Anthropic buys $3.4bn of gas generators for its data centres, FT reports
THE SO WHAT
Locking in 1.7GW of gas‑fired backup for $3.4B makes clear that hyperscale AI is moving toward self‑provisioned, fossil‑heavy power to de‑risk grid constraints. Enterprises with decarbonization targets need to reconcile AI expansion plans with the reality that their upstream suppliers may be leaning into gas, not renewables, to meet demand.
READ THE SOURCE
MORE FROM THE WIRE
Applied AIPartnering with Edelbyte: Data Infrastructure That Keeps Up With AI's Curiosity
If agents are going to hammer your internal systems with orders of magnitude more queries than humans ever did, the real constraint becomes data infra cost and latency, not model quality. Teams building agents should be budgeting for a new class of OLAP-style infra like Edelbyte’s, not assuming the existing warehouse can take the load.
Applied AISources: John Ternus wants more urgency in Apple's AI efforts, pushing Craig Federighi to move beyond Siri AI tweaks to agentic and other advanced capabilities
If Apple leadership is explicitly pushing from “Siri tweaks” to agentic capabilities, the default UX on a billion-plus devices is being re-scoped from Q&A to task execution. If you build consumer apps in Apple’s ecosystem, assume background agents and delegated workflows become a first-class surface and design for being called—not tapped.
Applied AITencent AI chief warns models will soon pose biology risks, SCMP reports
When a major lab’s chief scientist says AI models may soon pose biology risks comparable to current cyber risks, the bar for internal governance and red-teaming just moved. If you touch life sciences, chemicals, or bio data, assume regulators and counterparties will start asking for concrete AI misuse controls, not just generic security language.
Applied AIUnfold CEO Ran Cherniavsky on breaking down enterprise data barriers to AI adoption
The real drag on enterprise AI isn’t model quality—it’s data locked in legacy systems and proprietary apps with no modern interfaces. Before you chase the next model upgrade, inventory the systems where your critical data is effectively unreachable and budget for extraction and normalization work.