Liqid Inc

Liqid Inc is an enterprise infrastructure company that develops composable disaggregated infrastructure (CDI) platforms for dynamically allocating and managing data center resources such as Graphics Processing Unit (GPU), Non-volatile Memory Express (NVME) storage, Field Programmable Gate Array (FPGA), and networking.

• Composable disaggregated infrastructure platform for data center resource pooling and orchestration (infrastructure virtualization)
• Software-defined resource management for GPUs, NVME storage, networking, and accelerator devices (infrastructure automation)
• Support for on-premises (on-prem) and edge environments with rack-scale composability (data center infrastructure)
• Integration with existing x86 server ecosystems and PCIe-based fabrics (infrastructure integration)
• Tools for policy-based configuration, monitoring, and lifecycle management of composed systems (IT operations management)

More About Liqid Inc

Liqid Inc focuses on composable disaggregated infrastructure (CDI), enabling enterprises and institutions to pool compute, GPU, storage, and networking resources and assign them dynamically to workloads. Its platform is used in data centers that run Artificial Intelligence (AI), High performance computing (HPC), analytics, and other accelerator-intensive applications, where fixed server configurations can lead to underutilization of expensive components. By abstracting physical resources into a shared pool, Liqid allows administrators to compose and recompose servers in software according to workload requirements.

The company’s core technology operates at the infrastructure virtualization and automation layer, using PCIe-based fabrics to disaggregate GPUs, NVME storage, FPGAs, and other devices from host servers. Through a software control plane, IT teams can allocate these resources to bare-metal servers without physical recabling or downtime. This approach aligns with software-defined data center concepts, where policy-based management and programmatic control replace static hardware provisioning practices.

Liqid’s offerings are positioned for organizations that maintain on-prem or edge environments but want some of the elasticity and resource efficiency associated with cloud architectures. In contrast to traditional server-centric models, where each node contains fixed ratios of Central Processing Unit (CPU), memory, storage, and accelerators, Liqid’s composable model allows independent scaling of each resource domain. For example, GPU pools can be expanded or rebalanced across servers running AI training workloads, while NVME pools can be allocated to data-intensive analytics jobs.

The platform typically integrates with standard x86 servers and uses PCI Express (PCIe) fabrics as the interconnect layer, enabling attachment of external GPU and NVME enclosures to multiple hosts. Software APIs and management interfaces allow integration with existing data center workflows and tools, supporting operations teams that manage mixed environments with virtualization, containers, and bare-metal workloads. Policy controls and telemetry help ensure that performance-sensitive applications receive the required accelerator, storage, and network capacity.

Within an enterprise IT directory or marketplace taxonomy, Liqid Inc fits primarily under composable infrastructure, data center resource orchestration, and AI/HPC infrastructure optimization. Its technology is relevant in environments where GPU and accelerator utilization, NVME performance, and flexible resource assignment are priorities, including AI/ML clusters, research computing environments, and large analytics platforms. By enabling resource pooling and software-defined composition, Liqid provides an alternative to static server designs and contributes to more granular capacity planning and infrastructure lifecycle management.