45k user reviews & 1 million votes, over 29k web hosting brands & 85k plans since 2004!
Menu
WHTop → 📋 News → MCP Servers Expose New Infrastructure Risks Across Enterprise AI
MCP Servers Expose New Infrastructure Risks Across Enterprise AI
📅 - Enterprise adoption of Model Context Protocol is opening a new infrastructure security gap as AI agents connect to servers beyond conventional cloud governance, including foreign networks, consumer connections and abandoned domains. Research commissioned by OX Security, covering 15,465 published MCP servers found weak controls around hosting location, domain ownership and persistent permissions, complicating data residency and supply chain oversight for organizations.
The issue is becoming harder to dismiss as Model Context Protocol moves deeper into developer tooling and enterprise automation. Created by Anthropic in late 2024, MCP gives agents a standardized way to connect with [...][... Check source for end of article ...]
📅 - OVHcloud Questions Business Case for Europe’s AI Gigafactories - OVHcloud is questioning whether Europe's emerging AI Gigafactory model can support the economics required to justify massive GPU investment, arguing that current public commitments leave operators carrying much of the commercial risk. The concerns were laid out in a newly published blog article by OVHcloud founder, Chairman and CEO Octave Klaba, who challenges several assumptions underpinning the proposed infrastructure model.
Klaba argues that single-site designs, nationally bounded demand, and the rapid depreciation of GPU hardware could leave multibillion-euro infrastructure projects struggling to maintain sufficient utilization. His argument is not that Europe lacks a need for [...]
📅 - GTZHost Releases Comprehensive Engineering Guide to Resolving "CUDA Out of Memory" Errors in AI Training - As enterprise investment in Generative Artificial Intelligence and Large Language Models (LLMs) continues its unprecedented acceleration, data science teams are increasingly running into severe infrastructure bottlenecks. The most ubiquitous and frustrating hurdle encountered by machine learning engineers during model training and fine-tuning is the notorious RuntimeError: CUDA out of memory. To assist developers in optimizing their deep learning pipelines and maximizing their infrastructure efficiency, GTZHost has published an extensive technical tutorial detailing how to systematically diagnose, manage, and resolve Video RAM (VRAM) saturation on GPU servers.
📅 - Liquid Cooling vs. Air Cooling for Next-Gen GPU Servers - Next-generation artificial intelligence workloads have fundamentally broken traditional data center physics. As NVIDIA's Blackwell architecture and equivalent high-performance silicon enter enterprise infrastructure, hosting providers are confronting a strict thermal wall. Managing modern GPU clusters is no longer just a software engineering problem—it is a direct exercise in thermodynamics.
The Physics of Extreme TDP Escalation Thermal Design Power (TDP) for enterprise GPUs has escalated rapidly. Previous generations hovered around 300W, while Hopper (H100) reached 700W. NVIDIA Blackwell (B200) shatters limits past 1,000W to 1,200W per chip. Packing eight high-TDP GPUs into a single server
📅 - Oracle Invokes Force Majeure on New Mexico AI Data Center - Oracle has invoked force majeure protections tied to Project Jupiter, the $165 billion New Mexico data center underpinning its AI infrastructure expansion with OpenAI and SoftBank, according to Bloomberg. The move is intended to limit payment exposure if the 2.45 gigawatt facility misses its planned 2028 opening, adding contractual pressure to a strained financing and permitting picture.
The notice does not mean Oracle is walking away from Project Jupiter as its main tenant. Rather, Oracle is seeking protection from payment obligations if delays outside its control prevent the facility from becoming operational on schedule. That shifts attention from whether the campus will be built to [...]
📅 - IonQ Adds Quantum Acceleration to Industrial HPC Workflows - IonQ and Synopsys have tested a hybrid quantum computing method that cut runtime in engineering simulations by as much as 14.6%, placing quantum processing inside an established computer aided engineering workflow rather than beside it. The result gives cloud and HPC operators a concrete view of how quantum systems could function as specialized accelerators for classical workloads.
The interesting part is where the quantum system sits. It does not replace the classical cluster running a crash simulation, fluid model or mechanical analysis. Instead, quantum processing is used before the computationally expensive solving stage to reorganize large systems of equations so the classical [...]
📅 - DigiCert Targets Post-Quantum Migration Across Cloud Infrastructure - DigiCert has made Quantum Central generally available, giving enterprises a management layer for discovering cryptographic assets, setting post-quantum policies and coordinating remediation across cloud and on-premises environments. The launch moves post-quantum preparation closer to routine infrastructure operations, where certificate inventories, key management, ownership and change control may prove harder than selecting new cryptographic algorithms themselves at scale.
The problem is increasingly operational rather than theoretical. Enterprises can decide they need post-quantum cryptography years before they know where vulnerable encryption actually sits across applications, [...]
📅 - 
📅 - 
📅 - 
📅 -