TikTok $400M Privacy Settlement: EU Impact & AI Market

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

Bob: And I'm Bob Russell. Today: TikTok settles a massive US children's privacy case for four hundred million dollars, the UN forecasts the global AI market will hit four point eight trillion dollars by 2033, Micron pours ten billion into a new AI memory research lab, Brazil builds its own sovereign AI supercomputer, and to close out, two things you can actually use today: the encrypted European messaging app Element, and the free open-source home inventory tool Homebox. Let's get into it.

TikTok agrees to a landmark four hundred million dollar settlement over US children's privacy violations.

Samantha: TikTok has agreed to pay four hundred million dollars to settle a US Department of Justice lawsuit accusing it of breaking federal children's privacy laws. This is one of the biggest ever penalties under the Children's Online Privacy Protection Act, known as COPPA. The Justice Department and the Federal Trade Commission sued back in 2024, arguing that TikTok collected personal data from children under thirteen without asking their parents, and then failed to properly inform those parents about what was being gathered. Three hundred million is being paid immediately, with another hundred million to follow once a previous consent decree against Musical dot ly, TikTok's predecessor, is formally vacated.

Bob: And this is not just an American headache. The size of the number matters, because it sets a benchmark that European regulators can now point to when they push back on the same platforms. The Irish Data Protection Commission has already fined TikTok multiple times over how it handles children's data under GDPR, and the EU's Digital Services Act adds another layer of pressure specifically around minors. What this settlement really does is confirm the direction of travel: platforms that treat under-thirteens as just another audience segment are running out of runway. Expect European enforcement actions to cite this figure repeatedly.

Samantha: The practical question for European parents is whether TikTok's changes actually reach beyond US borders. Historically the company has updated its systems globally when forced to comply with one major jurisdiction, because running two separate versions is expensive and legally messy. So there's a decent chance the tightened age gates and consent flows will show up here too, even without a direct European ruling. The bottom line is that the cost of ignoring children's privacy just went up sharply, and that recalibrates every social platform's risk model.

The UN projects the global AI market will hit four point eight trillion dollars by 2033, a twenty-five-fold jump.

Bob: A new report from UN Trade and Development, UNCTAD, forecasts the global artificial intelligence market will reach four point eight trillion dollars by 2033. That's a twenty-five-fold increase in just ten years. The report highlights genuine upside: AI already helping rural health workers, personalising education in under-resourced schools, and improving climate disaster forecasting. But UNCTAD also flags a serious concentration problem. A small group of economies and roughly one hundred firms dominate AI research and computing power, and up to forty percent of jobs worldwide could be affected by AI-driven automation.

Samantha: For Europe this cuts two ways. On one hand, the bloc has positioned itself as the global standard-setter on AI governance with the AI Act, so a report calling for inclusive governance plays directly into Brussels' hands. On the other hand, when you look at who actually owns the computing power and the frontier models, Europe barely features on that list of one hundred firms. So the continent risks being a rule-maker without being a market-maker. UNCTAD's call for global cooperation is polite diplomatic language for: the current distribution is unsustainable.

Bob: And the jobs number is the one that will hit political conversations hardest. Forty percent of jobs affected does not mean forty percent disappear, it means transformed, restructured, or displaced. That's still a huge social challenge, especially for economies with rigid labour markets. The report essentially argues that without deliberate policy choices, the benefits of this four point eight trillion dollar market concentrate in a handful of places while the disruption spreads everywhere. That is a warning European finance ministers should be reading this weekend.

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.

Micron commits ten billion dollars over a decade to a new AI memory research lab in Idaho.

Bob: Micron Technology is investing ten billion dollars over the next ten years in a new research facility in Boise, Idaho. It's being called Micron Research Labs, and the focus is specifically on advanced memory technologies and the compute systems that AI accelerators depend on. Groundbreaking is planned for 2027, and the lab is designed to host hundreds of researchers from industry, academia, and government. The context here is high-bandwidth memory, or HBM, which has become one of the most sought-after components in the AI supply chain. Every AI accelerator built by Nvidia, AMD or the hyperscalers needs it.

Samantha: For European industry this matters because memory has quietly become one of the biggest bottlenecks in the whole AI infrastructure story. Chip fabrication gets the headlines, but without enough advanced memory, those chips just sit there waiting for data. Europe has been trying to build its own semiconductor resilience through the Chips Act, but almost all serious HBM production sits in South Korea, the US, and increasingly Japan. A ten billion dollar research investment in Idaho reinforces where the deep expertise is being built, and it isn't in Europe.

