Amazon Bets on AI Phones While Blue Origin Launches Orbital Data Centers

Amazon is building another smartphone. Ten years after the Fire Phone disaster-a $170 million writedown that lasted 14 months-the company is back, this time betting that AI has fundamentally changed what a mobile device should be. Codenamed "Transformer," it's being developed by a dedicated team called ZeroOne inside Amazon's devices division, led by J Allard, the former Microsoft executive behind Xbox. The difference from 2014 isn't just the product. It's the premise: instead of packing in gimmicks like 3D displays, this phone would be AI-native, synced to Alexa, potentially bypassing app stores entirely. Amazon hasn't even approached wireless carriers yet-which raises a deeper question about what this device actually is.

The Infrastructure Play

That question matters more when you look at what Amazon has been quietly building. The company has two separate satellite projects now running in the same division: Leo (Project Kuiper), its broadband network with over 7,700 satellites approved by the FCC, and Sidewalk, its mesh networking protocol that uses Echo and Ring devices to create low-bandwidth coverage. A phone without carrier dependence starts to look less like a gamble and more like a natural extension of infrastructure Amazon already owns. The company is reportedly exploring both a conventional smartphone and a stripped-down "dumbphone" with minimal features-hedging its bets on what AI-first mobile actually means.

The Orbital Data Center Race Heats Up

Meanwhile, Blue Origin is placing its own hardware bet-on a massive scale. Jeff Bezos' space venture asked the FCC this week for permission to launch up to 51,600 satellites called Project Sunrise, designed to operate as orbital data centers. The idea is straightforward but radical: bypass the power and cooling constraints of Earth-bound AI infrastructure by moving compute into space. These satellites would use laser links to communicate with each other and with TeraWave, Blue Origin's previously announced 5,408-satellite broadband constellation. SpaceX fired back immediately-objecting to the filing and pointing out that Amazon (Bezos' original company, distinct from Blue Origin) had opposed SpaceX's own orbital data center application. It's a three-way fight between Bezos' companies and Musk, with Google, Axiom Space, and others circling the same opportunity. The timing is telling: as AI models grow hungrier for compute and cooling becomes the bottleneck, companies are literally reaching for space.

In quantum research, synchronisation problems that once seemed insurmountable are cracking. A new full-mesh network called XCOM has achieved 100 picosecond synchronisation between quantum control boards-a precision that was previously unattainable without major hardware upgrades. The breakthrough matters because coordinating multiple control boards has been a key bottleneck in scaling qubit counts. With better timing, larger and more complex quantum systems become possible. Separately, researchers have used low-power lasers (just 10 picowatts) to control material vibrations in ultra-thin layers, a breakthrough that could lead to new ways of manipulating materials with light. These aren't commercial products yet, but they're signals that quantum engineering is moving past dead ends.

What ties these stories together isn't just ambition-it's a recognition that the next generation of computing infrastructure will be fundamentally different. AI needs different hardware. Quantum needs different approaches to control. Mobile needs rethinking. The companies making those bets are the ones positioning themselves for what comes next.