Singapore’s living-neuron data center & Apple’s Siri delays hit hardware - Tech News (Mar 10, 2026)

First up: a headline that blurs the line between computing and biology. Singapore-based data center developer DayOne is teaming up with Cortical Labs to build what they’re calling a “bio data centre”—a facility designed to run some AI workloads on living neurons grown from stem cells. The pitch is simple and ambitious: dramatically lower energy use compared with traditional silicon. The early plan is small—starting inside a university setting with a prototype rack—then scaling only if performance, safety, and regulation line up. Beyond the wow factor, the real story is pressure: AI is pushing power grids and sustainability rules to the limit, and regions like Singapore are hungry for computing that doesn’t come with an equally massive energy footprint.

Staying with AI—but shifting to consumer tech—Apple’s smart home display is reportedly delayed again, and the reason isn’t the screen or the hardware. It’s Siri. Sources say the device has been largely ready, but the product relies on a new generation of Siri-driven personalization that pulls from a user’s own data to make the display feel like a true home hub. The new target is now later this year, hinging on whether Siri’s broader AI upgrade finally lands on time. The interesting takeaway is organizational, not gadgety: AI features are becoming the critical path, and they’re increasingly dictating when entire product categories can ship.

Now to policy: governments are moving toward tougher age checks online, and the trend is accelerating. After Australia’s teen-focused social media restrictions, regulators across Europe, Brazil, and multiple U.S. states are exploring stronger “age assurance” requirements—not just for social networks, but also for AI chatbots and adult sites. The argument is that tools have improved enough to make large-scale age gating feasible, using a mix of signals, verification steps, and sometimes face-based estimation. But the hard parts haven’t gone away: accuracy around legal cutoffs, uneven performance across different users and cameras, and the privacy risk of building identity-style checks into everyday browsing. If this wave continues, the internet may look a lot more like a set of gated venues—and a lot less like a single open lobby.

In a closely related “who gets to set the rules” story, Anthropic is suing the U.S. Department of Defense after the Pentagon labeled the company a “supply chain risk.” Anthropic says the designation is being used improperly—essentially as punishment for objecting to uses of AI in areas like mass surveillance or autonomous weapons. The stakes are big: that label can shut a vendor out of defense work, and Anthropic says contracts are already being canceled. What makes this especially notable is the precedent it could set. If supply-chain authorities become a pressure tool in AI policy disputes, it could reshape how AI firms speak up on safety—and how the government manages competition among vendors.

From government pressure to creator pressure: thousands of authors have released an “empty” book titled “Don’t Steal This Book,” containing little more than contributor names. It’s a pointed protest aimed at AI training practices and copyright policy, timed around the UK government’s upcoming work on how generative AI should be allowed to use copyrighted texts. The fight is over default rules: should AI companies be able to train unless writers opt out, or should permission and payment be the starting point? Publishers are also circling collective licensing schemes that could become the practical middle ground. Either way, this is one of the clearest signals yet that creators are organizing at scale, not just suing one company at a time.

Over in space—and more specifically, space industry pay—Blue Origin is reworking its stock option program after years of employee frustration. Reports say an older plan left many long-tenured staff with options that ultimately expired without value, largely because liquidity events never came. A new plan is expected to start granting options soon, and it’s reportedly designed with some form of liquidity opportunity in mind. This matters because space is in a talent war, and equity is a major recruiting lever—especially when competitors have already delivered payouts through private sales or other mechanisms. The subtext: if Blue Origin wants to keep pace, it needs compensation structures that look like the rest of the high-growth space sector.

Let’s zoom out to geopolitics and industrial strategy. China’s latest economic blueprints, unveiled at the National People’s Congress, put extra emphasis on building up domestic demand while also doubling down on technological breakthroughs. The long-term goal is clear: strengthen self-sufficiency in areas like AI, chips, robotics, biotech, and next-generation networks—especially as export-heavy growth runs into tariff risk and broader trade friction. The balancing act is tricky: subsidies and rapid scale can create oversupply, which then pushes more exports, which can trigger even more backlash abroad. But the direction of travel is unmistakable: Beijing is treating advanced tech as both an economic engine and a strategic shield.

Back in the U.S., the FAA has approved a set of pilot programs to expand real-world testing for electric aircraft, including the vertical-takeoff category often pitched as air taxis. The key significance here is tempo. Certification is slow and expensive, and these pilots are meant to gather operational data sooner—across cargo, emergency response, regional travel, and urban concepts. If the projects produce credible safety playbooks, they could help the sector graduate from flashy demos to routine operations. If they don’t, they’ll still answer a question investors and cities keep asking: is this category ready to behave like aviation, not just like a prototype showcase?

Now for a regulatory fight that sounds futuristic but has very real near-term consequences. Amazon has asked the FCC to reject SpaceX’s request tied to an “orbital datacenter” concept—essentially, compute infrastructure in space at an enormous satellite scale. Amazon argues the filing lacks core details and raises concerns about interference, debris risk, and environmental impact. Behind the paperwork is a larger battle over who gets to define the next phase of orbital infrastructure: not just connectivity, but potentially computation as well. Even if space-based data centers never become economical, the FCC’s response could set a precedent for how aggressively regulators demand specifics before approving mega-constellations.

And finally, a neuroscience story that’s equal parts fascinating and unsettling. Researchers have reconstructed short, grainy “movies” from recordings of neural activity in mice while they watched brief video clips. The results aren’t crisp—more like a pixelated hint than a clear playback—but the method points to a future where scientists can test theories of perception in a more direct, data-driven way. The upside is huge for basic research: understanding how brains represent the world, and maybe even what animals experience in dreams or illusions. The longer-term shadow is also obvious: as these techniques improve, “neural privacy” stops being a philosophical concept and becomes a policy problem.