Category
Design & Build
-
Six Months, Not Fourteen: Schneider Electric’s Power Modules Tackle AI Bottlenecks
Schneider Electric launches standardised AI power Modules to cut deployment time to 6 months. Schneider Electric recently announced a new range of standardised Prefabricated Power Modules and Power Skids, aimed at hyperscale, neocloud and colocation data center operators building out AI infrastructure. The modules can deliver up to 2.5 megawatts of capacity each, which Schneider…
-
Liquid Cooling Moves From Component to Data Center Architecture
LiquidStack is shifting the liquid-cooling conversation from individual components to whole-system design as AI rack densities rise and operators prepare for successive generations of high-density compute
-
Civo fires starting gun on 40-site UK AI infrastructure push
When open in 2027, the 8MW Hertfordshire facility will herald the rollout of a national distributed edge network.
-
Oracle Invokes Force Majeure on $165B Blue Owl Jupiter Data Center project
New Mexico AI Campus Hits Regulatory Wall Oracle Corp has sent a force majeure notice to Blue Owl Capital, the developer behind Project Jupiter, it’s 2.45-gigawatt data center campus under construction in New Mexico, according to Bloomberg, which cited people familiar with the matter. The move is designed to let Oracle delay lease payments, not…
-
Why Vertiv is Consolidating the AI Factory
The scaling bottleneck for artificial intelligence has officially left the chip lab and entered the physical world. Vertiv’s rapid string of acquisitions and high-capacity NVIDIA certifications reveal an industry-wide truth: the future of hyperscale compute belongs to those who control the physical mechanics of power and fluid
-
ABB Bets on Direct Current to Break the AI Power Wall
The AI power race is shifting from generation to the final hundred metres of the data centre. ABB’s new Infinitus direct current architecture aims to rewrite the white space by bypassing legacy alternating current infrastructure entirely to reclaim crucial compute capacity.
-
GridLock: Why Hyperscalers Are Buying Their Own Substations
While tech giants pour billions into advanced AI chips, their multi-gigawatt growth roadmaps are stalling at the local power grid. To survive the infrastructure crunch, cloud operators are transforming into heavy industrial procurement giants buying up transformer supply lines and building their own utility-scale substations.
-
AI Data Center Debt Just Got More Expensive, Meta’s Own Bond Sale Proves It
Bloomberg reports AI data center debt yields are climbing across the market, with Meta’s $12.5 billion El Paso bond offering pricing at a materially higher risk premium than a nearly identical deal from a year earlier
-
Nordic AI Campus Moves From Paper to Steel
3E Network’s Finnish subsidiary, Aurora Core Technology, entered active construction on a green-powered AI data center project this week — a milestone increasingly hard to reach given how many announced campuses stall at the permitting stage
-
Cornelis Networks to Build Fabric that Keeps GPUs From Sitting Idle
Cornelis Networks raised $205 million and unveiled Active Compute Fabric, a scale-up interconnect built on open UALink and ESUN standards a direct challenge to Nvidia’s grip on GPU-to-GPU networking