Amalgamy: Unified AI and HPC Operating System

Turn fragmented AI and HPC infrastructure into a governed capability

Amalgamy delivers the right workload to the right compute at the right time. Run mixed AI and HPC workloads with security, efficiency, and control across your existing environments such as Kubernetes, Slurm, Run:ai and more.

The Challenge

Sovereign AI capability is hard

AI and HPC compute is scarce, expensive, and hard to build and operate at scale. Key operational challenges include:

Utilization Illusion

Clusters appear to be fully utilized when GPUs are allocated. But much of that GPU capacity is idle as jobs wait on data movement, approvals, or facility constraints.

Compliance is an afterthought

Security and compliance usually step in only at the end, as an approval gate. That makes compliance feel like a roadblock instead of part of how work gets done.

Infrastructure sharing is not trusted

Institutions can't share infrastructure with partners without losing control, trusting blindly, or defaulting to hyperscaler models.

The Control Layer

From AI Infrastructure to Operational Capability

Amalgamy is the only AI/HPC operating system built to operationalize sovereign, multi-institution infrastructure. It brings intelligent workload placement, automated governance, and existing environments together under one control layer.

Sovereign AI Blueprint

Amalgamy unifies AI and HPC operations across the sovereign AI stack.

Plan • Build • Activate

Workloads

World Models • Model Training • Inference • Agentic Workflows • HPC Jobs

Amalgamy

Unified AI + HPC Operating System

Orchestration • Utilization Intelligence • Access Control • Resource Optimization

Existing Environments

Slurm • Kubernetes • Run:ai • Linux • Cloud and On-Prem Systems

Infrastructure

Facilities • Power • Cooling • GPU/CPU • Storage • Networking

Policy • Governance • Workforce Enablement

Download the Amalgamy White Paper

Get the full architecture, deployment models, and the economic framework behind Amalgamy, including how institutions turn scarce compute into governed, measurable capability.

Download Amalgamy White Paper

Capabilities

Key Amalgamy Capabilities

Intelligent Workload Scheduling

Routes every job to the right compute, automatically, so utilization improves.

Amalgamy reduces the Utilization Illusion by preventing data-starved jobs, prioritizing the newest GPUs for parallel work, and backfilling older clusters with suitable workloads.

The outcome is higher utilization from existing hardware and lower pressure to buy more capacity.

Policy-Driven Security and Compliance

Enforces policy at the workload level, so approved work moves faster.

Amalgamy builds security and compliance directly into how jobs are scheduled and isolated, removing the need for late-stage approval steps.

The outcome is faster secure deployment and lower compliance risk across sensitive environments.

Integration with Existing Infrastructure

Works with your environments, so workflows are unchanged.

Amalgamy coordinates across Slurm, Kubernetes, Run:ai, and existing hardware, so researchers submit jobs the same way while operators gain a unified control layer.

The outcome is faster time-to-value and no rip-and-replace penalty.

Utilization and Operating Intelligence

Turns workload activity and utilization into telemetry, so that decisions are data-based.

Amalgamy reports on the difference between allocated time and productive time. This helps with capacity planning, grants, chargebacks, and shared governance reporting.

The outcome is data-based decisions, with metrics like Grant-Dollar Efficiency and Yield per Watt.

The Dynamic Tenant Perimeter

Amalgamy runs each workload within a tenant whose perimeter expands and contracts dynamically in real time across shared infrastructure. As a result, each tenant gets the maximum capacity the cluster can provide.

When priorities shift, Amalgamy handles it automatically. Capacity is reallocated, lower-priority jobs are evicted or scaled back, and policy is enforced in real time, without anyone having to make that call manually.

It's zero trust, but built for the entire cluster, not just for compute and networks.

Use Cases

Mixed Workloads, One Governed Environment

Run AI, HPC, inference, education, and partner workloads within one governed operating model. Use cases include:

Research acceleration

Help faculty, labs, and research teams run AI and HPC workloads with less operational friction and more predictable access to the resources their work requires.

AI production and inference

Support low-latency and production-oriented AI workloads alongside traditional HPC and research workloads.

Education and workforce development

Give students, operators, and technical teams access to real infrastructure patterns so training becomes a path to durable workforce capability.

Industry collaboration

Create governed environments where commercial partners can run approved workloads while the institution preserves control over infrastructure, data policy, and administrative boundaries.

Personas

Who Amalgamy is built for

AI devs and researchers

Researchers spend more time on discovery, experimentation, and publication, and less time navigating storage permissions, data movement, and infrastructure handoffs.

How Amalgamy helps

Amalgamy runs approved AI and HPC work through familiar workflows while coordinating data readiness, policy requirements, and compute placement behind the scenes.

HPC and AI operators

Operators support more users and more workload types with better visibility and coordination across the environment without changing existing tools and workflows.

How Amalgamy helps

Amalgamy adds an operating layer that works across your existing storage, networking, and schedulers creating a single control and visibility layer.

CIOs, CTOs, and Infra leaders

Leaders can demonstrate AI/HPC investments turn into measurable institutional capability and better ROI across research, production, education, and partnerships.

How Amalgamy helps

Amalgamy unifies compute, data, policy, and utilization intelligence so leaders can see what infrastructure is actually producing, where capacity is being wasted, and how workload types can expand over time.

Security, compliance, and governance teams

Governance teams enable approved users and partners to move faster without weakening institutional controls or losing visibility into sensitive work.

How Amalgamy helps

Amalgamy applies policy-driven security, workload-level isolation, and audit telemetry across mixed AI/HPC environments.

Public-sector, consortium, and partnership leaders

Partners can securely share AI infrastructure across institutions, agencies, and commercial partners while preserving local control.

How Amalgamy helps

Amalgamy enables governed collaboration models where institutions can share capacity, support partners, and preserve administrative boundaries.

Students and workforce development leaders

Educators can build durable local AI infrastructure capability by training people on realistic systems and workflows.

How Amalgamy helps

Amalgamy makes production-grade AI/HPC environments more accessible for education, workforce development, and hands-on operational learning.

Collaboration

Federated Infrastructure

Amalgamy enables governed, auditable access across federated environments through temporary, policy-controlled access mechanisms.

Core Deployment Models

Public-Private Compute Partnerships

Securely bridge commercial networks and state-level infrastructure.

Regional AI Consortiums

Safely pool and scale available compute power across allied networks.

Academic Collaboration Networks

Build high-utilization research networks across universities without sacrificing data ownership.

Shared Sovereign AI Initiatives

Coordinate multi-site resources while enforcing national and institutional data governance policies.

Economics

Summary of Amalgamy economic benefits

Increase Revenue

Better Yield per Watt: generate more productive work from the same physical infrastructure.
Faster time to production: move from bare metal to governed workloads faster.
Higher Grant-Dollar Efficiency: turn research infrastructure into more usable research output per dollar.

Reduce Costs

Lower idle compute cost: reduce time lost to data-starved or mis-placed workloads.
Avoid rip-and-replace cost: add intelligence to the environments already in use.
Reduce compliance and deployment drag: shorten the gap between installed infrastructure and secure production use.

Improve Productivity

Keep teams focused on workloads: reduce manual coordination across data, policy, and placement.
Preserve familiar tools and workflows: improve throughput without retraining everyone onto a new stack.
Embed policy into operations: make governance part of how work runs, not a late-stage gate.

Contact us

Whether you are deploying a new AI cluster or federating existing infrastructure into a regional compute network, Amalgamy provides the orchestration, governance, and operational layer required to scale securely.

Design, deploy, and operationalize AI infrastructure with Amalgamy.