📅 - USD.AI has provided a $34 million three-year debt facility to NexGen Cloud to finance NVIDIA B200 GPU deployment in Sweden, turning AI compute into collateral for European infrastructure expansion. The deal gives NexGen non-dilutive capital for its Hyperstack platform while testing whether GPU-backed lending can become a serious funding model for sovereign AI capacity.
The interesting part is not simply that another European AI cloud provider is buying more GPUs. Everyone with access to capital, power, and supply allocation is doing some version of that. The more revealing shift is financial engineering around the machines themselves.
USD.AI, developed by Permian Labs, is treating [...][... Check source for end of article ...]
📅 - JetCool Turns Dell's AI Server Into a Turnkey Liquid-Cooled System - AI servers are becoming easier to buy. Making them operational inside existing data centers remains considerably harder. JetCool is targeting that deployment gap with a liquid-cooled version of Dell's PowerEdge XE7745, combining hardware, cooling infrastructure, installation, and lifecycle support into one package instead of leaving enterprises to integrate those components independently after delivery.
The announcement builds on Dell's earlier introduction of the PowerEdge XE7745 as a GPU-dense platform for AI inferencing, fine-tuning, and HPC workloads. Rather than redesigning the server itself, JetCool has wrapped Dell's hardware in a fully integrated direct-to-chip liquid cooling [...]
📅 - IBM Debuts Sub-1nm Chip Technology For AI Compute - IBM has debuted sub-1 nanometer chip technology built on a 0.7 nm nanostack transistor design, claiming a path to denser, faster and more energy-efficient processors within five years. For businesses, the signal is not immediate procurement. It is a warning that compute roadmaps, AI costs and data center power assumptions may shift again sooner than planned today.
The claim is big. Almost uncomfortably big. A chip node below 1 nm, or 7 angstroms, sounds like the industry has found another hidden drawer in Moore's Law. IBM says the technology can place nearly 100 billion transistors on a chip roughly the size of a fingernail, close to twice the density of the 2 nm chip it showed in [...]
📅 - General Intuition Raises $300M For Gameplay-Trained AI Agents - General Intuition is seeking roughly $300 million at a valuation exceeding $2 billion, turning a rejected acquisition offer for gaming data into one of AI's most closely watched infrastructure plays. Rather than selling the data itself, the startup wants to build AI agents trained on billions of first-person gameplay clips collected through Medal, its affiliated gaming platform instead.
The fundraising says as much about AI economics as it does about technical ambition. Eighteen months ago, reports suggested OpenAI was willing to spend around $500 million to acquire Medal, whose business revolves around gamers uploading short clips from their gameplay sessions. The offer never became a [...]
📅 - The Strategic Blueprint: Why London Reigns as Europe's Premier Data Center Location - For global enterprises, high-frequency trading firms, and IT professionals looking to dominate both the European and North American digital landscapes, choosing the right geographical deployment zone is the single most critical infrastructure decision you can make. At iDatam, we continuously analyze global network topologies, and our latest findings reinforce a well-known industry truth: London is not just an emerging tech hub; it is a fully matured, hyper-optimized digital command center. Deploying dedicated servers in London provides unparalleled strategic advantages that simply cannot be replicated by inland European facilities.
The primary architectural benefit of a London data center l
📅 - Exploring the Best Data Center Locations: The Madrid Advantage for Europe and LatAm - For global enterprises, SaaS providers, and IT professionals looking to dominate both the European and Latin American digital landscapes, choosing the right geographical deployment zone is the single most critical infrastructure decision you can make. At iDatam, we continuously analyze global network topologies, and our latest findings reinforce a massive shift in the industry: Madrid has transformed from a secondary regional market into the undisputed digital and economic engine of Southern Europe. Deploying dedicated servers in Madrid provides unparalleled strategic advantages that simply cannot be replicated by traditional northern "FLAP" (Frankfurt, London, Amsterdam, Paris) data centers
📅 - Microsoft Says Datacenters Use Less Water as AI Load Grows - Microsoft says it has cut the water intensity of its datacenters nearly 90% since the early 2000s, a claim arriving as AI infrastructure strains power grids, local politics and resource planning. The company reported 2025 average water usage effectiveness of 0.27 liters per kilowatt-hour, down from 2.3, while capacity demand keeps rising for cloud and AI services.
For buyers of cloud infrastructure, the number is not cosmetic. Water is becoming a site-selection variable, a permitting constraint, a community relations problem and, increasingly, a line item in risk models. Microsoft is trying to show that datacenter growth can be separated from water consumption. Not eliminated. [...]