The global in-memory computing market was valued at USD 24.00 billion in 2025. It is estimated at USD 27.36 billion in 2026 and is projected to reach USD 94.53 billion by 2036, growing at a CAGR of 13.2% over the 2026-2036 forecast period. This report assesses the market in depth, covering the technology landscape, DRAM and HBM supply and pricing, the open-source and licensing landscape, regulation and security, pricing and business models, the value chain, and the competitive landscape.
The study covers software, hardware, and services that keep and process data mainly in main memory, or that carry out computation inside or next to memory, instead of shuttling data back and forth between storage, memory, and processors. Software spans databases like SAP HANA and Oracle Database In-Memory, grid and cache products like Redis, Valkey, Hazelcast, and GridGain, analytics, and stream processing. Hardware spans processing-in-memory (PIM) parts that place logic in DRAM or HBM, compute-in-memory (CIM) chips that run AI math inside SRAM or non-volatile arrays, and appliances. Services span consulting, implementation, migration, and managed services. Plain memory chips and servers sold without in-memory software or functions are excluded.
SAP's results show how closely enterprise demand is tied to ERP modernization and AI, as the company, whose S/4HANA ERP is built on HANA, reported that its 2025 cloud revenue grew 23% to EUR 21.02 billion and that its cloud backlog reached a record EUR 77.29 billion, adding that Business AI appeared in two-thirds of fourth-quarter cloud orders. On the hardware side, energy is a growing concern: the International Energy Agency projects that data center electricity use will more than double from about 415 TWh in 2024 to around 945 TWh by 2030, which adds pressure to cut the energy spent moving data between memory and processors.
Rising DRAM prices are the most immediate restraint, since in-memory systems depend on large memory footprints; Samsung was reported to have raised some memory chip prices by up to 60% in December 2025, and both Samsung and SK hynix reported more 2026 orders than they could supply. The 2024 Redis license change and the resulting Valkey fork created uncertainty over licensing, compatibility, and long-term support. Migrating legacy databases and ERP systems takes years of data migration, code redesign, and testing, and in-memory hardware is still immature, facing issues with precision, device variability, density, and software support. Further challenges include keeping data durable in volatile memory after Intel wound down its Optane persistent memory business, securing exposed in-memory stores, fitting processing-in-memory into programming models and standards, and a shortage of skilled architects and consultants.
Vector search and AI feature serving offer the largest opportunity, as retrieval-augmented generation, recommendation engines, and AI agents need fast access to embeddings and context; SAP reported that 90% of its 50 largest fourth-quarter 2025 deals included AI or its Business Data Cloud. Instant payment systems such as India's UPI, Brazil's Pix, and the U.S. FedNow service need transactions authorized and screened within seconds, and the push for low-power AI inference in devices and data centers opens room for compute-in-memory chips. Key trends include in-memory platforms adding vector search and AI features, wider use of open-source forks and fully managed services, Compute Express Link (CXL) for adding and sharing memory across servers, and PIM inside AI accelerators.
By component, the market is divided into software, hardware, and services, each split into detailed product and service types, from databases and caches to PIM devices, CIM accelerators, and migration and managed services. Software is expected to hold the largest share in 2026, as in-memory databases, caches, and analytics platforms are already used widely across enterprise and cloud applications, from ERP and payments to web sessions, and are now being extended with vector search for AI. Hardware is expected to grow at the highest CAGR, as AI's memory bottleneck pulls PIM devices and CIM accelerators out of the lab and into first commercial shipments.
By deployment, the report covers cloud, on-premises, and hybrid models. Cloud deployment is expected to account for the largest share in 2026, supported by the rapid growth of managed in-memory database and caching services from the major cloud providers and by the shift of ERP to cloud subscriptions, as seen in SAP's Cloud ERP Suite growth. Hybrid deployment is expected to grow at the highest CAGR, as data sovereignty rules, including GDPR in Europe, and the need to link on-premises systems with cloud services lead organizations to run in-memory workloads across both environments.
By hardware technology, the market is divided into DRAM & HBM-based processing-in-memory, digital SRAM-based compute-in-memory, and analog non-volatile compute-in-memory. Digital SRAM-based compute-in-memory is expected to hold the largest share of the hardware market in 2026, as it works with standard CMOS manufacturing, avoids the precision and noise issues of analog designs, and has more mature commercial products, such as Axelera AI's edge AI chips. Analog non-volatile compute-in-memory is expected to grow at the highest CAGR, because it could deliver far more AI inference per watt than other designs, with companies such as EnCharge AI and Mythic and IBM Research advancing analog in-memory designs. Memory-based PIM from Samsung and SK hynix depends on progress in software and standards.
