Hanyuan-2: China's dual-core 200-qubit quantum computer

Samantha: Welcome to The State of Tech, The European Edition, Sunday May ten, 2026. I'm Samantha Lawrence.

Bob: And I'm Bob Russell. Today: China unveils a dual-core quantum computer with 200 qubits, Apple opens its devices to third-party AI models, large language models cut preoperative anxiety in cancer patients, Taiwan jumps into the global top 20 for AI adoption, the Philippines courts UAE money for an AI industrial hub, and the boss of Baykar calls for a technological solidarity alliance. Let's start with quantum.

China's Hanyuan-2 launches as the world's first dual-core quantum computer with 200 qubits.

Samantha: China's CAS Cold Atom Technology has unveiled Hanyuan-2, and the headline number is 200 qubits across a dual-core architecture. It's being called the world's first dual-core quantum machine, and the design lets the two cores run in parallel, either splitting workloads or handling error correction simultaneously.

Bob: The really interesting bit is the hardware approach. Hanyuan-2 uses neutral atoms instead of superconducting circuits, which means you don't need the extreme cryogenic cooling that most western quantum systems rely on. The developers claim world-class qubit lifespan and significantly lower energy consumption, which matters because energy bills have been a quiet killer for quantum labs.

Samantha: And the dual-core point isn't just marketing. If one core runs the algorithm while the other handles error correction in real time, you're closer to fault-tolerant quantum computing, which is the threshold where this stuff becomes commercially useful rather than a physics experiment.

Bob: It also signals where China is placing its bets. Neutral-atom platforms have been quietly gaining ground globally, and Beijing seems to have decided this is the route to scale. Two hundred qubits on a more energy-efficient platform puts pressure on IBM, Google, and the European players working on different architectures.

Samantha: For Europe specifically, this lands at an awkward moment. The EU's quantum flagship has made progress, but funding is fragmented across member states, and there's no single European champion at this scale. Germany, France and the Netherlands all have strong research bases, yet pulling them into one industrial pipeline is the perennial challenge.

Bob: There's also the sovereignty angle. Quantum will eventually break a lot of current encryption, and whoever gets there first sets the rules. European regulators are already pushing post-quantum cryptography standards, but the underlying hardware race is one Europe cannot afford to watch from the sidelines.

Samantha: Industrial deployment is the phrase the Chinese team keeps using. They're talking pharma, materials science, logistics optimisation, the kind of problems where even modest quantum advantage translates into real money.

Bob: The bottom line is that the energy-efficiency claim alone could reshape procurement decisions. If you can run a 200-qubit machine without a cryogenic plant attached, suddenly universities and mid-sized firms become potential buyers, not just national labs.

Apple opens up, letting users pick ChatGPT, Gemini or Claude as their default AI.

Bob: Apple is reportedly about to do something genuinely out of character. Reports indicate the company will let users replace its native AI with third-party models, specifically ChatGPT, Gemini, and Claude. For a company famous for its walled garden, this is a meaningful shift.

Samantha: It's a tacit admission that Apple Intelligence hasn't kept pace. Siri's revamp has been delayed, the on-device models are competent but not cutting-edge, and rather than keep losing ground, Apple is turning the AI layer into a choice rather than a fixed feature.

Bob: There's a strategic logic here. If Apple becomes the neutral platform where you pick your preferred model, it stops competing on a battlefield it's losing and starts taxing the battlefield instead. Every API call routed through an iPhone is leverage.

Samantha: And for users, the practical effect is real. You could ask Siri something and have Claude answer, or summarise a document with Gemini, or draft an email with ChatGPT, all within the same operating system. That's a different relationship with your phone.

Bob: For Europe this matters on two fronts. First, the Digital Markets Act has been pushing exactly this kind of openness for two years. Brussels will quietly take credit, even if Apple frames it as user choice. Second, European AI startups suddenly have a path onto a billion-plus iPhones, assuming Apple's selection criteria aren't too restrictive.

Samantha: Mistral being the obvious example. If Apple's framework is genuinely open, a European model could be the default for European users. That changes the competitive map.

