MemVerge
MemVerge is a privately held AI infrastructure software company (founded 2017, Milpitas, CA) that builds big-memory computing software and the open-source MemMachine AI memory layer, helping enterprises and cloud providers overcome the GPU memory wall and add persistent context to AI agents.
- Company typePrivate
- Founded2017
- HeadquartersMilpitas, United States
- Headcount11–50
- GTM typeB2B
- OfferingSoftware
What MemVerge does
MemVerge, Inc. is a privately held AI infrastructure software company headquartered in Milpitas, California, founded in 2017 by Charles Fan and Shuki Bruck. The company develops big-memory computing software for GPU-centric AI workloads, organized around two product lines. The Memory Machine family (Memory Machine AI, Memory Machine Cloud, Memory Machine X for CXL, Memory Machine Batch for AWS, and the Checkpoint Engine) provides container orchestration, GPU scheduling, transparent checkpointing, and CXL-based memory pooling and multi-node sharing to address the "memory wall" in AI infrastructure. In September 2025, the company launched MemMachine, an open-source AI memory layer providing episodic, profile, and procedural memory for AI agents, which achieved an 84.87% LoCoMo benchmark score and was expanded in December 2025 with native MCP support, Python SDKs, and integrations with LangGraph, AWS Strands, Claude Code, n8n, and GPTStore.
The company targets financial services (American Express, Intesa Sanpaolo, and unspecified major banks), scientific computing (leading universities and research institutes such as TGen), cloud service providers (ByteDance, Tencent Cloud, Alibaba Cloud, MemX), and emerging verticals including healthcare and enterprise productivity. Its commercial model combines freemium open-source distribution (MemMachine), a $2,500 per month White Glove service tier, and custom-priced multi-year enterprise contracts with dedicated account managers and solution engineers; the product is also available through the AWS Marketplace. MemVerge has assembled a strategic investor base of Broadcom, Cisco Investments, Intel Capital, Samsung, SK hynix, NetApp, and Lightspeed, alongside technology partnerships with NVIDIA (Inception Program), AMD, Micron, XConn, Cadence, Sentieon, AWS, Google Cloud, VMware, HPE, and Red Hat. Headcount remains in the 11-50 range, with revenue not publicly disclosed.
MemVerge firmographics
Firmographics- Name
- MemVerge
- Legal name
- MemVerge, Inc.
- Website
- https://memverge.ai
- Company type
- Private
- Founded year
- 2017
- Operating status
- Operating
- Headcount range
- 11–50 employees
- Short description
- MemVerge is a privately held AI infrastructure software company (founded 2017, Milpitas, CA) that builds big-memory computing software and the open-source MemMachine AI memory layer, helping enterprises and cloud providers overcome the GPU memory wall and add persistent context to AI agents.
- Ownership category
- akta.pro rank
MemVerge industry classification
Industry- Product category
- AI Infrastructure Software
- NAICS
- Software Publishers (5132), Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (518210)
- SIC
- Services-Prepackaged Software (7372), Computer Storage Devices (3572)
- akta.pro primary industry
- AI Memory & Storage for Training (HBM/DDR/SSD/Object Storage) (HDAAAAAF)
- akta.pro secondary industries
- AI Compute Virtualization & Scheduling (GPU virtualization, cluster schedulers) (HDAAAAAG), AI Compiler, Runtime & Kernel Optimization Software (CUDA/ROCm/XLA, graph compilers) (HDAAAAAI), Memory Modules & Subsystems (DIMM/SO-DIMM/HBM stacks) (HDAHAKAH)
Keywords
Where MemVerge is headquartered
LocationHeadquarters
- HQ city
- Milpitas
- HQ country
- United States
- HQ region
- North America
Offices1 record
Markets served
MemVerge business model
Business model- GTM type
- B2B
- Offering type
- Software
- Cost components
- Personnel, Technology or R&D, Marketing or Sales, Operations, Infrastructure
Revenue model
- Software Licenses and Subscriptions: Memory Machine software products sold as enterprise licenses with subscription pricing. Products include MemMachine Enterprise, Memory Machine AI, Memory Machine X for CXL, and Memory Machine Batch for AWS. Pricing varies by deployment scale and features.
