FMS 2026: Liqid launches the industry's most advanced CXL memory pooling platform for AI and scientific discovery. With Micron Technology in powering the Pacific Northwest National Laboratory's large-scale memory-centric computing platform. Liqid, a player in software-defined memory and GPU pooling infrastructure, announced its role as a strategic technology partner in Abaco, the memory-centric computing system designed by Micron, a subcontractor to the US Department of Energy's (DOE) Pacific Northwest National Laboratory (PNNL).Liqid is providing the scale-up memory pool that allows PNNL to expose hundreds of terabytes of coherent active memory to data-intensive AI for Science workloads. Solving the Memory Wall for Scientific AI For decades, memory architectures have limited applications such as computational chemistry, molecular biology, and AI-driven simulation by restricting access to the large memory capacity they require. The Abaco system pioneers CXL-enabled memory pooling and sharing, shattering the capacity limits of individual servers to support massive-scale AI workloads. Liqid delivers the industry's first and only fully disaggregated, software-defined memory pooling solution. In a Liqid rack-scale configuration, up to 160TB of DRAM can be dedicated to a single host or dynamically allocated and shared across as many as 16 server nodes. Real-time allocation of terabytes of memory enables zero stranded capacity and massive performance scaling when needed. What Liqid Contributes to PNNL * Liqid EX-5410C Memory Platform: CXL 2.0 memory expansion system supporting up to 40TB of DRAM per chassis and scaling to a unified memory pool of more than 160TB for memory-intensive AI, HPC, and database workloads * Liqid Matrix Software: The orchestration and management layer that turns pooled memory into a programmable resource. Liqid Matrix delivers the industry's only unified interface for real-time deployment, management, and orchestration of composable GPU, memory, and storage - and integrates with Kubernetes, Slurm, Ansible, and other tooling already used in scientific computing environments * Liqid CXL 2.0 Fabric: A dedicated CXL 2.0 switch and HBAs deliver ultra-low-latency, high-bandwidth connectivity between hosts and the memory pool Memory Sharing and Benchmark Results Liqid and Micron have also collaborated to enable the industry's first architecture that supports true memory-sharing capabilities. Micron's famfs Linux kernel enhancement enables advanced memory sharing across systems with software cache coherence, laying the foundation for future performance gains through larger shared memory pools, reduced data movement, and more efficient access to massive datasets. With famfs, disaggregated memory pools from the Liqid system become exposed to the host operating system as a file system, requiring no application layer changes to harness the latency and bandwidth advantages of pooled memory. Using the famfs Linux kernel in the Abaco system with Micron high-capacity RDIMMs, Micron lab benchmarks using representative workloads demonstrated performance improvements. Results showed up to 30x faster processing time for graph analytics workloads and up to 7x more tokens per second for certain KV cache workloads, validating the power of CXL memory pooling for AI and scientific computing, with additional optimization and performance gains expected as the platform enables new features. Executive Perspectives "PNNL is exactly the kind of environment that proves what composable CXL memory can do at the frontier of science. PNNL and Micron have built something singular: a memory-centric system designed for the workloads that could define the next decade of AI for Science. Liqid is honored to extend that platform with the scale, orchestration, and composability that turn pooled memory into a programmable resource. This is what tokens per dollar and tokens per watt look like in a national laboratory environment," said Sumit Puri, founder and CTO, Liqid. "PNNL's work demonstrates how CXL-based memory pooling can extend system memory resources beyond the limits of individual servers. By combining Micron high-capacity DRAM with a disaggregated memory architecture, the Abaco system provides researchers with access to larger pools of memory for scientific computing workloads. This project highlights an innovative approach to addressing memory challenges as computing requirements continue to grow." added Vijay Nain, senior director, emerging memory and ecosystem development, Micron Technology. "The Abaco system testbed design increases memory capacity by orders of magnitude, which is critical for scientific workloads. We look forward to quantifying that performance improvement in HPC, AI inference, and post-training applications," concluded James A. Ang, Ph.D., chief scientist, computing, IDSD, Pacific Northwest National Laboratory. Why It Matters for AI for Science PNNL will leverage Abaco to advance AI for Science, with AI-integrated computational chemistry as the first user-facing workload. By extending the lab's pooled memory with Liqid's CXL memory fabric, the system can keep larger structures, larger parameter spaces, and heavier intermediate data resident in active memory, directly attacking the memory bottleneck that has historically constrained scientific AI. PNNL's testbed is available to DOE-funded researchers from national laboratories and academia through the Advanced Memory to Support Artificial Intelligence for Science (AMAIS) initiative, funded by the DOE Advanced Scientific Computing Research program within the DOE Office of Science. Read also: Switchtec PCIe Switches and Micron 9650 NVMe SSDs leverage next generation interface technology to deliver increased bandwidth, low latency, and scalable storage connectivity Industry veteran joins Liqid's executive team to lead product management, strategy, and engineering as company's software-defined memory and GPU pooling solutions gain traction Company launches employee matching and community seed funding, reinforcing Micron's broader U.S. investment and workforce development strategy Supply commitment supports next-generation GM vehicle platforms with reliable, high-performance memory and storage solutions Generating $41.6 billion, up 74.3% QoQ and up 347.3% YoY Solutions deliver high-performance, agile, and efficient GPU, memory, and storage scale-up and scale-out to support AI inference and other power-hungry, compute-intensive workloads.

