IEA expects global data center investment to hit $4T by 2030; fewer than 100 major AI projects would require a combined 50GW of power, matching peak UK demand
THE SO WHAT
A projected $4T in data center capex by 2030 and sub-100 AI projects needing 50GW—UK peak demand—turn AI from a software story into a grid-planning problem. If you’re building AI-heavy products, power availability and siting now belong in your strategic planning alongside model choice.
READ THE SOURCE
MORE FROM THE WIRE
Deep & Emerging TechGoogle-Tied Prepaid Bond Deals Delayed Due to Lackluster Demand
The delay of $3B in Alphabet-linked prepaid energy deals suggests even marquee tech credits can’t fully insulate structured energy financing from rate and volatility risk. If you’re counting on prepaid or project bonds to underwrite data center power, build in timing and pricing slack.
Deep & Emerging TechGoogle Quantum AI tests quantum contextuality on up to 105 qubits
Demonstrating contextuality across 105 qubits without claiming classical advantage is a hardware and verification milestone, not a commercial one. If you’re in HPC or crypto, this is a nudge to keep quantum risk on the 5–10 year horizon, not to re-architect workloads this quarter.
Deep & Emerging TechNscale’s ByteDance Deal Faces Scrutiny Under US Chip Curbs
Regulators scrutinizing Nscale’s AI compute sale to ByteDance under new US chip curbs shows export control is now about services and capacity, not just hardware. If you sell cloud or AI infrastructure with any China exposure, you need a live map of regulatory risk by customer and region.
Deep & Emerging TechChase Would Like You to Know Its Latest Cybersecurity Issue Is a Whole New Problem
Another major bank breach tied into a broader Mastercard incident underscores how tightly coupled financial data supply chains have become. Treat your exposure as ecosystem-level—vendor, processor, and partner security posture now matters as much as your own.