Bob: There are privacy questions too. Apple has built its brand on data minimisation, and routing queries to OpenAI or Google introduces external data flows. Expect detailed disclosure prompts and probably regional defaults that comply with GDPR.

Samantha: The key takeaway is that the assistant wars just got a new venue, and the device maker is now the referee rather than a player. That's a structural shift, not just a feature update.

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.

Large language models cut preoperative anxiety and physician workload in cancer trial.

Bob: New research from Fudan University Shanghai Cancer Center shows AI-assisted patient communication can measurably reduce preoperative anxiety. In a randomised trial with prostate cancer patients, the LLM-driven system produced lower anxiety scores, better disease understanding, and higher patient satisfaction.

Samantha: And critically, it shortened the time physicians spent on these conversations. The model identified concerns that doctors had been missing, the kind of unspoken worries patients don't raise in a fifteen-minute consultation.

Bob: That last point is where I think this study breaks new ground. We've seen plenty of AI-in-healthcare papers about diagnostics, but this is about empathy and communication, areas where most people assumed humans had a permanent edge. Turns out a well-prompted LLM can surface concerns patients don't articulate to a busy surgeon.

Samantha: The researchers also flagged that the system worked across different languages and cultural settings, which is crucial for any kind of international deployment. Healthcare communication is heavily context-dependent, so generalisability is the hard test.

Bob: For European health systems, this is genuinely interesting. The NHS, the Dutch hospitals, the German Krankenhäuser, they're all dealing with chronic staff shortages and burnout. If you can shave fifteen minutes off pre-op consultations while improving patient outcomes, that's not a marginal gain.

Samantha: There are caveats. Medical AI in Europe falls under both the AI Act and the Medical Device Regulation, so any deployment needs serious validation. You can't just plug ChatGPT into a hospital workflow and call it a day.

Bob: And there's a liability question nobody has fully answered. If an LLM provides reassurance that turns out to be misleading, who carries the risk? The hospital, the physician who signed off, or the model provider?

Samantha: Still, the direction of travel is clear. AI as a communication assistant, not a replacement, fits the European preference for human-in-the-loop systems. And patient anxiety is a real clinical problem with real outcomes attached.

Bob: The bottom line is healthcare AI is moving from the lab into the clinic faster than the regulatory frameworks can keep up. Studies like this will accelerate that, for better or worse.

Taiwan enters global top 20 for AI adoption, overtaking the US and Germany.

Samantha: Taiwan has cracked the global top 20 for AI diffusion in the first quarter of 2026, according to a Microsoft think tank report. Almost thirty-two percent of Taiwan's working population is now using generative AI, up from 23rd place in the second half of last year.

Bob: And it's worth pausing on who Taiwan just overtook. The United States, Denmark, and Germany are all behind it on adoption rates. That's a striking reversal, given how much of the AI conversation is dominated by American firms.

Samantha: The report attributes a lot of this to improved AI capabilities in Asian languages. For years, English-language models worked beautifully in English and clunkily everywhere else. As Mandarin, Japanese, and Korean performance has caught up, adoption in those markets has surged.

Bob: That's a lesson Europe should absorb. Generative AI in Dutch, Italian, Greek, or Polish has historically lagged the English experience. If linguistic quality is the bottleneck, then European AI adoption depends partly on whether models like Mistral, or fine-tuned versions of larger ones, deliver native-level performance.

Samantha: Germany being below Taiwan is the headline that will sting in Berlin. Germany has been investing heavily in industrial AI, but the diffusion measure here is about everyday workplace use, which is a different metric entirely.

Bob: It also tells you something about workplace culture. Taiwan's tech sector is dense and highly interconnected, so tools spread fast. European workplaces tend to be more cautious, more compliance-driven, and frankly slower to adopt anything that might touch sensitive data.

Samantha: That caution isn't entirely unwarranted. GDPR concerns, works councils, sectoral rules, they all slow rollout. But there's a productivity gap opening up, and at some point regulators and employers need to decide whether they're protecting workers or just delaying the inevitable.

