IBM Quantum Chip Factory: $1B CHIPS Act Deal Explained

Samantha: Welcome to The State of Tech, The European Edition, Friday May twenty-two, 2026. I'm Samantha Lawrence.

Bob: And I'm Bob Russell. Today: IBM lands a billion dollars from Washington to build America's first quantum chip factory, a one trillion dollar AI investment wave that's set to lift the global economy, Israel's military unveils a brand new AI combat unit, China's AI boom rewrites the graduate job market, SpaceX tries again with its biggest rocket ever, and to close out, two practical things you can actually try this weekend. Let's get into it.

IBM lands one billion dollars to build America's first quantum chip factory in New York.

Bob: So this is the big one today. IBM is getting a one billion dollar package from the Trump administration to set up a new business called Anderon, which will manufacture quantum computing chips in New Albany, New York. It's the first specialised quantum chip plant in the United States, and it's part of a wider two billion dollar push for the whole quantum sector.

Samantha: And the money flows through the CHIPS and Science Act from 2022, the same pot of cash that's been rewiring American semiconductor manufacturing. The framing here is clear: quantum is being treated as critical national infrastructure, right next to AI and defence.

Bob: Which is a real shift. Quantum computing has been the perpetual five-years-away technology for a decade. Now Washington is putting hard cash behind it, betting that whoever masters quantum chips first gets a lasting edge in cryptography, drug discovery, and AI training.

Samantha: And the geopolitical signal matters as much as the technology. The US wants this manufactured at home, not in Taiwan, not in South Korea, not in the Netherlands. They want the whole supply chain on American soil.

Bob: Quick reality check though: a quantum chip is not like a normal processor. It needs to operate near absolute zero temperature, the production is incredibly delicate, and only a handful of companies in the world can do this at all. IBM is one of them, and now they get government backing to scale it up.

Samantha: For Europe this is uncomfortable reading. We have serious quantum research, in Delft, in Munich, in Paris, in Copenhagen. But research is not the same as a billion-dollar factory. The European Chips Act exists, but it's mostly aimed at conventional semiconductors, not quantum hardware.

Bob: And if the US locks down quantum chip manufacturing the way it has tried to with AI chips, European companies and universities could end up depending on American export licences for the hardware they need. We've seen that movie before with Nvidia's AI accelerators.

Samantha: So expect this to land on the desks of European policymakers very quickly. The question is no longer whether quantum matters, it's whether Europe is willing to put public money behind it at the same scale.

Bob: One detail worth noting: the facility will employ thousands over time. It's not just chips, it's an industrial base. That's the real prize.

Global AI investment is set to hit one trillion dollars in 2026, lifting world growth.

Samantha: Story two, and this is the kind of number that makes you blink. Commonwealth Bank economists are projecting that AI investment will add roughly half a percentage point to US economic growth this year, and a quarter point to global growth. Worldwide AI spending could touch one trillion dollars in 2026.

Bob: And here's the striking detail: about 58 percent of that American AI capital flows out to other economies, through supply chains, equipment orders, components, services. So when Microsoft, Amazon, Google, or Meta build a new data centre in Arizona, a lot of that money ends up in Taiwan, South Korea, Japan, and yes, parts of Europe.

Samantha: The hyper-scalers, the companies running the largest cloud data centres, are doing most of the spending. They're buying chips, cooling systems, networking gear, and increasingly, power generation capacity. It's a building boom on a scale we have not seen since the original internet build-out in the late nineties.

Bob: Though there's a nervous undertone. Some analysts are warning this looks like an investment bubble, with capacity being built ahead of confirmed demand. If AI revenue doesn't catch up, a lot of those data centres become very expensive empty buildings.

Samantha: That's the central question for 2026. Is the spending justified by actual productivity gains, or are we watching a classic over-build cycle?

Bob: For Europe the picture is mixed. Dutch lithography giant ASML benefits enormously, because every advanced AI chip on the planet still depends on its machines. German industrial firms supplying cooling and power systems get a tailwind too.

Samantha: But the headline economic uplift, the productivity, the new companies, the high-paying jobs, that's largely happening in the United States. Europe is supplying the picks and shovels, not striking the gold. And that's a strategic problem we keep failing to address.

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.

Israeli military launches Alumot, a new AI and cyber division built for the battlefield.

Bob: The Israeli Defence Forces have officially launched a new technological and operational division called Alumot. It mixes combat soldiers with AI researchers, data specialists, and cyber experts, all in one unit, with the explicit goal of delivering real-time AI tools to troops in the field.

Samantha: The framing from the IDF is that information superiority is now a battlefield weapon, on par with firepower or armour. Alumot is designed to push AI-driven targeting, sensor analysis, and decision support straight into active operations.

Bob: This is a significant organisational move, because most militaries keep their tech researchers separate from operational units. Putting them in the same chain of command means decisions move faster, but it also raises serious ethical questions about autonomous targeting and human oversight.

Samantha: And Israel has been criticised before for the role of AI systems in target selection during operations in Gaza. Independent reporting last year suggested AI tools were already being used to generate target lists at high speed. Alumot formalises and expands that approach.

Bob: The wider context is that every major military is now experimenting with battlefield AI. The US, the UK, France, Germany, China, Russia. The difference here is that Israel is institutionalising it openly, with a named unit and a public mandate.