- Professional Services: White Glove Service tier at $2,500/month includes up to 10 hours of professional services, dedicated Slack support, cloud-based test-dev environment, and named account manager for smaller organizations seeking accelerated AI agent deployment.
- Open Source Community: Open source MemMachine product drives ecosystem adoption and land-and-expand motion into enterprise deployments. Community features (Discord, GitHub, playground) serve as top-of-funnel for commercial product sales.
Pricing tiers
| Model | Billing | Price |
|---|---|---|
| Freemium | Free | Playground - Free tier for experiencing AI memory |
| Subscription | Monthly | White Glove Service - $2,500/month for accelerated development |
| Subscription | Multi-year contract | Enterprise Solution - Custom pricing |
Go-to-market motion3 records
Distribution channels4 records
Marketing channels11 records
MemVerge product offering
Product offeringCore offering
MemVerge develops and sells AI infrastructure software that optimizes big memory computing for GPU-centric AI workloads. Its Memory Machine product family provides intelligent memory tiering, GPU orchestration, transparent checkpointing, and CXL-based memory pooling and sharing, while MemMachine provides an open-source AI memory layer that gives AI agents persistent episodic, profile, and procedural memory. Products are sold as enterprise licenses and subscriptions with a freemium open-source entry tier.
Product overview
MemVerge offers a comprehensive AI infrastructure software portfolio centered around Big Memory Computing for GPU-centric AI. The core Memory Machine product family includes Memory Machine AI (container orchestration for data-intensive workloads), Memory Machine Cloud (cloud automation), Memory Machine X for CXL (memory expansion and sharing), and Memory Machine Batch for AWS (cloud batch processing). The MemMachine AI Memory Layer is the intelligent memory system providing episodic, profile, and procedural memory for AI agents, available as open source and enterprise editions. Supporting initiatives include AI Infra Forum, Memory Fabric Forum, and AI Memory Forum for community education. Solutions like WaveRider for Sentieon address specialized scientific computing use cases.
Differentiator
Problem solved
Functional benefit
Brands
- MemMachine: Open source AI memory layer system for advanced AI agents that provides persistent, intelligent memory across sessions, models, and environments.
- Memory Machine AI
- Memory Machine Cloud
- Memory Machine X for CXL
- Memory Machine Batch
- Checkpoint Engine
Products and services
- MemMachine Open Source Open-source AI memory layer for advanced AI agents that enables applications to learn, store, and recall data and preferences from past sessions to enrich future interactions; memory persists across multiple sessions, agents, and large language models. Targeted at developers and AI agent builders adopting via memmachine.ai free playground.
- MemMachine Enterprise Commercial version of MemMachine providing enterprise-class security, compliance, and scalability with customized, secure deployment in customer's choice of environment. Sold to enterprises requiring persistent, intelligent AI agent memory with enterprise controls.
- Memory Machine AI Suite of container orchestration services for running data-intensive pipelines such as bioinformatics and interactive computing applications. Includes GPU scheduling, memory tiering, and memory sharing to maximize GPU utilization and break the memory wall for AI workloads. Targets enterprise AI/ML engineering teams.
- Memory Machine Cloud Cloud-based container orchestration platform for automating AI workload pipelines on-premises and in the cloud, enabling GPU-as-a-Service deployment within enterprises and delivering 50%+ cloud cost savings for scientific computing customers.
- Memory Machine X for CXL Software for CXL memory expansion and multi-node memory sharing, enabling big memory data center redesigns and acceleration of AI workloads through Memory Machine technology. Demonstrated 100 TiB-scale CXL memory pools for AI inference.
- Memory Machine Batch for AWS Batch processing solution for AWS environments that provides transparent checkpointing and restore of jobs running on AWS Batch, enabling high-quality-of-service on unreliable Spot instances by capturing entire running state for hot restarts.
- Memory Machine Checkpoint Engine Powerful checkpointing and restore engine designed for integration with Kubernetes, job schedulers, and AWS Batch environments. Captures entire running state of CPU and GPU jobs for hot restarts without losing work progress.