Bob: The key takeaway is that AI adoption is now a competitiveness indicator, not just a curiosity statistic. Falling behind on diffusion translates directly into falling behind on output per worker.

Philippines courts UAE money for a four-thousand-acre AI-native industrial hub.

Bob: The Philippines is pitching hard to UAE investors for its AI ambitions. Following its accession to the US-led Pax Silica initiative, Manila is talking to Masdar and DAMAC Digital about a 4,000-acre AI-native industrial acceleration hub in New Clark City, in Luzon.

Samantha: Pax Silica is essentially the American attempt to build a secure, allied supply chain for AI chips and semiconductors, deliberately sidestepping Chinese components. The Philippines joining gives it geographic and political weight in Southeast Asia.

Bob: The hub itself is ambitious. Four thousand acres designed from the ground up for AI workloads, meaning purpose-built data centres, dedicated power infrastructure, and digital connectivity. That's a different proposition from retrofitting old industrial parks.

Samantha: The Emirati involvement is the diplomatic story. UAE capital has been flowing into AI infrastructure across multiple continents, from G42's deals in the United States to investments in African data centres. The Philippines is the latest stop, and Masdar's clean-energy expertise fits the power-hungry nature of these facilities.

Bob: For Europe, the implications are mixed. On one hand, European tech firms could supply equipment, software, and expertise to the hub. On the other, this is another example of AI infrastructure scaling up outside Europe, with American and Gulf capital, while European projects struggle for permits and power connections.

Samantha: Energy is the recurring constraint. AI hubs need gigawatts, not megawatts, and Europe's grid is already strained. The Philippines is essentially saying it will build the power infrastructure as part of the deal, which is something European member states find politically difficult.

Bob: There's also the geopolitical alignment angle. Pax Silica is explicitly an allied initiative, and Europe sits awkwardly within it, wanting strategic autonomy while depending on the same supply chains. Watch whether European firms get carve-outs or end up boxed out.

Samantha: The bottom line is that AI infrastructure is becoming a geopolitical asset class, and the deals being signed today will shape the capacity map for the next decade.

Baykar's chief calls for a technological solidarity alliance against tech monopolies.

Samantha: And we'll close with something a bit different. Selçuk Bayraktar, the CEO of Turkish defence firm Baykar, has called for what he's branding a Technological Solidarity Alliance among allied and developing nations, aimed at countering global tech monopolies.

Bob: He made the pitch at the SAHA 2026 International Defence and Aerospace Exhibition. His argument is that the dominance of a few corporate giants threatens national sovereignty, privacy, and human freedom. He framed it as a moral choice, talking about ethics, justice, and human dignity.

Samantha: It's a striking speech coming from a defence drone manufacturer. Baykar is best known for the Bayraktar TB2, which has shown up in conflicts from Ukraine to the Caucasus, so this isn't your typical civil-society critique of big tech.

Bob: That's what makes it interesting. When a successful defence-tech entrepreneur starts talking about technological sovereignty in moral terms, it suggests a coalition forming outside the usual US-China binary. Turkey, Gulf states, parts of Africa, Southeast Asia, all looking for a third way.

Samantha: The European resonance is obvious. Brussels has spent years talking about digital sovereignty, the AI Act, the Data Act, the Digital Markets Act, all aimed at exactly the kind of monopoly concentration Bayraktar is describing.

Bob: Whether a formal alliance ever materialises is another question. Solidarity alliances sound great at trade shows and rarely survive contact with actual procurement budgets. But the rhetoric matters because it shapes what counts as legitimate.

Samantha: And there's a delicious irony in a drone manufacturer warning about technology threatening human freedom. The speech writes itself for the cynics.

Bob: True, but the underlying point isn't wrong. If five companies control the cloud, the models, and the chips, the room for sovereign decision-making narrows considerably. Whether you're in Ankara, Amsterdam or Manila, that's a real concern.

Samantha: Today we covered: China's dual-core quantum computer, Apple opening up to third-party AI, AI cutting preoperative anxiety, Taiwan's surge in AI adoption, the Philippines courting UAE money for an AI hub, and Baykar's call for a technological solidarity alliance.

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.