Search

Factories take energy into their own hands. The problem starts later

Translated title. Original: Fabryki biorą energię we własne ręce. Problem zaczyna się później

Industrial companies now manage energy like their core operations, using a mix of long-term PPAs, spot purchases, and storage to hedge volatility; renewables help control costs, while specialized teams and AI-driven tools manage a more complex, hybrid energy supply landscape.

DARPA Backs Project to Measure AI Energy Use at Every Stage of the Workflow | OilPrice.com

DARPA-backed project E-CALC will fund researchers at Michigan Engineering, UIUC, and BAE Systems with $1.83 million to measure and optimize AI energy use across the entire workflow, addressing high uncertainty in current energy estimates and potential 30% waste from software-hardware misalignment, with implications for civilian and defense sectors.

OpenAI deploys a Control Room to stop its AI models

Translated title. Original: OpenAI déploie une Control Room pour arrêter ses modèles d'IA

The text says AI needs more energy, and demand is rising faster than the United States has ever produced it.

Energize - Pilot your stock, master your budget

Translated title. Original: Energiz - Pilotez vos stocks, maîtrisez votre budget

An energy-tracking app that automatically analyzes consumption, forecasts depletion dates using recent history and local weather, integrates Netatmo sensors, groups costs by heating season, and offers a portable, full-screen mobile experience without manual data entry.

How data center operators hide information about energy consumption | heise online

Translated title. Original: Wie Betreiber von Rechenzentren Angaben zum Energiebedarf verstecken | heise online

The page notes that an English version is available and that its translation received technical assistance and editorial review.

Sustainable Compute in US Data Centers: AI Efficiency

Sustainable compute aims to deliver more computing work per unit energy in US data centers by optimizing hardware, models, and workloads, acknowledging that efficiency does not automatically reduce total energy demand if demand grows rapidly. Effective implementation combines multiple measures—hardware, software optimization, utilization, scheduling, and thermal management—to maximize useful output within grid and infrastructure limits.

Water usage at AI scale - DCD

Water usage at AI scale is influenced by cooling design, operating temperatures, and heat rejection. A Schneider Electric whitepaper compares four 100 MW cooling designs across climates, highlighting how liquid cooling and temperature optimization can cut water use while boosting energy efficiency.

Energy intelligence: the future energy system foresees and acts - technocrat | technocrat

Translated title. Original: Energiaintelligencia: a jövő energiarendszere előre lát és cselekszik - technokrata | technokrata

Energy intelligence aims to unite energy data, processes, and assets to forecast and prevent grid issues, automate responses, and integrate variable renewables. It requires synchronized development of generation, networks, storage, and consumption, aided by AI and digital twins, to create a reliable, adaptive energy system.

Energy Shortage: Microsoft Exits Nscale’s UK Loughton AI Data Center Deal

Microsoft has pulled out of Nscale's Loughton AI data center in the UK, leaving the project with no publicly announced customers as power supply remains uncertain until the 2030s. Nscale continues to pursue alternative power sources while maintaining its broader partnership with Microsoft and funding efforts ahead of a potential public listing.

Aydın October 20 Electricity Outage! Which districts will lose power on Tuesday?

Translated title. Original: Aydın’da 20 Ekim Elektrik Kesintisi! Salı Günü Hangi İlçelerde Elektrikler Kesilecek?

The document warns residents of planned electricity outages in Aydın on Tuesday, 20 October 2026, due to maintenance and grid upgrades. It lists affected districts and neighborhoods and advises residents to monitor official notices, prepare by charging devices and backing up data, and verify schedules from the distribution company before assuming exact outage times.