- WaveRider for Sentieon Joint solution with Sentieon that accelerates next-generation sequencing (NGS) workflows in the cloud by leveraging Memory Machine technology.
- MemVerge.ai Intelligent Memory (AWS Marketplace) MemVerge.ai Intelligent Memory software available in AWS Marketplace AI Agents and Tools category, enabling cloud-native enterprise customers to discover and procure the offering through AWS infrastructure.
Quantifiable outcome
- Double GPU utilization for AI workloads at major banks
- +9 more outcomes
Companies that use MemVerge
Customer profileNamed customers11 records
Segments5 records
Ideal customer profiles4 records
MemVerge technology and API
TechnologyTechnology focussed Yes
API detail
- Has API
- Yes
- API docs
- API detail
Core technology
AI maturity
App detail
Integration8 records
AI capability6 records
Feature8 records
MemVerge partnerships and signals
Strategic signalPartnerships
15 partnerships are on record, tiered core and minor.
- XConn TechnologiescoreJoint demonstration of 100 TiB-scale CXL memory pool for AI and inference workloads at SC25 in St. Louis. Partnership addresses the memory wall challenge in AI infrastructure through scalable CXL memory pooling solution enabling KV cache offload and prefill/decode disaggregation.
- XConn TechnologiescoreDemonstrated CXL memory pool for KV Cache using NVIDIA Dynamo at 2025 OCP Global Summit in San Jose. Joint solution showcases scalable memory up to 100 TiB addressing memory wall challenge in AI applications.
- LangGraphcoreEcosystem integration with LangGraph framework enabling developers to build context-aware AI agents with persistent memory using LangGraph's agent orchestration capabilities.
- AWS Strands AgentscoreDeeper ecosystem integration with AWS Strands Agents framework for building AI agents on AWS infrastructure with native MemMachine memory capabilities.
- Claude CodecoreEcosystem integration enabling Claude Code developers to leverage MemMachine's persistent memory capabilities for code generation and development workflows.
- n8ncoreIntegration with n8n workflow automation platform enabling AI agents built on n8n to utilize MemMachine memory for persistent context across workflow executions.
- MicroncoreJoint solution leveraging intelligent tiering of CXL memory to boost GPU utilization by 77% and more than double speed of OPT-66B batch inference. Addresses memory wall in AI infrastructure.
- AMDcoreMemVerge brings intelligent scheduling and resource management to AMD Instinct GPUs. AMD's cloud infra team validated MemVerge's solution for mature, intuitive user interface and enterprise-grade scheduling capabilities.
- SamsungcoreJoint demonstration of memory pooling and sharing for 'Endless Memory' with Samsung and XConn, showcasing multi-node CXL memory sharing technology.
- SK hynixcoreEndless Memory partnership for memory pooling and sharing software development with CXL Flight Simulator. SK hynix integrated HMSDK into Linux for CXL 2.0 memory optimization.
- AWScoreAWS Partner Network membership. MemVerge products available on AWS Marketplace. Memory Machine Batch designed for AWS Batch environments.
- Google CloudminorGoogle Cloud Partner network membership for cloud deployment and go-to-market alignment.
- VMwareminorVMware Partner network membership for enterprise virtualization and cloud infrastructure integration.
- HPEminorHPE Partner Ready program membership for enterprise hardware and solutions integration.
- NVIDIAcoreNVIDIA Inception Program member. Joint solutions demonstrating CXL memory for NVIDIA GPU environments. Checkpoint Engine integration for NVIDIA GPU workloads.
Scale indicators6 records
Recent moves6 records
Expansion highlights7 records
MemVerge competitors and assessment
Company assessmentDirect peers
- Weka: Weka offers a GPU-optimized data platform for AI workloads, sitting in nearly the same architectural layer as MemVerge's Memory Machine AI — accelerating AI/ML pipelines with high-throughput, low-latency storage.
- Vast Data: Vast Data provides a unified data platform combining storage, database, and compute services for AI workloads, addressing the same memory and data tiering bottlenecks MemVerge targets for GPU-centric AI infrastructure.
