Today’s Tech Brief is dominated by one central thread: the infrastructure and platform race behind AI—who pays for the power, who controls the chips, and who gets distribution at the device and national-security layers.

Microsoft laid out a plan to absorb incremental electricity costs tied to its data-centre buildout and to emphasize local transparency and tax-base support, a bid to de-risk community backlash while scaling capacity.

Meta is reshuffling priorities by cutting part of its metaverse team and leaning harder into data centres and its Ray-Ban smart glasses line, with talk of sharply increasing production volumes of the hardware used to access voice-based AI.

A potential US–Taiwan trade deal is being discussed that would reduce tariffs on Taiwan and expand TSMC’s planned chip production footprint in the United States, with additional Arizona plants described as part of the long-range buildout.

The strategic tension is that deepening US–Taiwan chip alignment can complicate already-fragile trade dynamics with China, at a moment when rare-earth supply and broader supply-chain continuity remain high-stakes leverage points.

In Iran, a nationwide internet shutdown entered a fifth day as President Trump called for Starlink to help restore communications; Iranian authorities have declared Starlink illegal, leaving the operating conditions and outcomes uncertain.

Malaysia moved to take legal action involving Elon Musk’s X and xAI shortly after the country banned Grok over sexually explicit images involving women and children, putting content governance and platform compliance in the crosshairs.

SK hynix pledged to build a new memory-chip packaging plant targeted for completion by 2027, underscoring that memory and packaging capacity—not just GPUs—are becoming critical bottlenecks in AI systems.

Intel and AMD shares rose on a note that both are sold out of CPU allocation, a reminder that servers need large volumes of CPUs alongside accelerators—and supply constraints can quickly become a growth limiter.

Major banks are explicitly tying near-term spending to AI adoption: JPMorgan described ongoing investment to drive efficiency and keep pace, with the industry still early in translating deployment into hard ROI.

Apple is moving to embed Google’s AI into parts of Siri functionality, a distribution-level shift that signals urgency: access to best-in-class models and tooling matters more than owning every layer of the stack, especially at iOS scale.