45k user reviews & 1 million votes, over 29k web hosting brands & 85k plans since 2004!
Menu
WHTop → 📋 News → Microsoft Deploys 3M Expanded Beam Optics in Azure Networks
Microsoft Deploys 3M Expanded Beam Optics in Azure Networks
📅 - Microsoft will deploy 3M's Expanded Beam Optical technology in Azure data centers, making Microsoft the first publicly announced hyperscale cloud provider to deploy the connector system. The partnership also extends into enterprise software, with 3M planning to use Microsoft's AI and digital tools across its internal customer service, finance, sales, marketing, and order management operations over time.
The infrastructure element carries more weight for data center operators. Optical connectors are minor components by cost, yet they sit inside deployment workflows that become harder to control as AI clusters consume more fiber, more ports, and more technician hours. 3M's EBO design uses [...][... Check source for end of article ...]
📅 - Macquarie Secures Sydney Site for a 200MW Data Center Campus - Macquarie Technology Group has exercised an option to acquire a Macquarie Park site measuring 34,200 square meters for A$240 million, advancing plans for a technology campus alongside a 200 megawatt data center development. The purchase, funded with cash and corporate debt, still faces settlement, planning, power, construction, customer and financing hurdles before operations targeted for late 2029.
The option was exercised after the site's subdivision process was completed, moving the transaction toward standard settlement over the next few weeks. For infrastructure buyers, that is the first irreversible part of a project that remains highly conditional. Macquarie now controls a large [...]
📅 - Stop Wasting GPU Cycles: Multi-Model Serving with SGLang - As artificial intelligence applications scale, the cost and efficiency of deploying Large Language Models (LLMs) become paramount. A common bottleneck for AI developers and enterprise IT teams is highly inefficient GPU utilization. VRAM fragmentation and sluggish inference times can severely limit your hardware’s potential, often forcing you to purchase additional, expensive GPU resources unnecessarily.
iDatam’s latest technical tutorial tackles this exact issue, introducing a highly effective solution: deploying SGLang on a bare-metal GPU server.
Why SGLang is a Game-Changer for AI Inference SGLang is designed specifically to optimize how multiple models interact with your server's VRAM.
📅 - Tech Sovereignty Gap Leaves the US and China Firmly in Front - Forrester expects global technology sovereignty to barely improve through 2030, despite governments pouring money into semiconductors, AI and domestic infrastructure. Its new 14-country index projects the average score rising from 39% in 2025 to just 40%, leaving the United States and China far ahead while Europe and other economies remain dependent on foreign technology suppliers and platforms.
The forecast puts numbers against a problem governments have spent several years trying to solve with subsidies, procurement rules and increasingly explicit industrial policy. Technology independence is expensive. Worse, the underlying supply chains are interconnected enough that replacing one [...]
📅 - Dimension AI Wins $71.4M GPU Capacity Contract From Exascale - Exascale Labs has signed a three-year agreement to buy roughly $71.4 million of dedicated GPU compute capacity from Singapore-based Dimension AI, expanding the supply behind its GPU-as-a-Service platform. The contract gives Exascale longer-term access to scarce infrastructure, but leaves undisclosed the GPU models, delivery schedule, utilization commitments, customer demand, and economics supporting the purchase over its term.
The immediate change is access, not ownership. Exascale describes itself as an asset-light operator that sources GPU capacity from third-party data centers and resells reserved or on-demand compute through a software-defined platform. The Dimension AI contract [...]
📅 - Crusoe and Lancium Scale Texas AI Infrastructure Toward 1GW+ - Crusoe and Lancium plan to build a 1 gigawatt AI data center campus in Childress, Texas, extending a partnership already underway in Abilene as hyperscale computing demand accelerates. Construction is scheduled to begin during the third quarter of 2026, with the project targeting a major cloud customer and substantial regional infrastructure investment alongside long-term employment opportunities.
Texas continues attracting another wave of AI infrastructure, although the scale keeps growing faster than the surrounding debate. A one gigawatt campus would rank among the largest dedicated AI computing developments announced in the United States, reflecting how hyperscale operators [...]
📅 - U.S. Firms Move Microsoft AI From Pilots to Operating Models - U.S. companies are moving beyond isolated artificial intelligence deployments and restructuring Microsoft environments around integrated AI operations, according to new research from Information Services Group. The shift reflects mounting pressure to justify technology spending as enterprises demand measurable returns, tighter financial oversight, and operational models capable of scaling AI without multiplying infrastructure complexity or long-term costs.
The latest assessment suggests the conversation around enterprise AI has become noticeably less experimental. Boards still want automation. Business units still want generative AI. Finance teams, however, increasingly dictate the pace. [...]
📅 - 
📅 - 
📅 -