Menu

Liqid: The Right Place, Right Time for Composable AI Stack

Liqid: The Right Place, Right Time for Composable AI Stack

The News: Liqid brings in tech industry veteran Edgar Masri as CEO, focusing the company’s efforts on flexible and efficient GPU infrastructure for AI applications. Sumit Puri will continue to contribute to Liqid’s success as a member of the Board of Directors and the roles of President and Chief Strategy Officer (CSO). Read the press release here.

Liqid: The Right Place, Right Time for Composable AI Stack

Analyst Take: We have followed Liqid since its early days as a PCIe switch. The company evolved into a composable stack understanding the needs of the HPC market to maximize the use of CPU, GPU, memory, storage, and all the elements needed to drive large computational computing. Today, the company finds itself in a sweet spot where the need and requirements for this type of technology are beyond the national labs and major research centers and becoming part of large enterprises. As has been quoted many times, luck follows preparedness, and Liqid is in this space.

What is Liqid? In shorthand, it is a PCIe switch-based architecture with direct-connects resources to servers that can compose multiple GPUs, FPGAs, and flash modules to single compute node. The company ships this using industry-standard storage, servers, and GPUs, making this solution more flexible than proprietary modular chassis.

The Liqid Matrix software composes granular pools of resources into bare-metal systems within a few seconds. These systems are custom configured for the workloads they will support. Matrix, which runs on a low-cost server or embedded into a switch connected to the fabric, orchestrates this process at the device level through a GUI, CLI, or a set of RESTful APIs. In addition, this software manages scaling of resource pools, multitenancy, remote configuration, migration, etc. The Matrix software extends the capabilities, composing across multiple fabric types, including PCIe Gen 3, PCIe Gen 4, Ethernet, and InfiniBand (IB). Most recently, Liqid announced and is shipping the UltraStack, quoted as “up to 20-way systems” using NVIDIA L40S GPU and Dell Technology PowerEdge servers.

Why does this matter? Liqid’s systems can dynamically reconfigure to meet the demands of jobs in real time. When the resources (i.e., GPUs) are no longer needed, Liqid can move them back into the pool and make them available for other workloads. There is also elimination of overprovisioning and less manual intervention to make changes in the environment to meet the needs of the data scientists.

In a timely fashion, Liqid brings a considerable powerhouse of a system into the traditional data center, with configurability that to a degree is like dynamic cloud operations.

Disclosure: The Futurum Group is a research and advisory firm that engages or has engaged in research, analysis, and advisory services with many technology companies, including those mentioned in this article. The author does not hold any equity positions with any company mentioned in this article.

Analysis and opinions expressed herein are specific to the analyst individually and data and other information that might have been provided for validation, not those of The Futurum Group as a whole.

Other Insights from The Futurum Group:

Dell AI PoC for Transportation & Logistics

Intel Gaudi Performance – Beats NVIDIA?

Quantum in Context: A Qubit Primer

Author Information

Camberley Bates

Now retired, Camberley brought over 25 years of executive experience leading sales and marketing teams at Fortune 500 firms. Before joining The Futurum Group, she led the Evaluator Group, an information technology analyst firm as Managing Director.

Her career spanned all elements of sales and marketing including a 360-degree view of addressing challenges and delivering solutions was achieved from crossing the boundary of sales and channel engagement with large enterprise vendors and her own 100-person IT services firm.

Camberley provided Global 250 startups with go-to-market strategies, creating a new market category “MAID” as Vice President of Marketing at COPAN and led a worldwide marketing team including channels as a VP at VERITAS. At GE Access, a $2B distribution company, she served as VP of a new division and succeeded in growing the company from $14 to $500 million and built a successful 100-person IT services firm. Camberley began her career at IBM in sales and management.

She holds a Bachelor of Science in International Business from California State University – Long Beach and executive certificates from Wellesley and Wharton School of Business.

Related Insights
Can Prisma SASE Actually Secure Agents It Cannot See?
March 29, 2026

Can Prisma SASE Actually Secure Agents It Cannot See?

Palo Alto Networks extended Prisma SASE to govern agentic AI workloads, but structural mismatches between SASE design and dynamic agent behavior raise critical enforcement questions....
Prisma AIRS 3.0: Does Palo Alto Own the Agentic AI Security Stack?
March 29, 2026

Prisma AIRS 3.0: Does Palo Alto Own the Agentic AI Security Stack?

Palo Alto Networks unveiled Prisma AIRS 3.0, a purpose-built security platform for autonomous AI agents. As enterprises deploy agentic systems across cloud and SaaS, control of the agentic security stack...
Will Supermicro's Legal Crisis Shift Server Market Share to New Dell and HPE GPU Platforms?
March 27, 2026

Will Supermicro’s Legal Crisis Shift Server Market Share to New Dell and HPE GPU Platforms?

Brendan Burke, Research Director at Futurum, shares insights on how Supermicro's export crisis creates a GPU allocation opening for Dell and HPE, reshaping the AI server competitive landscape post-NVIDIA GTC...
Infosys Bets on P&C Insurance Depth With Stratus Acquisition
March 27, 2026

Infosys Bets on P&C Insurance Depth With Stratus Acquisition

Infosys agreed to acquire Stratus, a New Jersey tech firm specializing in property and casualty insurance, signaling a strategic shift toward domain-specific AI in regulated industries....
Can Palo Alto's Agentic NetOps Actually Kill the NOC Ticket?
March 27, 2026

Can Palo Alto’s Agentic NetOps Actually Kill the NOC Ticket?

Palo Alto Networks' agentic AI Troubleshooting Agent in Prisma SD-WAN enables autonomous network remediation, reducing resolution time from hours to minutes, marking AI-native networking's transition to production....
Infosys Bets on Anthropic to Survive the Automation Wave It Helped Build
March 27, 2026

Infosys Bets on Anthropic to Survive the Automation Wave It Helped Build

Infosys expands its Anthropic partnership to develop enterprise AI agents, signaling that its labor arbitrage model faces disruption and reflecting an urgent pivot toward AI-first service delivery....

Book a Demo

Newsletter Sign-up Form

Get important insights straight to your inbox, receive first looks at eBooks, exclusive event invitations, custom content, and more. We promise not to spam you or sell your name to anyone. You can always unsubscribe at any time.

All fields are required






Thank you, we received your request, a member of our team will be in contact with you.