In this exclusive VMblog interview, Matthew Shaxted, CEO of Parallel Works, discusses how the company's ACTIVATE platform is helping enterprises navigate the growing trend of cloud repatriation while optimizing performance and costs. Spun out of Argonne National Laboratory in 2015, Parallel Works has evolved from pioneering hybrid multi-cloud orchestration to addressing the complex challenges of enterprise AI workloads. As organizations face rising costs of public cloud services and evolving licensing models, Shaxted explains how ACTIVATE's new OpenStack virtualization integrations provide a unified control plane for distributed computing resources—whether on-premise, in the cloud, or in hybrid environments. The platform's ability to bridge traditional virtualization with emerging cloud-native technologies offers enterprises a path to migrate workloads on-premises while maintaining the flexibility and scalability they need for compute-intensive applications and AI models.

VMblog:  Can you provide an overview of Parallel Works and the types of challenges your technology helps customers solve?

Matthew Shaxted:  Parallel Works was spun out of Argonne National Laboratory in 2015 to commercialize supercomputing workflow technology. Early on, we pioneered hybrid multi-cloud orchestration capabilities before the enterprise sector fully grasped how important this type of software tooling would be for enterprise AI workloads. By addressing complex orchestration needs for enterprise research, Parallel Works anticipated similar demands emerging in enterprise AI and quantum computing.

At the core of our innovation is Parallel Works ACTIVATE, a unified control plane for distributed computing resources. ACTIVATE empowers  teams that depend on compute-intensive applications and AI models by providing a seamless way to provision, manage, and share computing resources - on-premise, in the cloud, or in hybrid environments. The technology is built to scale, ensuring organizations can efficiently handle the demands of simulation, analytics, and AI-driven workloads.

For example, if you have used the weather app on your phone you have used the output of the large-scale computing that ACTIVATE enables. We provide the last mile end user experience to access what are often diverse and complex computing environments within and external to an organization. Many organizations we work with are straddling a variety of on-prem, batch scheduler, virtualized, hyperscale and accelerator clusters, and we help form the bridge to increase utilization and accessibility of the systems.

VMblog:  What industry changes drove Parallel Works to enhance ACTIVATE with new OpenStack virtualization integrations, and how does this address current market challenges in cloud resource management?

Shaxted:  The rising cost of public cloud services and evolving licensing models for proprietary virtualization solutions have led enterprises to reevaluate their cloud strategies. Many companies are turning to private cloud and hybrid cloud environments to reduce total cost of ownership (TCO) and gain greater control over their resources. Parallel Works enhanced ACTIVATE with OpenStack support to provide a cost-effective alternative that enables businesses to migrate virtualized workloads on-premises while maintaining scalability, security, and continued operational agility to hyperscale cloud providers The integration aligns with the growing trend of cloud repatriation, helping companies optimize cloud spending and resource management.

VMblog:  With the rising trend of enterprises repatriating workloads from public cloud providers to private or on-premises data centers, how does ACTIVATE help businesses navigate this shift while optimizing performance and costs?

Shaxted:  ACTIVATE streamlines the transition to private cloud environments by integrating OpenStack support, allowing enterprises to move workloads off expensive public clouds while maintaining flexibility. The platform provides turnkey cluster provisioning, dynamic cost control, and seamless storage management, making it easier for organizations to deploy and manage HPC and AI workloads in-house. Additionally, ACTIVATE's real-time usage tracking and automated budgeting alerts ensure cost transparency, helping businesses avoid over-provisioning and unnecessary spending while maintaining high performance.

We essentially make our customer's OpenStack environments look and feel like AWS, Google, Azure or other hyperscale cloud providers, by allowing org admins to set costs to their Virtual Machine shapes, and track these costs in an operational model largely identical to the public clouds.

VMblog:  How does ACTIVATE bridge the gap between traditional virtualization and emerging cloud-native technologies like Kubernetes?

Shaxted:  While many agree containerization and microservices are the future of cloud-native computing, many organizations still rely on traditional virtual machines (VMs) and are not yet ready for a complete shift to container-based schedulers such as Kubernetes. ACTIVATE bridges this gap by supporting both virtualized and containerized environments, enabling businesses to gradually adopt Kubernetes without disrupting existing workflows. By integrating OpenStack and Kubernetes under a unified control plane, ACTIVATE allows enterprises to run hybrid environments, transitioning at their own pace while benefiting from automation, scalability, and cost efficiencies.

VMblog:  One of ACTIVATE's key features is its dynamic cost control and usage tracking. Can you elaborate on how this functionality helps organizations maintain financial oversight and optimize their cloud spending?

Shaxted:  ACTIVATE provides granular financial oversight through real-time cost tracking, automated alerts, and billable VM assignments to OpenStack resources. These features allow enterprises to set custom budgets, monitor resource consumption in near-real time, and receive notifications when usage exceeds predefined limits. The platform's predictive analytics further help IT teams make data-driven decisions, ensuring that workloads are efficiently distributed across public and private cloud environments to minimize unnecessary spending while maintaining optimal performance.

VMblog:  How has the shift from HPC to AI and machine learning impacted the services and solutions you provide for your clients?

Shaxted:  As enterprises increasingly prioritize AI and machine learning workloads, Parallel Works has evolved ACTIVATE to support GPU-intensive tasks, dynamic resource scaling, and advanced workload orchestration. The platform now features fractionalized GPU resource allocation, enabling businesses to efficiently manage AI workloads while avoiding underutilization. Additionally, ACTIVATE's one-click launch capabilities for AI tools like Jupyter notebooks and RStudio provide seamless access to computational resources, accelerating innovation in AI-driven research and operations.

VMblog:  What future developments can we expect from Parallel Works regarding AI and HPC resource management, and how do you see OpenStack evolving within enterprise IT strategies?

Shaxted:  Parallel Works is focused on expanding ACTIVATE's automation and AI-driven resource management capabilities, improving performance optimization for hybrid cloud and AI-driven workloads. Future developments will likely include deeper Kubernetes integrations, AI-powered predictive scaling, and enhanced multi-cloud orchestration tools to help enterprises seamlessly manage workloads across diverse environments.

##