Gartner: AI drives semiconductors to $1.6T in 2026

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

Bob: And I'm Bob Russell. Today: a huge semiconductor forecast driven by the AI boom, humanoid robots hitting factory floors in record numbers, the UK's telecoms bottleneck threatening its AI ambitions, the US Federal Reserve putting AI at the centre of monetary policy, and two things you can actually use today: the offline Europe-wide navigation app Magic Earth, and the open-source writing tool Zettlr. Let's get into it.

Gartner forecasts global semiconductor revenue will nearly double to one point six trillion dollars in 2026.

Samantha: We start with a number that is genuinely hard to get your head around. Gartner is forecasting that global semiconductor revenue will hit one point six trillion dollars this year. That is a ninety-two percent jump from 2025. Almost a doubling of an already enormous industry in a single year. And the driver, unsurprisingly, is AI. Memory chips alone are expected to bring in eight hundred and thirty-seven billion dollars, more than half of the entire market, and roughly double their share from a year ago.

Bob: And the memory piece is the eye-catcher, because memory used to be the boring, cyclical corner of the chip world. Prices go up, prices go down, factories overshoot, everyone waits it out. Now suddenly memory is the star, because AI systems in data centres need enormous amounts of fast memory to keep those accelerators fed. Gartner also flags that the AI data centre ecosystem, so the servers, the accelerators, the networking around them, jumps from about thirty-six percent of the semiconductor market this year to over fifty-three percent by 2030. That is a fundamental reshaping of what this industry is even for.

Samantha: And that matters for Europe on two fronts. First, supply chains. If memory pricing is running this hot, every European company that builds hardware, from cars to industrial equipment to consumer electronics, is going to feel it in their bill of materials. Second, investment priorities. The European Chips Act was built around logic manufacturing, the classic processor factories. This forecast suggests the real money and the real strategic leverage might actually sit in memory and in the packaging around AI accelerators, where Europe has much less of a footprint.

Bob: There's also a concentration story hiding in these numbers. Nearly doubling the market in one year means the winners take a disproportionate slice. Analysts point out that a handful of companies, mostly in Taiwan, South Korea and the United States, capture most of that growth. For European policymakers who have been talking for years about strategic autonomy in chips, this is the moment where the gap between rhetoric and reality gets uncomfortable. The bottom line is that the AI boom is now the semiconductor cycle, and Europe is largely a customer in that story, not a supplier.

Humanoid robot shipments top twenty-two thousand units in the first half of 2026 as industry deploys them at scale.

Samantha: Next up, humanoid robots are having their breakout moment. In the first half of 2026, more than twenty-two thousand humanoid robots were shipped worldwide. That is not a research number, that is real units being deployed, mostly in industrial settings. Manufacturing lines, warehouses, logistics hubs. The kind of places where the work is repetitive, physically demanding, and increasingly hard to staff.

Bob: And twenty-two thousand in six months is the kind of figure that quietly signals a phase change. A year ago the conversation was still whether humanoids would ever leave the demo video. Now production capacity is expanding, buyers are placing real orders, and the industry is talking about embodied intelligence, meaning AI that does not just answer questions in a chat window but actually moves through the physical world and picks things up. The interesting question is not whether these robots work. It is where they land first, and at what cost per hour compared to a human shift.

Samantha: For European industry, this is a double-edged shipment. On one side, manufacturers in Germany, Italy, Poland, and the Nordics are exactly the kind of buyers who could use these systems, aging workforces, high labour costs, tight margins. On the other side, most of the leading humanoid makers today sit in China and the United States. Europe has a strong industrial robotics tradition through companies like KUKA and ABB, but the humanoid form factor is a different race, and Europe has not really put a flagship player on the board yet.

Bob: And there is a regulatory piece that is going to bite quickly. If tens of thousands of humanoid robots start working alongside people on factory floors, European labour law, workplace safety rules, and liability frameworks will need to catch up. The current safety standards were written for caged industrial arms, not for a two-legged machine walking through a warehouse aisle next to a human colleague. Expect Brussels to be forced into that conversation sooner than it planned.

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.

UK executives warn lagging telecoms infrastructure could derail Britain's AI ambitions.

Bob: Story three takes us to the United Kingdom, where industry leaders are sounding a pretty blunt alarm. The warning is that Britain risks falling behind in the global AI race, not because of a lack of models or data centres, but because its telecoms infrastructure simply cannot handle what is coming. The UK currently ranks fifty-seventh globally on mobile network performance. That puts it well behind other G7 countries and most of its EU neighbours, and executives say that gap is now a strategic problem.

Samantha: And the mismatch is striking. On paper the UK is doing a lot of the right things at the top of the stack. There are big data centre investments, government AI strategies, sovereign compute initiatives. But AI is not just about the data centre, it is about the pipes that move data to and from users and devices. If your 5G coverage is patchy and your fixed networks are uneven, then every AI application that depends on real-time data, from autonomous vehicles to industrial sensors to consumer assistants on phones, is going to run into a wall.