Samantha: For Europe this lands awkwardly. The EU has been working on an AI Act that imposes strict rules on high-risk AI, but defence applications are largely exempt. So while European companies face heavy compliance for civilian AI, military AI sits in a different category entirely.

Bob: And European defence ministries are watching closely. France and Germany have their own AI defence initiatives, the UK has the Defence AI Centre. But none of them have gone as far as creating a single fused unit. Expect that to change in the next eighteen months.

China's AI boom is rewriting the graduate job market and squeezing every other industry.

Samantha: Story four. In China, the AI sector is now the dominant force in graduate hiring. Tech companies are aggressively competing for AI-literate graduates, and the ripple effect is spreading to every other industry. Banking, manufacturing, retail, healthcare, all of them are expected to recruit workers who can use AI tools day-to-day.

Bob: And the numbers behind this are striking. Chinese universities have rapidly expanded AI and machine learning programmes, and the government has made AI literacy a national strategic priority. We're talking millions of graduates each year, with AI skills baked into the curriculum.

Samantha: The wider implication is that the definition of a competent worker is shifting. In China, knowing how to prompt an AI model, automate a workflow, or interpret model output is becoming as basic as knowing how to use email.

Bob: And this puts pressure on employers everywhere. If Chinese graduates arrive in the workforce already AI-fluent, and European graduates are still being taught spreadsheets, the productivity gap widens quickly. That's not a hypothetical, it's measurable within a few years.

Samantha: For Europe the question is whether our universities adapt fast enough. Some are moving, Imperial in London, TU Delft, ETH Zurich, but it's uneven. And many vocational education tracks barely touch AI at all.

Bob: There's also a labour market angle. European employers are starting to demand AI skills in job adverts that two years ago required none. The supply is lagging, which means wages for AI-literate workers are climbing fast. If you're a graduate this year, that's an opportunity. If you're a policymaker, that's a warning sign.

SpaceX retries the first Starship V3 launch today after a last-second scrub.

Bob: Moving on to something a bit lighter, and concrete. SpaceX scrubbed the first launch attempt of its upgraded Starship V3 yesterday, due to a hydraulic pin issue on the launch tower's chopstick arms, those are the mechanical arms designed to catch the rocket on return. They're trying again today, Friday, from Starbase in Texas.

Samantha: This is the first Starship flight of 2026, the first test in seven months, and the debut of the V3 design. If you want to actually watch it, SpaceX streams the whole thing live on their website and on X, usually starting about thirty minutes before liftoff. The launch window opens late afternoon European time, so it's perfectly timed for a Friday evening.

Bob: And it's worth tuning in even if you're not a space nerd. Starship is the largest and most powerful rocket ever built, and the V3 version is supposed to be a meaningful step toward actual cargo and eventually crew missions. The visuals are spectacular regardless of whether the test succeeds or ends in what SpaceX cheerfully calls a rapid unscheduled disassembly.

Samantha: Which is engineer-speak for a very big explosion.

Bob: Exactly that. And the practical thing here for European listeners: the stream is free, no account needed, and SpaceX has good on-screen telemetry so you can follow altitude, speed, and engine status without needing a physics degree.

Samantha: One small tip: if you want the best commentary, the official SpaceX stream is dry and technical, but there are independent streams on YouTube, channels like Everyday Astronaut or NASASpaceflight, that add context in plain language. Worth having both open in two tabs.

Bob: And if it scrubs again, which is genuinely possible with a rocket this complex, they'll usually announce a new window within a few hours. So it's a low commitment way to enjoy some Friday evening drama.

A practical AI shift to try this weekend: running smart tools directly on your laptop, no cloud needed.

Samantha: To close out, something more grounded. The wider AI industry is shifting toward what's called edge deployment, meaning AI that runs on your own device rather than in a distant data centre. The practical version of this you can actually try today is a free app called LM Studio.

Bob: It's available for Windows, Mac, and Linux, completely free, and what it lets you do is download and run AI chat models directly on your own computer. No subscription, no data sent to OpenAI or Anthropic, no internet connection required once it's set up.

Samantha: Installation takes about ten minutes. You download the app, pick a model from their built-in library, something like Llama or Mistral, and you've got your own private chatbot running locally. It will not be as sharp as ChatGPT, but it's surprisingly capable for everyday writing, summarising, or brainstorming.

Bob: And the big appeal is privacy. If you work with sensitive documents, client information, legal notes, medical paperwork, running an AI locally means none of that leaves your machine. For lawyers, doctors, journalists, anyone with confidentiality requirements, that's a meaningful difference.

Samantha: The catch is hardware. You'll want at least sixteen gigabytes of memory in your laptop, and ideally a recent graphics chip. On older machines it will work but slowly. On a modern Mac or a decent gaming laptop, it's genuinely impressive how fast it runs.

Bob: So that's the weekend project. Download LM Studio, pick a small model to start, and within an hour you've got a fully private AI assistant running on your own hardware. Free, offline, yours.

Samantha: Today we covered: IBM's billion-dollar quantum factory deal, the trillion-dollar global AI investment wave, Israel's new battlefield AI unit, China's AI-driven graduate hiring boom, SpaceX's Starship V3 retry, and running private AI on your own laptop with LM Studio.

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.