Tech Layoffs, TSMC Profits, Google AI for Science & Pentagon Ethics

Samantha: Welcome to The State of Tech, The European Edition, Friday April seventeen, 2026. I'm Samantha Lawrence.

Bob: And I'm Bob Russell. Today: the AI chip boom keeps roaring as both TSMC and ASML raise their forecasts sky-high. AI models that spot cancer years before doctors can. Forbes names its AI 50 list — and Europe shows up big. Physical AI hits an inflection point for robotics. IBM and Illinois go deep on quantum-AI supercomputing. And finally, the surprising shift happening on that Forbes list that might just make you rethink who's really winning the AI race. Let's start with the chips.

The semiconductor giants powering AI just delivered blockbuster numbers — and the forecasts keep climbing.

Samantha: So if you want to understand where AI is heading, follow the money. And right now, the money is flowing through semiconductors like never before. Both TSMC and ASML reported this week, and the numbers are staggering.

Bob: TSMC came out first. The world's biggest contract chipmaker now expects sales to grow more than thirty percent this year in dollar terms. Their first quarter revenue hit eighteen point one billion dollars. Profits jumped fifty-eight percent. And management actually upgraded their forecast from earlier estimates. The reason? AI demand is not slowing down. It's accelerating.

Samantha: And then ASML, right here in Europe, the Dutch giant that makes the machines that make the chips — they raised their own outlook too. They now expect net sales between thirty-six and forty billion euros for 2026. That's up from previous guidance, driven by the same AI chip hunger.

Bob: What's remarkable is how these two companies sit at completely different points in the supply chain but tell the exact same story. TSMC manufactures the chips. ASML builds the lithography equipment that makes advanced chip manufacturing possible. When both are raising forecasts simultaneously, that's not a blip. That's a structural shift in demand.

Samantha: And here's the thing — ASML is one of Europe's most strategically important tech companies. There is literally no alternative supplier for the extreme ultraviolet lithography machines they make. None. Every cutting-edge chip on the planet depends on ASML's technology.

Bob: Which puts Europe in a fascinating position. The continent often gets criticized for lacking big tech platforms, but in the physical infrastructure layer of AI — the actual hardware that makes it all possible — Europe is absolutely essential. ASML's raised outlook means more investment, more jobs, more R&D spending in the Netherlands and across its European supply chain.

Samantha: For European policymakers pushing digital sovereignty and the Chips Act, these results are validation. The semiconductor ecosystem here isn't just surviving — it's thriving because of AI. And TSMC's growth matters for Europe too, because so many European automotive, industrial, and tech companies depend on TSMC's foundries for their chip supply. A healthy TSMC means more stable supply chains for European industry.

Bob: The big question is whether this demand is sustainable. And every signal right now says yes. Data centers are expanding, AI training requires more compute, inference workloads are growing. Both companies see no ceiling yet.

Samantha: So the AI boom isn't just a software story. It's a hardware story. And Europe is right at the heart of it.

---

AI that can spot cancer years before a radiologist — and it's already being tested in hospitals worldwide.

Bob: Let's shift from chips to something that could save your life. Researchers at UC Berkeley and UCSF have developed AI models that can identify people at high risk for breast and lung cancer years before traditional imaging catches anything.

Samantha: Years. Not months — years. The models are called Mirai and Sybil, and they analyze medical imaging data to flag patients who might develop cancer long before tumors become visible to the human eye. The team has also launched a startup called Voio, which built a model called Pillar-1 that can detect an even broader range of medical threats and generate draft reports for radiologists.

Bob: And this isn't some lab experiment anymore. Over ninety hospitals across thirty countries are currently running studies or trials using these models. That's a serious global footprint. And because the models are open source, adoption has been fast.

Samantha: For Europe, this is huge. We've got a well-documented radiologist shortage. The workload keeps growing as populations age, and screening programs expand. These AI tools don't replace radiologists — they augment them. They help prioritize which scans need urgent attention and which patients need closer monitoring.

Bob: Think about it from a patient's perspective. If you're going in for a routine mammogram or lung scan, and an AI flag catches something that wouldn't have been noticed for another three years — that's potentially the difference between early-stage treatment and a much harder fight.

Samantha: European hospitals are actively participating in these trials, which is encouraging. But there's also the regulatory dimension. The EU's medical device regulations and the AI Act both apply here. These models will need to clear some serious compliance hurdles before they're deployed at scale across European healthcare systems.

Bob: Still, the direction is clear. AI-augmented healthcare isn't coming — it's here. And for European patients and health systems under pressure, this could be genuinely transformative.

---

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.

Forbes just named the most important AI companies of 2026 — and Europe has real representation.

Bob: The Forbes 2026 AI 50 list dropped, and there's a fascinating trend running through it. The era of just throwing bigger models at every problem? That's fading. What's rising instead is independence, efficiency, and industry-specific solutions.

Samantha: So less "one model to rule them all" and more "the right tool for the right job." Companies on this year's list are breaking their dependence on a handful of dominant AI platforms. They're building their own efficient, customizable models tailored to specific industries — healthcare, finance, logistics, manufacturing.

Bob: And the geographic spread is notable. Forbes specifically calls out the growing influence of companies from Paris, London, and Stockholm. European AI startups aren't just participating anymore. They're competing at the top tier.

Samantha: That shift toward AI independence resonates strongly with European values around digital sovereignty. If companies can run capable, specialized models without routing everything through Big Tech cloud platforms, that changes the power dynamic. European businesses get more control over their data, their costs, and their AI strategies.

Bob: The list also highlights AI agents in the workplace as a major theme. We're talking about AI systems that don't just answer questions but actually take actions — booking meetings, processing invoices, managing supply chains. That's a big leap from chatbots.

