RALEIGH, N.C., March 13, 2025 /PRNewswire/ -- Coherently, Inc., a new startup backed by NetActuate, today officially introduced Coherently AI, a revolutionary software platform designed to simplify, scale, and monetize AI infrastructure. At the core of Coherently AI is the Global Command Manager (GCM), a multi-tenant solution that empowers telcos, enterprises, and cloud service providers (CSPs) to rapidly deploy AI environments—and even resell GPU capacity—across hosted, self-hosted, multi-tenant, and air-gapped deployments.
Coherently AI addresses the fundamental complexity of AI infrastructure
Coherently Global Command Manager - CPU View
Coherently Global Command Manager - Shell View
By combining best practices from the NVIDIA Reference Architecture with an automated "zero trust" networking approach, Coherently AI delivers "cloud in a box" functionality. Customers can install entire AI datacenters—from networking hardware to server firmware and OS layer—in a fraction of the time typically required, while seamlessly extending white-label AI services to their own customers (referred to as "tenants"). This eliminates the steep learning curve, high cost, and security headaches often associated with AI deployments.
Key Benefits of Coherently AI and GCM
Secure Multi-Tenancy at Scale
Dynamic Isolation: Tenants are granted secure access to bare metal servers or virtual machines in seconds, with strict zero-trust enforcement handled by BlueField-3 DPUs and Spectrum-X Cumulus Linux switches powered by NVIDIA.
Tenant Lifecycle Management: Real-time provisioning and deprovisioning of GPU resources ensures on-demand flexibility, whether you have a handful of tenants or thousands.
White-Label and Resellable GPU Services
White-Label Branding: Operators and enterprises (super-admins) can integrate logos, custom themes, domains, and even custom HTML/CSS/JS.
Revenue Generation: Resell unused GPU services to external customers (tenants) on demand. Coherently's built-in marketplace and brokering tools enable you to capture new revenue streams when AI hardware is underutilized.
API-Driven Architecture
Full Control & Customization: Every function, from provisioning bare metal to configuring GPUs, is available through a secure JSON-based API.
Deep Integration: Operators can tie Coherently into existing billing, authentication, or internal portals, creating a seamless user experience for tenants and super-admins alike.
Automated Infrastructure Management
Day 0 to Day 2 Operations: From networking switches to server firmware, GCM automatically discovers and configures every component in real time.
Reduced Complexity: Pre-validated drivers, container toolkits (such as NVIDIA Container Runtime), and specialized library stacks eliminate the trial-and-error of building AI environments from scratch.
High-Performance AI Workloads
Optimized for NVIDIA RA: Adheres to NVIDIA's reference architecture, supporting up to 8 GPUs per server and low-latency, and RDMA-based connectivity at 400 Gbps via BlueField SuperNICs.
Bare Metal & VM Options: Choose raw hardware access or near-native virtualization performance for generative AI, inference models, or telco-specific applications (e.g., vRAN).
"Coherently AI addresses the fundamental complexity of AI infrastructure," said Drew Hastings, CTO and Co-founder of Coherently. "By blending automated installation, high-performance networking, and robust multi-tenancy, we're giving telcos and enterprises a true 'cloud in a box' solution. With faster time to market, direct ownership of data, and the ability to resell AI factories with white-label branding, our customers leverage new levels of flexibility and speed while maximizing their ROI."
Designed for Telcos, Enterprises, and CSPs
Telcos: Leverage existing datacenter footprints to stand up AI clusters rapidly. Generative AI chatbots, advanced analytics, and edge inference can be delivered as white-labeled services to customers.
Enterprises: Own your data and meet regulatory needs without building AI infrastructure from scratch. Coherently's automated workflows shorten deployment cycles from weeks to days, or even hours.
Cloud Service Providers (CSPs): Expand your service portfolio with branded AI offerings. Coherently's built-in marketplace allows you to provision GPUs and HPC resources to customers on demand.
Built for Modern AI & Networking
Zero Trust Security: BlueField-3 DPUs isolate tenant traffic and automatically enforce per-tenant network isolation with VXLAN, ensuring bulletproof separation.
Network Automation: GCM automatically manages the entire datacenter network topology while maintaining the highest possible scalability and lowest latency by leveraging EVPN and RDMA.
Seamless Integration & Global Reach
Coherently AI supports on-premises, hosted, and completely air-gapped deployments, meeting sovereign data regulations and government security requirements. Through its partnership with NetActuate's 40+ global data center sites and extensive network connectivity, Coherently can accelerate large-scale AI rollouts worldwide, from day-one facility setup to ongoing optimization.
About Coherently, Inc.
Coherently, Inc. is a NetActuate-backed startup devoted to simplifying AI infrastructure for telcos, enterprises, and CSPs. Its flagship platform, Coherently AI, provides an automated, secure, and highly scalable "cloud in a box" solution—enabling customers to deploy AI workloads quickly, own their data, and resell GPU capacity under their own brand.
NetActuate is a leading provider of highly available, low latency custom network and infrastructure services that reach every major global market. From the datacenter to the last mile, we help providers take their products and services to the global edge faster. Our customers can rapidly scale without fear of high costs or devastating performance issues. We built one of the world's largest global networks by number of peers, and it serves as the foundation for our performance BGP anycast platform that powers over 25 billion transactions a day.
SOURCE NetActuate, Inc
WANT YOUR COMPANY'S NEWS FEATURED ON PRNEWSWIRE.COM?
Share this article