Top Stories
All Sections
Finance
About
20 results for "data center"
A new colocation model is placing AI data centers directly at renewable energy generation sites, bypassing multi-year grid interconnection queues and concentrating AI compute near power sources rather than population centers.
Cornelis Networks raised $205M for its Active Compute Fabric — GPU-agnostic networking that processes AI data in transit rather than stalling GPUs on synchronization overhead.
BloombergNEF data shows co-located solar-plus-storage investment hit a record $25 billion in H1 2026 — nearly double H2 2025 — as AI data centers race to secure reliable clean power without waiting years for grid interconnection.
New research argues that physics-informed AI—models constrained by actual power system laws—outperforms standard neural networks for electricity grid planning as data-center demand and climate volatility push grids beyond historical data ranges.
Federal regulators are expected to rule by June 2026 on who pays for AI data center grid infrastructure—a decision that will determine whether developers or ratepayers absorb billions in transmission costs.
OpenAI has issued a formal RFP to US manufacturers seeking suppliers for robotics components, data center hardware, and consumer electronics — a major signal of its intent to anchor AI production domestically.
AI is accelerating office vacancy while driving the largest data center construction boom in 30 years — a violent reallocation of capital within commercial real estate that CBRE CEO Bob Sulentic compares to the 1990s outsourcing wave.
SpaceX is pricing its $1.75 trillion IPO tomorrow, and the AI1 orbital data center satellite in the filing is a real hardware program — but analysts say the power and thermal economics of running AI workloads from low Earth orbit remain unproven.
AI data center construction spending has reached parity with general office construction in the U.S., while organized community opposition is intensifying in Louisiana, Arizona, Michigan, and Texas over water, power, and land use.
Nvidia and startup Emerald AI partnered with AES, Constellation, Invenergy, NextEra, Nscale, and Vistra to operate AI data centers as dispatchable grid resources that throttle power consumption in real time — reframing AI infrastructure from grid burden to grid tool.
A Harvard Belfer Center analysis warns AI workloads will consume 9–11% of total U.S. electricity by 2030 and that corporate renewable pledges are structurally inadequate to prevent grid instability.
AI grid management tools are cutting operating reserve costs by up to 15% and dramatically reducing renewable curtailment — as aging infrastructure buckles under AI data center demand and accelerating electrification.
U.S. data center construction hit $50 billion in April 2026 — 28% above last year and the first time on record that AI infrastructure spending outpaced transportation. AI has crossed from sector trend to macroeconomic force.
An Oaklandside investigation documents how AI firm expansion is driving rent increases and displacing small businesses, with AI companies already occupying 17 million square feet across the Bay Area.
A UN University report warns AI data centers consumed 4.5 trillion litres of water in 2025 and could hit 9.3 trillion by 2030 — rivaling global drinking water needs.
AI automation is shrinking office space demand while data center outlays have risen to match general office construction spending for the first time, driving 'Powered Land' near substations to $150,000 per acre.
The European Commission launched AI.grids, a coordinated initiative to develop AI models specifically for grid management and planning, estimating demand-side flexibility could cut EU electricity costs by more than 71 billion euros per year.
China added 430 gigawatts of renewable energy in 2025 alone and is co-locating AI data centers in surplus clean-power zones, creating a structural electricity cost advantage over US AI infrastructure that export controls and chip restrictions cannot address.
SoftBank pledges €75 billion for 5 gigawatts of AI data center capacity in France—Europe's largest AI infrastructure commitment—brokered directly between Masayoshi Son and President Macron.
Grid-scale battery storage is filling the gap between AI data centers' round-the-clock power demands and intermittent renewables — reshaping how hyperscalers structure power agreements.