. Not that they have controls on paper, but that those controls are effective when audited. That gap between adoption speed and governance capacity is the real story.
Samantha: And the risk areas the report flags are quite concrete. Communications, data retention, legal holds, risk workflows. These are the everyday operations where AI is being quietly plugged in, often without central IT knowing. Think of an employee using an AI assistant to draft client emails, or a compliance team using AI to sort documents. If nobody can prove what the AI did or why, you have a serious auditability problem the first time a regulator or a court comes knocking.
Bob: For Europe this hits particularly hard because the AI Act is already phasing in real obligations. Companies deploying so-called high-risk AI systems need documentation, monitoring, and human oversight. If this report is accurate, most organisations are nowhere near ready. And the awkward truth is that many boards approved AI rollouts last year without asking hard questions about governance, because everyone was afraid of falling behind competitors.
Samantha: What this shows is a widening gap between the marketing story about AI and the operational reality. The bottom line for European companies: the compliance clock is ticking, and having a policy document is not the same as having working controls.
Nubia unveils NaviX Ultra, a smartphone designed from the ground up for AI agents.
Bob: At the World AI Conference in Shanghai this weekend, Chinese manufacturer Nubia unveiled a device it calls the NaviX Ultra, and it's being pitched as one of the first smartphones designed specifically around AI agents. The key idea is that instead of you tapping through five apps to book a trip, the phone understands what you want in plain language and executes the whole chain of tasks itself. Booking, calendar, messaging, payment, all handled by an AI agent that works across apps rather than inside one.
Samantha: This is a real shift in how smartphones are being conceived. For fifteen years the model has been an app grid: you decide, you tap. Nubia is arguing that the interface of the future is intent-based. You say what you need, and the phone figures out the how. Google and Apple have both been moving in this direction with their AI assistants, but they've been retrofitting existing phones. Nubia claims to have built the hardware and software around this idea from the start.
Bob: For European consumers this raises some real questions. If an AI agent on your phone is booking things, paying for things, and reading your messages to understand context, that's an enormous amount of data flowing through one system. Under European privacy rules, users need to understand what's being processed and where. Chinese manufacturers have faced tough scrutiny in Europe before, and a device that ingests this much behavioural data will get a close look from data protection authorities.
Samantha: The broader signal from the Shanghai conference is that AI is moving from the chatbot phase into the everyday-device phase. Whether Nubia's specific product takes off in Europe or not, the direction of travel is clear: the smartphone as an assistant that acts on your behalf, not just a slab of glass with apps on it. European brands, from Nokia's successor HMD to smaller players, will need to decide whether to follow or specialise.
Lithium polymer battery breakthroughs from Pknergy point to longer-lasting portable electronics.
Bob: To close out, two things worth your attention today. First one is on the hardware side. Shenzhen-based Pknergy, part of the PKCELL group, showed off its latest lithium polymer battery technology at the Canton Fair this week. The headline is high energy density cells for the next generation of portable and industrial electronics. In plain terms: batteries that pack more power into the same space, which matters for everything from earbuds to medical devices to drones.
Samantha: What makes this worth flagging as something to check out is the transparency angle. The company opened up its manufacturing lines and testing labs to visitors, which is unusual for the battery industry. If you're a European hardware maker, a designer building a wearable, or even a hobbyist working on a project, this is one of the suppliers now openly competing for European contracts. Their customisation offer targets small and mid-sized clients, not just giant OEMs, which is a shift in a market that used to be dominated by a few opaque players.
Bob: For end consumers the practical takeaway is that the batteries inside your next set of wireless headphones, your next e-bike light, or your next smartwatch may well come from suppliers like this one. Longer runtime, thinner form factors, and better safety margins are all downstream effects. It's not something you install today, but it's the kind of shift that quietly reshapes what portable electronics can do over the next two years.
Samantha: And it fits a broader European trend. With energy density improving, product designers can push into new categories, from medical patches to compact power tools, without the battery being the limiting factor. Worth watching if you're in the hardware space.
Block launches Buzz, an agentic workspace for human-AI collaboration in finance.
Bob: And the second one to check out is Buzz, launched by Block, the payments company. Buzz is what's being called an agentic workspace: a shared environment where humans and AI agents work side by side on the same tasks, particularly in finance and operations. Think of it as a project room where the AI can pick up a piece of work, do a first pass, hand it back to a human, and log everything in one place.
Samantha: What makes Buzz interesting is that it's launching at exactly the moment when that Arctera governance report we talked about is warning nobody can prove their AI controls work. Buzz is essentially pitched as an answer to that: a controlled environment where AI actions are visible, auditable, and tied to specific human workflows. Whether it delivers on that promise remains to be seen, but the framing is smart.
Bob: For European finance and operations teams, this is worth exploring even if you don't adopt it. It shows one direction the market is taking: not AI as a magical replacement for humans, but AI as a colleague with a name tag, whose work you can see and correct. That model fits much better with European regulatory expectations around human oversight than a black-box AI making decisions in the background.
Samantha: The practical test is whether platforms like Buzz let compliance officers actually reconstruct what an AI did on a given transaction. If yes, this becomes a genuine tool for regulated industries. If not, it's just marketing on top of the same governance gap.
Samantha: Today we covered Japan's national push into physical AI and robotics, Fireworks AI's one and a half billion dollar funding round, the governance gap exposed by the Arctera report, Nubia's AI-agent smartphone the NaviX Ultra, and two things to try yourself: Pknergy's lithium polymer batteries, and Block's agentic workspace Buzz.
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.