MetalSoft for AI FactoriesMetalSoft for AI Factories

MetalSoft is an intelligent orchestration platform specifically built for on-premises. It orchestrates bare metal servers, network, storage and more for on-premises environments, focusing on banking, telco, and government as well as service providers.

MetalSoft has unmatched hardware orchestration capabilities and as such it provides safe mechanisms for the AI to optimize and manage infrastructure for specific workloads, diagnose failures and replace failed nodes.

Yes. Through our Hardware Abstraction Layer, MetalSoft's AI treats your entire on-premises fleet as a single pool of resources. It understands the unique performance characteristics of different vendors and optimizes the placement of workloads—Bare Metal or VM—to ensure maximum efficiency and performance.

No, MetalSoft manages the hardware directly and can deploy Bare Metal, VMs and containers in a very lightweight and efficient manner, optimizing for NUMA node placement and many other factors, reducing the complexity and increasing performance.

On the contrary, it strengthens it. AI-assisted operations ensure that security configurations do not "drift" over time. If a manual change violates a compliance boundary (like DORA's resilience requirements), the AI-enabled monitoring triggers an immediate isolation or remediation protocol.

MetalSoft deploys SLURM directly on the bare metal, without any kubernetes or virtualization layer. This makes it familiar to everybody that used SLURM before, eliminates complexity and provides absolute performance.

MetalSoft can help clients deploy VLLM or other inference runtimes on bare metal, VMs or containers depending on the tenancy model, maximizing performance and simplifying the setup.

No, we believe customers can run their own Jupyter Notebooks and manage their dependencies and there are many fantastic tools out there that can help with this.

MetalSoft can help deploy qdrant or other clusters on demand via our extension mechanism.

MetalSoft leverages AI for two primary functions: Monitoring and Autonomous Optimization. By analyzing telemetry data from switches, firewalls, and servers simultaneously, the AI performs real-time Root Cause Analysis (RCA), distinguishing between transient noise and critical system failures. It also assists in "Intent-Based" operations, where the system automatically optimizes hardware settings for specific workload requirements.