Bob: The executives quoted are essentially arguing that you cannot bolt world-class AI onto second-tier telecoms. And there is a European lens here, because the UK has been positioning itself post-Brexit as a lighter-touch, faster-moving AI hub compared to the EU. This warning suggests that regulatory speed does not compensate for physical infrastructure gaps. Meanwhile countries like Germany, France and the Nordics have been investing more consistently in fibre and 5G rollout, and they may end up being the more attractive base for AI-heavy industries.

Samantha: What this makes clear is that the AI race is not only about who has the biggest model. It is also about who has the boring, expensive, decade-long infrastructure to actually deliver it to users. For the UK government, the pressure now is to move telecoms upgrades up the priority list, or watch investment quietly drift toward markets with faster networks.

The US Federal Reserve puts AI at the centre of its economic policy deliberations.

Bob: Story four, and this one is more subtle but arguably has the longest tail. The US Federal Reserve is increasingly treating artificial intelligence as a central factor in its economic policy discussions. What used to be a side topic in speeches is now, according to reporting, a starring role in how officials think about growth, employment, inflation, and financial stability.

Samantha: And that shift matters, because central banks generally do not talk publicly about a technology unless they think it can move the numbers they are paid to manage. The Fed is looking at AI on multiple fronts. On productivity, whether it lifts output enough to change potential growth. On employment, whether it displaces workers faster than the labour market can absorb. On inflation, whether it puts downward pressure on services prices. And on financial stability, whether the enormous capital spending going into AI data centres is starting to look like a bubble.

Bob: For Europe, the read-across is that the European Central Bank and national central banks are almost certainly having the same internal conversations. If the Fed publicly integrates AI into its forecasting models, expect Frankfurt and the Bank of England to follow. And that has practical consequences. If policymakers start treating AI-driven investment as a structural growth story, that changes how they think about interest rates and how they assess whether markets are overheated.

Samantha: There is also a supervisory angle. European banking regulators have been increasingly vocal about concentration risk in AI infrastructure lending, where a small number of hyperscalers are absorbing enormous amounts of capital. If the Fed formalises AI as a macro-relevant force, that gives European counterparts more room to do the same. The bottom line is that AI has quietly moved from a tech story to a monetary policy story, and that shift will show up in central bank speeches on both sides of the Atlantic in the coming months.

Two things worth trying today, the offline European maps app Magic Earth and the open-source writing tool Zettlr.

Bob: To close out, two things you can actually use today. The first one is a navigation app called Magic Earth. It is a European-built maps and navigation service, developed by a team in Romania, and it is free on iOS and Android. What makes it stand out is that it is genuinely offline-first. You download the maps for the countries you want, and after that, turn-by-turn driving, walking and cycling navigation works with no data connection at all. Users report that coverage across Europe is strong, including rural areas where big-name maps apps get patchy.

Samantha: The privacy angle is the other reason it is worth a look. Magic Earth's developers say the app does not collect personal data or build a profile of your movements, which is a genuinely different posture from the dominant navigation apps. Reviewers highlight that traffic information is anonymised, and that there is no account or login required to use it. So if you are planning a road trip, a hike, or a weekend somewhere with spotty coverage, this is one of the few apps where the offline mode is not an afterthought.

Bob: Setup takes about five minutes. Install the app, open the map, and download the regions you want. For most European countries that is a few hundred megabytes each. After that you can put the phone in airplane mode and still navigate. It also handles lane guidance, speed limits, and voice instructions offline. Widely used by cyclists and campervan travellers who spend real time away from mobile signal.

Samantha: And it costs nothing, no subscription, no in-app purchases. For a Sunday, that is a pretty low-friction thing to install and have ready for the next trip.

The free open-source writing tool Zettlr helps you organise long documents and research from your own laptop.

Bob: The second thing to try is a writing tool called Zettlr. It is free, open-source, and it runs on Windows, Mac and Linux. Zettlr is aimed at people who write long, structured documents. Think theses, reports, articles, book chapters, anything where you need to keep track of sources and structure across days or weeks of work.

Samantha: What makes it interesting is that it combines a calm, distraction-free writing surface with proper research features underneath. You can link notes to each other, tag them, and pull them into a longer document later. It handles footnotes, citations and reference managers like Zotero out of the box. Academics and journalists are the group that tends to swear by it, but reviewers point out that it works equally well for anyone drafting a long report or a business proposal.

Bob: The privacy angle here is that everything lives on your own machine. There is no cloud sync, no account, no company reading your drafts to train a model. If you want to back up your work, you use your own folder in Dropbox or on a drive. That is a genuinely different bargain from the big cloud-based writing tools, and it will appeal to anyone handling sensitive material.

Samantha: Today we covered: Gartner's one point six trillion dollar semiconductor forecast, over twenty-two thousand humanoid robots shipped in the first half of 2026, the UK's telecoms bottleneck threatening its AI ambitions, the US Federal Reserve putting AI at the centre of monetary policy, and two things to try yourself: the offline European maps app Magic Earth, and the open-source writing tool Zettlr.

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.