#ai-infrastructure

Oregon regulators approved a new large-load tariff for Portland General Electric that shifts more infrastructure and interconnection costs to big data centers and ties approvals to clean-energy rules.

A Utility Dive opinion argues that power companies should treat the AI data center surge as a systemwide grid modernization challenge, not a queue of isolated projects.

Elon Musk’s xAI has added 19 more portable gas turbines at its Southaven, Mississippi, data center site, intensifying scrutiny over whether the company is sidestepping air permitting rules while scaling AI infrastructure

A sharply framed Electrek report argues that gas use for AI infrastructure clashes with long-standing solar rhetoric, underscoring a widening debate over how the AI buildout will actually be powered.

A newly approved Utah data center project has become a flashpoint in the debate over AI infrastructure, power demand, water stress, and local environmental impact.

Google says its new Missouri data center project will be paired with support for added power capacity, a $20 million energy impact fund, and local workforce training programs.

Nvidia says it will split data center revenue into hyperscalers and a broader enterprise category, a reporting shift aimed at showing investors its business is not tied only to the largest AI cloud buyers.

Researchers say algorithm-designed copper cold plates could reduce cooling’s share of data center electricity use from roughly 30% to about 1.1%, addressing one of AI computing’s fastest-growing energy burdens.

A new BloombergNEF outlook described by TechCrunch argues that solar is on track to become the world’s largest source of power by 2035, even as AI data centers create new demand that could keep gas and coal in the mix.

Rising military use of fiber-guided drones and surging demand from AI-linked data center construction are driving sharp increases in fiber-optic cable prices.

A new NEMA forecast says annual U.S. electricity consumption will rise more than 55% by 2050, with data centers driving a major share of near-term growth and electric transportation surging over the longer run.

Startup SPAN is testing a distributed data center model that would place liquid-cooled GPU nodes alongside homes, offering residents subsidized electricity, internet access, and backup batteries in exchange for hosting.

A Wired analysis argues that Nvidia’s enduring advantage in AI comes less from chips alone than from CUDA, the software platform that made those chips indispensable to parallel computing.

OpenAI, working with AMD, Broadcom, Intel, Microsoft, and NVIDIA, has released a new networking protocol through the Open Compute Project aimed at reducing congestion and failures in giant AI training clusters.

Ahead of the AI & Big Data Expo in San Jose, HP used an interview with its AI and Data Science business development lead to frame its enterprise focus around AI, processing, and data strategy.

NERC has issued a rare Level 3 alert after data centers unexpectedly dropped load or rapidly oscillated demand, warning that computational loads pose immediate reliability risks for the power grid.

Nicolas Sauvage says the best venture bets can take four years to look obvious. That thesis has shaped TDK Ventures’ push into AI inference, batteries, and grid hardware long before those areas became crowded.

A reported venture tied to Coatue is buying land near major power sources for potential data center development, highlighting how AI demand is redrawing the relationship between capital, energy and real estate.

Fusion startup Zap Energy says it will also develop fission reactors, arguing the route to grid-relevant power can move faster as AI-driven electricity demand climbs.

Google engineers say a new compression method called TurboQuant can cut AI working-memory needs by up to six times without sacrificing model performance, potentially easing one of the infrastructure burdens of large chat