Musk's orbiting data centres create more problems than they solve
Plus, Russia cracks down on messaging apps to drive users towards towards the Kremlin's new 'superapp', Max
Welcome to the latest edition of ASPI’s Cyber & Tech Digest.
Each week, ASPI curates and contextualises the most important developments in cyber, technology, and geopolitics — highlighting what matters and why.
This edition covers the period: 7 February 2026 to 13 February 2026.
Follow the Australian Strategic Policy Institute on Bluesky, LinkedIn, and X.
What We’re Tracking
Elon Musk moves to merge xAI and SpaceX for orbital AI data centres
What happened: Elon Musk told employees at xAI that the company needs a factory on the moon to build A.I. satellites, along with a mass driver to launch them into space, according to the New York Times. He said xAI is merging with SpaceX to support the creation of A.I. data centres in outer space.
Musk also described a self-sustaining city on the moon as a steppingstone to Mars. In the same remarks, he discussed plans to grow X and add features aimed at increasing daily active users.
The Financial Times reports that Musk aims to launch solar-powered A.I. data centres into orbit within three years. The concept would rely on satellites cooled by the vacuum of space to generate computing power.
Other technology leaders, including Jeff Bezos and Google, are exploring similar space-based A.I. initiatives, with prototype satellites planned for launch by 2027.
Why we’re tracking this: Moving advanced computing infrastructure into orbit would test the economic logic of off-world industrialisation, not just its technical feasibility.
It would also expand competition over who controls the physical backbone of frontier A.I. systems.
What people are saying:
Industry experts told the Financial Times that the economics of orbital data centres may become viable before the technology is fully ready.
The Financial Times notes high launch costs, radiation shielding and cooling as persistent constraints.
The New York Times reports that Musk framed a lunar factory as necessary to scale A.I. satellite production.
My view: The technical feasibility of placing computing infrastructure in orbit is not the central question; the economic logic is. Terrestrial data centres may face resource and land-use constraints, but orbital facilities introduce many more additional costs and risks, including launch expense, debris avoidance and structural scaling challenges, without a clearly defined problem they uniquely solve. Based on the reporting so far, there is no demonstrated use case that offsets this added complexity. That assessment would change only if proponents show a durable economic advantage, not merely a working prototype. Musk has proven his detractors wrong before, so I wouldn’t rule his latest scheme out entirely. But a healthy dose of skepticism is rightly warranted.
— Stephan Robin, CTS
What We’re Watching
A weekly scan of notable developments we’re tracking across technology, policy, and geopolitics.
⸻
🛰️ Strategic competition
The European Commission proposed banning all cryptocurrency transactions with Russia as part of a broader sanctions package targeting banks, the digital rouble, and sanction-evasion routes through third countries.
Meanwhile, EU officials and chip executives warned that the United States remains dependent on Europe’s advanced semiconductor tooling, particularly ASML’s extreme ultraviolet lithography machines, as Brussels doubles down on its Chips Act ambitions.
A Dutch appeal court upheld the suspension of Wingtech’s leadership at Nexperia, reinforcing European control over the Chinese-owned chipmaker amid supply security concerns.
⸻
📡 Cyber posture
Singapore’s government removed the China-linked group UNC3886 from the networks of its four major telcos after an 11-month operation targeting core infrastructure. Separately, security researchers reported that the same threat actor exploited zero-day vulnerabilities to siphon technical network data from Singtel, StarHub, M1, and Simba.
At the same time, Microsoft patched multiple actively exploited zero-day flaws in Windows and Office, including vulnerabilities enabling high-privilege malware execution.
Elsewhere, analysts argued that Chinese cyber operations often prioritise persistent observation over disruption, quietly mapping adversary systems for strategic insight.
Chainalysis reported that cryptocurrency flows linked to suspected human trafficking services rose 85% year over year in 2025, with stablecoins increasingly used in Southeast Asia–based scam compounds.
Meanwhile, Israeli authorities charged a reservist and civilian for allegedly using classified military information to place bets on Polymarket.
⸻
🛡️ Surveillance states
Iran’s government expanded its National Information Network, strengthening its capacity for internet shutdowns, behavioural profiling, and domestic traffic control.
Meanwhile, Roskomnadzor began throttling Telegram as Russia promotes a state-backed “super-app” and tightens pressure on foreign messaging platforms.
⸻
⚖️ Platform accountability
Communications Minister Anika Wells sought an urgent meeting with Roblox and asked the eSafety Commissioner to assess enforcement powers after reports of grooming and explicit content on the platform. Separately, the Australian government demanded a review of Roblox’s PG rating, as regulators test compliance with private-by-default settings and restrictions on adult contact with children.
Julie Inman Grant, Australia’s eSafety Commissioner, was named to Time’s top 100 influential leaders in health for shaping Australia’s under-16 social media ban. Meanwhile, advocacy groups warned that the same ban risks isolating children with disabilities who rely on digital communities for connection.
In Brussels, the European Commission issued a warning to Meta over policies that block rival AI assistants on WhatsApp, signalling potential interim measures under EU competition rules.
⸻
🧠 AI power plays
OpenAI began testing ads in ChatGPT for free-tier users in the US, framing the move as a way to fund broader access while keeping conversations private. Meanwhile, a former OpenAI researcher resigned in protest over advertising incentives, warning the shift risks replicating the privacy trade-offs of Facebook’s ad-driven model.
At the same time, OpenAI granted the Pentagon access to ChatGPT through the Genai.mil program for “all lawful uses,” despite internal debate over military deployment. Separately, OpenAI disbanded its mission alignment team, redistributing members as part of a broader organisational reshuffle.
Elsewhere, Anthropic committed $20 million to a super PAC backing candidates favouring stronger AI guardrails, escalating its policy contest with OpenAI-aligned groups.
⸻
🚀 Space compute
Elon Musk announced a reorganisation at xAI following a merger with SpaceX, consolidating ambitions to integrate AI and launch infrastructure. Separately, Musk outlined plans for an AI satellite factory and data centres on the moon, positioning space-based compute as a strategic frontier.
At the same time, the Financial Times reported that multiple tech leaders are racing to deploy AI data centres in orbit within three years despite technical and economic hurdles.
⸻
🏗️ Infrastructure surge
Australian businesses doubled planned data centre investment to nearly $52 billion in six months as AI demand accelerates. Meanwhile, policymakers and developers debated the environmental and grid impacts of a proposed 1GW mega data centre in Western Sydney.
Elsewhere, major US tech companies prepared to spend more than $US660 billion this year on AI infrastructure, prompting a wave of expected bond issuance.
⸻
🎭 Culture
Religious leaders and technologists experimented with AI-powered chatbots and avatars in spiritual settings, blending generative systems with ritual and doctrine.
That’s all for this week. For more timely analysis and commentary, check out The Strategist and ASPI’s Stop the World podcast—or our other Substack newsletters:
