'Makes it even more disappointing': Microsoft backs fossil fuel big time with $7 billion deal in race for AI supremacy
THE SO WHAT
A $7B methane gas deal to feed AI data centers makes the tradeoff explicit — hyperscale AI is currently being underwritten by fossil infrastructure. If you’re building on top of this stack while claiming climate alignment, you need a real energy strategy and disclosures, not just offsets and marketing copy.
READ THE SOURCE
MORE FROM THE WIRE
Deep & Emerging TechBlue Origin one step closer to launching New Glenn from Vandenberg Space Force Base
A second heavy-lift provider at Vandenberg reshapes West Coast launch capacity — polar and sun-synchronous orbits get more competitive pricing and better cadence. If you’re planning EO or defense constellations, revisit launch assumptions and negotiate multi-provider terms while capacity is still being allocated.
Deep & Emerging TechBYD Is Rolling Out 5-Minute Charging on Several Popular EVs
10% to 70% in five minutes turns charging from a planning problem into a coffee-break problem — that’s a structural shift in EV behavior and infrastructure design. Fleet operators and retailers should be modeling what happens when dwell time collapses and throughput becomes constrained by grid and parking, not battery tech.
Deep & Emerging Tech'Maybe it’s not science fiction': Solar panels are causing rainwater to fall in one of the driest places on Earth
If large solar farms are measurably changing local rainfall patterns, energy siting becomes a climate-engineering decision, not just a land-use one. Utilities and developers need to model atmospheric impacts alongside grid and tax economics or risk a backlash that slows deployment in exactly the deserts they’re counting on.
Deep & Emerging TechASML Won’t Be an AI Bottleneck, Says Berenberg's Qiu
If ASML isn’t the constraint, the choke points move further down the stack—to power, packaging, and system integration. Infra planners need to stop using “lithography bottleneck” as an excuse and start modeling where the real limits to AI capacity will show up in 2027 and beyond.