Generate $6 trillion in annual revenue or face a major collapse — AI data center costs are doubling every 12 months
THE SO WHAT
When regulators yank a 'completeness' finding after 564 changes to a power application, the constraint on AI buildout is clearly grid and permitting, not GPUs. If your roadmap assumes cheap, continuous AI capacity, you need a parallel plan for power delays, local inference, and workload triage.
READ THE SOURCE
MORE FROM THE WIRE
Deep & Emerging TechChina wants solid-state batteries in use by 2030, and CATL rates them four out of nine
Beijing setting 2030 for large-scale solid-state use while CATL calls million-vehicle deployment before then 'very unlikely' is a gap between policy ambition and industrial readiness. Auto and storage operators should treat solid-state as an options bet — keep pilots and supplier relationships warm, but don't anchor 2030 product plans on it.
Deep & Emerging TechDigitalBridge CEO Marc Ganzi says his data center investment group will become SoftBank's "third-party infrastructure arm" after SoftBank's ~$4B takeover closed
SoftBank effectively stapling a $4B data center platform onto its portfolio turns it into an end-to-end AI capital + infra shop—owning both the equity and the racks. If you’re negotiating large-scale AI workloads, expect SoftBank-backed companies to come with bundled infra options and potentially different pricing power than pure software peers.
Deep & Emerging TechGarlic Peel Sensor Powers Low-Cost Home Security Sensor
Biowaste-based sensors that generate their own high voltages hint at ultra-cheap, low-power security and monitoring networks. Worth monitoring if you're in IoT or smart home—materials innovation like this can reset BOM assumptions and open up disposable or maintenance-free sensor designs.
Deep & Emerging TechRural Data Centers Are in for a Big Federal Tax Break
Federal tax breaks for rural data centers shift the siting equation from pure latency and talent access toward power, land, and incentives — but the fact that some hyperscalers are hesitant tells you permitting, workforce, and grid reliability still dominate the model. If you’re planning AI-heavy capacity, underwrite local infrastructure and political risk as hard as you underwrite the tax benefit.