Liqid Unveils CXL Memory Pooling Platform for AI and Scientific [...]. Liqid Unveils CXL Memory Pooling Platform for AI and Scientific Discover August 4, 2026 - Liqid is supplying the pooled memory layer for Abaco, a Micron designed system being built for the U.S. Department of Energy's Pacific Northwest National Laboratory. The platform can expose more than 160 terabytes of coherent memory across multiple servers, targeting AI and scientific workloads constrained less by processors than by how quickly datasets can be reached efficiently. The deployment gives Liqid a visible reference customer for a class of infrastructure that vendors have discussed for years but enterprises have adopted cautiously. Compute capacity has expanded rapidly. Memory capacity, bandwidth, and locality have not kept pace. Large graph models, scientific simulations, [...] [... Check source for end of article...]

Liqid names Daniel Berg Chief Product Officer. Industry veteran joins Liqid's executive team to lead product management, strategy, and engineering as company's software-defined memory and GPU pooling solutions gain traction. Liqid, a player in software-defined memory and GPU pooling infrastructure, announced the appointment of Daniel Berg as CPo. Berg joins Liqid with more than three decades of product and engineering leadership, including executive roles at Gorilla Logic, JumpCloud, NetApp, Skype, and Sun Microsystems. He will own product management, strategy and engineering for Liqid's entire product portfolio including Liqid Matrix and the company's software-defined memory and GPU pooling infrastructure. Berg will play a key role as the company deepens its OEM relationships and continues its focus on solving the AI infrastructure challenges facing Neoclouds and enterprises. Berg most recently served as the CEO of Gorilla Logic, a nearshore software services company and prior to that was the CTO of JumpCloud, a cloud-based identity and unified IT management platform. Before JumpCloud, Berg was Vice president and GM of the SolidFire business unit at NetApp, following NetApp's acquisition of SolidFire, where he had served as Chief Product Officer. Earlier in his career, Berg was CTO and head of Research and Development at Skype and spent 15 years at Sun Microsystems in engineering and management roles, where he was named a Distinguished Engineer. Berg has written several books and papers on Java, concurrent programming, and software development methodology and architecture, and has long championed diversity in engineering and continuous integration and deployment practices that improve how engineering teams communicate and ship. Liqid pools and dynamically allocates and scales CXL memory and GPUs across a datacenter to support workloads in real time, so enterprises stop paying for capacity that sits idle. The approach is gaining traction as AI inference, KV Cache, and in-memory databases push vs. the limits of legacy server architecture, where memory is locked to a single server and GPUs sit unused between jobs. Berg's product background, spanning storage, cloud identity, and real-time communications infrastructure, brings a unique insight and influence to Liqid's roadmap as the company works with OEM and ecosystem partners including, AMD, Cisco, Dell, Micron, NVIDIA, and Penguin Solutions to bring memory and GPU pooling infrastructure into mainstream neocloud and enterprise deployment. Executive Perspectives "Dan has built products at companies that had to get the technology right the first time. He knows how to take something technically hard and turn it into solutions that operations teams can and want to run. That's exactly the product leadership we need as our memory and GPU pooling solutions become part of standard buying cycles," said Rick Hegberg, CEO, Liqid. "The challenge I see today in deploying modern enterprise and neocloud AI infrastructure is that it is expensive, rigid, and poorly utilized. Liqid solves this at the fabric level, with solutions that work alongside the hardware neoclouds and enterprises already own. I've spent my career building product organizations that solve hard problems by bringing game-changing products to market, and I'm looking forward to doing that here," said Daniel Berg, CPO, Liqid. Read also: Solutions deliver high-performance, agile, and efficient GPU, memory, and storage scale-up and scale-out to support AI inference and other power-hungry, compute-intensive workloads.