Bob: The timing is telling too. Memory prices have been climbing sharply because AI demand has outstripped supply, and there's no easy short-term fix. Building a research lab is a ten-year bet, not a next-quarter fix. What it signals is that Micron sees AI memory demand as a durable, structural growth story, not a bubble. For European carmakers, industrial firms and cloud providers who all need this hardware, the practical takeaway is that securing supply contracts now, ahead of the crunch, is becoming a strategic priority.

Brazil moves to build its own national AI supercomputer to cut dependence on foreign providers.

Samantha: Brazil is pushing ahead with plans to acquire a national AI supercomputer, aiming to reduce its reliance on foreign providers for both storage and processing. The system will be made available to Brazilian companies, universities, research institutes and government agencies. The National Laboratory for Scientific Computing, known by its Portuguese abbreviation LNCC, will run the project. Alongside the hardware, there's a cooperation agreement to develop Brazilian alternatives for data infrastructure and to promote open-source models as part of a broader push on global AI governance.

Bob: This is part of a wider global pattern that Europe should be paying close attention to. Countries are increasingly treating AI compute the way they treat energy grids: as national infrastructure that you cannot afford to outsource entirely. France has been building EuroHPC systems, Germany has its Gauss Centre, and the European Commission is funding AI factories across member states. But Brazil's move shows that even mid-sized economies now consider a sovereign AI compute base a baseline requirement, not a luxury.

Samantha: The open-source angle is what makes this interesting beyond the hardware. Brazil is explicitly framing its supercomputer as a way to contribute to open AI models, rather than just consume closed ones from American or Chinese vendors. That aligns with what a lot of European researchers have been arguing: open weights and open datasets are the practical route to sovereignty for countries that cannot outspend the hyperscalers. Whether the LNCC actually delivers on that promise depends on execution, and public-sector tech projects have a mixed track record everywhere.

The encrypted European messenger Element gives you serious secure chat without a phone number.

Bob: To close out, two things you can actually use today. The first is Element, an encrypted messaging app built on the open Matrix protocol, which originated in the UK. It's used by the French and German governments for internal secure communication, so it's proven in serious environments. What makes it useful for regular users is that you don't need to hand over a phone number to sign up, unlike Signal or WhatsApp. You create an account with just a username, and everything is end-to-end encrypted by default.

Samantha: Element also supports group rooms, voice and video calls, and cross-device sync, and it runs on Windows, macOS, Linux, iOS and Android, plus straight in the browser. Because Matrix is a federated protocol, you can even self-host if you want to, though most users will just sign up on the default server. Users describe it as a middle ground between Slack and Signal: more collaborative than a pure messenger, more private than a corporate chat tool. Free to use, open source, and hosted in Europe.

Bob: One useful detail: because it's federated, different organisations can run their own Element servers but still talk to each other, similar to how email works between different providers. That makes it a genuine alternative for community groups, small businesses, or families who want a shared space that isn't owned by a big US platform. Worth trying this weekend if you've been meaning to move a group chat off WhatsApp.

Homebox turns your home inventory into a searchable database, entirely on your own machine.

Samantha: The second pick is Homebox, a free open-source home inventory tool. It's aimed at anyone who wants to keep track of what they own, whether that's for insurance, for organising a move, or just knowing which drawer the spare HDMI cable ended up in. You add items with photos, categories, locations, purchase dates, warranty info, and receipts. Everything stays on your own device or a small server you run yourself. No subscription, no cloud, no account.

Bob: What makes Homebox stand out is how lightweight it is. It runs in a small container on almost any computer, including a Raspberry Pi, and you access it through a normal web browser on your phone or laptop. Users often mention that it's genuinely useful the first time something breaks, gets stolen, or needs a warranty claim, because you have the receipt and serial number right there. It also handles labels and QR codes, so you can stick a code on a storage box and scan it later to see what's inside.

Samantha: It's the kind of tool that sounds boring until you actually need it. For anyone who has ever tried to file a home insurance claim from memory, Homebox is a much calmer way to prepare in advance. Free, private, and it works offline. A good project for a Saturday afternoon.

Samantha: Today we covered: TikTok's four hundred million dollar children's privacy settlement, the UN forecast that AI becomes a four point eight trillion dollar market by 2033, Micron's ten billion dollar AI memory research lab, Brazil building its own sovereign AI supercomputer, and two things to try yourself: the encrypted European messenger Element, and the open-source home inventory tool Homebox.

Bob: Want to know more or react? Visit stateoftech dot eu or email us at info at doorzetters dot net.

Bob: State of Tech, the tech world in 15 minutes.