- DDN (DataDirect Networks): DDN's EXAScaler and AI storage appliances are widely deployed for GPU clusters and HPC — directly competing with MemVerge's Memory Machine for high-throughput AI infrastructure pipelines.
- Hammerspace: Hammerspace offers a global data orchestration platform with strong GPU/HPC alignment, comparable in customer base (scientific computing, financial services) and in addressing data mobility across AI infrastructure.
- Lightbits Labs: Lightbits provides disaggregated NVMe-over-TCP storage with high performance for AI workloads, overlapping with MemVerge's tiered memory and storage efficiency value proposition.
- Mem0: Mem0 provides a memory layer specifically for AI agents/LLMs — the most direct open-source competitor to MemVerge's MemMachine in the AI agent memory category.
Emerging players
- Pliops: Pliops builds storage accelerators for AI/ML databases, attacking similar I/O and throughput bottlenecks in GPU clusters; smaller and more focused than MemVerge's full software stack.
- Pinecone: Pinecone is a managed vector database that, while focused on embeddings, competes for the same AI memory/retrieval budget as MemVerge's semantic memory features in MemMachine.
Broad incumbents
- Pure Storage: Pure Storage is a broad enterprise storage incumbent with active AI infrastructure positioning (AIRI); overlaps with MemVerge on AI workload acceleration but at a much larger scale and broader portfolio.
- NetApp: NetApp is a global enterprise storage leader and an investor in MemVerge. It offers AI-ready data management and ONTAP-based solutions that compete for the same enterprise AI infrastructure dollars.
Market position
Strengths4 records
Weaknesses4 records
Competitive moat6 records
Key risks6 records
Key highlights6 records
Customer concentration
MemVerge social profiles
Digital presenceMemVerge financial estimates
Financial estimateRevenue estimate
Valuation estimate
MemVerge leadership team
Management profileNumber of profiles
Profiles6 records
MemVerge funding detail
Funding detailFunding overview
Funding rounds5 records
Investors14 records
Funding detail is available on the Subscription and Enterprise plan.Contact sales →
MemVerge M&A and investment
M&A and investmentM&A
Investments
M&A and investment is available on the Subscription and Enterprise plan.Contact sales →
Frequently asked questions about MemVerge
What does MemVerge do?
MemVerge develops and sells AI infrastructure software that optimizes big memory computing for GPU-centric AI workloads. Its Memory Machine product family provides intelligent memory tiering, GPU orchestration, transparent checkpointing, and CXL-based memory pooling and sharing, while MemMachine provides an open-source AI memory layer that gives AI agents persistent episodic, profile, and procedural memory. Products are sold as enterprise licenses and subscriptions with a freemium open-source entry tier.
Is MemVerge a public or private company?
MemVerge is a private company. It is classified as venture growth investor backed and is currently operating.
When was MemVerge founded?
MemVerge was founded in 2017. It employs 11 to 50 people.
Where is MemVerge based?
MemVerge is headquartered in Milpitas, United States, in the North America region.
How does MemVerge make money?
Three revenue lines are on record. Software Licenses and Subscriptions are the primary driver. The others are professional Services and open Source Community.
Who are MemVerge's main competitors?
Direct peers on record are Weka, Vast Data, DDN (DataDirect Networks), Hammerspace, Lightbits Labs and Mem0. Emerging players are Pliops and Pinecone. Broad incumbents are Pure Storage and NetApp.
Does MemVerge have an API?
Yes. MemMachine offers Python SDKs for client and server, a more powerful REST API, and native Model Context Protocol (MCP) support to broaden compatibility across deployment environments. Developer documentation is at memmachine.ai.
What industry is MemVerge in?
MemVerge's product category is AI Infrastructure Software. Its primary akta.pro industry code is HDAAAAAF, AI Memory & Storage for Training (HBM/DDR/SSD/Object Storage), with a secondary code of HDAAAAAG, AI Compute Virtualization & Scheduling (GPU virtualization, cluster schedulers). Its NAICS code is 5132 and its SIC code is 7372.