Samantha: For European companies, this is an opportunity. The continent has deep domain expertise in sectors like automotive, pharma, industrial manufacturing. Combining that expertise with specialized AI — rather than relying on general-purpose models from Silicon Valley — could be a real competitive advantage.

Bob: The key takeaway from this list is that the AI landscape is fragmenting in a healthy way. It's no longer just about who has the biggest model. It's about who applies AI most effectively. And Europe is well-positioned for that race.

---

Robots are getting smarter — and the business world is finally ready to deploy them at scale.

Samantha: Next up — physical AI. Capgemini's research institute just published a report saying we've hit an inflection point. Robots with autonomous intelligence are moving from experiments to real-world deployment.

Bob: The numbers back it up. Seventy-nine percent of organizations surveyed are now engaging with physical AI in some form. Twenty-seven percent are already deploying or scaling solutions. That's not early adopter territory anymore — that's mainstream momentum.

Samantha: What's driving this? Foundation models, better simulation tools, and edge computing. Basically, the same AI breakthroughs powering chatbots and image generators are now being applied to robots operating in warehouses, factories, and farms. Robots that can adapt to unpredictable environments rather than just following rigid pre-programmed routines.

Bob: And Capgemini found that about two-thirds of European executives believe physical AI will enable robotics applications that were previously considered impossible or impractical. Think high-mix manufacturing, autonomous logistics in cluttered environments, precision agriculture on varied terrain.

Samantha: For Europe specifically, this matters because of the labor challenge. Aging populations, shrinking workforces in key industries — physical AI and intelligent robotics offer a path to maintaining productivity without relying solely on labor market growth.

Bob: It's also worth noting this report comes from Capgemini, headquartered in Paris. So this is a European institution framing the global conversation on physical AI. European companies like Siemens, ABB, and others are already major players in industrial robotics and automation.

Samantha: The bottom line is that physical AI is no longer science fiction or a tech demo. It's becoming operational infrastructure. And European industry, with its strong manufacturing base, stands to benefit enormously if deployment keeps accelerating at this pace.

---

IBM and the University of Illinois just doubled down on merging quantum computing with AI.

Bob: IBM and the University of Illinois Urbana-Champaign are expanding their Discovery Accelerator Institute. The new five-year plan focuses on quantum-centric supercomputing — essentially building architectures where quantum and classical systems work together seamlessly.

Samantha: So this isn't just about quantum computing in isolation. It's about integrating quantum processors with classical supercomputers and AI systems to tackle problems that neither could solve alone. Think drug discovery, materials science, complex optimization problems.

Bob: The university brings serious firepower — their supercomputing infrastructure is among the best in the world. IBM contributes its quantum hardware and software stack. Together they're developing new algorithms specifically designed for this hybrid approach.

Samantha: And they're also targeting next-generation AI systems and AI-driven engineering as part of the expanded mandate. So it's quantum plus AI, not quantum versus AI.

Bob: For Europe, this is both an inspiration and a competitive challenge. The EU has invested billions in quantum research through the Quantum Flagship program and various national initiatives. European institutions like the Fraunhofer Society, CERN, and several top universities are doing cutting-edge quantum work.

Samantha: But the scale of this IBM-Illinois collaboration is formidable. European quantum researchers will benefit from any open-source tools or published research that comes out of this institute. But it also underscores the need for Europe to forge similarly ambitious partnerships between industry and academia.

Bob: The race to practical quantum advantage is intensifying. And whoever cracks the quantum-classical-AI integration first could unlock capabilities we can barely imagine today. Europe has the talent. The question is whether it can match the scale and speed of these American partnerships.

---

The real surprise buried in the Forbes AI 50 list? It's not about who's on it — it's about what winning looks like now.

Samantha: Alright, let's end on something that made me genuinely smile. Remember that Forbes AI 50 list we mentioned earlier? There's a detail we didn't get to — and it says a lot about where the AI world is heading.

Bob: So here's the fun part. A few years ago, if you wanted to make the Forbes AI list, you basically needed the biggest model, the most parameters, the most compute. It was an arms race of sheer scale.

Samantha: And now? Some of the most celebrated companies on this year's list are winning because they figured out how to do more with less. Smaller models. Lower costs. Faster deployment. It's like the AI world finally discovered that efficiency is cool.

Bob: There's one trend in particular that caught my eye — companies building AI that runs on modest hardware, sometimes even on-device, without needing a massive cloud backend. That's a complete reversal from three years ago when everyone assumed you needed a data center the size of a football stadium to do anything interesting.

Samantha: It's the equivalent of going from needing a supercomputer to needing a decent laptop. And for smaller businesses, for startups in places without hyperscale infrastructure — including plenty of European companies — this levels the playing field dramatically.

Bob: What I love about this is it mirrors what happened in other tech waves. The early internet was dominated by a few giants with the biggest servers. Then it democratized. Mobile started with expensive devices and exclusive networks. Then it went everywhere. AI is following the same pattern — and we're watching it happen in real time.

Samantha: So the message from this year's AI 50? You don't need to be the biggest. You need to be the smartest about how you deploy AI. And that's a message that should make a lot of European entrepreneurs very happy.

Bob: I'll drink to that. Efficiency over excess — very European, if you ask me.

---

Samantha: That's our show. Today we covered: the AI chip boom lifting TSMC and ASML, AI models detecting cancer years early, Europe's rise on the Forbes AI 50, physical AI going mainstream in robotics, IBM and Illinois pushing quantum-AI supercomputing, and why smaller, smarter AI might just be the future.

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.