By application, the study covers real-time transaction processing, real-time analytics & business intelligence, caching & session management, fraud detection & risk management, AI/ML feature stores & vector search, edge AI inference, and ERP & supply chain. Real-time transaction processing is expected to lead in 2026, as banks, retailers, and telecom operators use in-memory databases for high-volume transactional systems where delays directly affect revenue and customer experience. AI/ML feature stores & vector search are expected to grow at the highest CAGR, driven by the fast rollout of generative AI and retrieval-augmented applications that need low-latency access to embeddings.
By organization size, the market is split into large enterprises and small & medium enterprises. Large enterprises run the most demanding in-memory workloads, including SAP S/4HANA ERP systems, trading and payment platforms, and large-scale analytics, and they lead most multi-year transformation programs. Small and medium enterprises are adopting in-memory databases and caches mainly as managed cloud services, which remove the need to run distributed clusters in-house and lower entry costs. The report compares adoption and deployment choices across the two groups.
By industry vertical, the report covers BFSI, retail & e-commerce, telecommunications & IT, manufacturing, healthcare & life sciences, government & defense, media, gaming & entertainment, and transportation & logistics. BFSI is expected to account for the largest share in 2026, as banks, insurers, and exchanges need instant payment handling, trading, risk control, and fraud checks, together with resilience rules such as the EU's Digital Operational Resilience Act. Retail & e-commerce is expected to grow at the highest CAGR, as online sellers use live personalization, stock tracking, and AI-based shopping features.
North America is expected to hold the largest share of the market in 2026, as it hosts the leading cloud providers and in-memory platform vendors, has large financial services and technology sectors, and adopted AI early. Most major platform vendors, including Oracle, Microsoft, Amazon Web Services, Google, Redis, and Aerospike, are based in the U.S., as are CIM startups such as d-Matrix and EnCharge AI, and real-time payment services such as FedNow and demand for fraud detection support further investment. Europe holds a significant share as the home of SAP, with legacy ERP migrations ahead of 2027 driving demand in Germany, France, the U.K., and the Nordic countries.
Asia-Pacific is expected to grow at the highest CAGR during the forecast period, driven by fast digital adoption, huge instant payment flows, and its position as the center of global memory chip production. India's UPI handles billions of transactions a month, South Korea's Samsung and SK hynix are developing processing-in-memory products, while banks, online retailers, and telecom operators in China, Japan, Singapore, and Australia are expanding their use of in-memory platforms. In Latin America, Brazil's Pix system and digital banks in Mexico, Colombia, Chile, and Argentina support demand, while in the Middle East & Africa, AI and digital government investment in the UAE and Saudi Arabia, Israel's technology base, and mobile money in South Africa and Kenya drive growth.
The market brings together major business software and database suppliers, hyperscale clouds selling hosted in-memory services, specialist database and data grid firms, open-source communities, memory makers developing processing-in-memory, and startups building compute-in-memory AI chips. Competition centers on latency and performance, scalability and resilience, AI and vector search features, licensing openness, cloud integration, and total cost of ownership, and in hardware on performance per watt and toolchain maturity. Leading companies are building vector and AI features into their platforms, shifting to subscriptions, answering open-source forks with new licenses and hosted offerings, supporting CXL and memory tiering, and funding in-memory chip development. Recent moves include the launch of Valkey with backing from Amazon Web Services, Google Cloud, and Oracle, the AGPL option added with Redis 8 in 2025, and EnCharge AI's 2025 funding round.
Key companies profiled in the report include SAP SE (Germany), Oracle Corporation (U.S.), Microsoft Corporation (U.S.), Amazon Web Services, Inc. (U.S.), Google LLC (U.S.), IBM Corporation (U.S.), Redis Ltd. (U.S.), Aerospike, Inc. (U.S.), SingleStore, Inc. (U.S.), GridGain Systems, Inc. (U.S.), Hazelcast, Inc. (U.S.), GigaSpaces Technologies Ltd. (Israel), Samsung Electronics Co., Ltd. (South Korea), SK hynix Inc. (South Korea), d-Matrix Corporation (U.S.), EnCharge AI, Inc. (U.S.), Axelera AI B.V. (Netherlands), and Mythic, Inc. (U.S.).
The study covers software, hardware, and services that keep and process data mainly in main memory, or that carry out computation inside or next to memory, instead of shuttling data back and forth between storage, memory, and processors. Software spans databases like SAP HANA and Oracle Database In-Memory, grid and cache products like Redis, Valkey, Hazelcast, and GridGain, analytics, and stream processing. Hardware spans processing-in-memory (PIM) parts that place logic in DRAM or HBM, compute-in-memory (CIM) chips that run AI math inside SRAM or non-volatile arrays, and appliances. Services span consulting, implementation, migration, and managed services. Plain memory chips and servers sold without in-memory software or functions are excluded.
SAP's results show how closely enterprise demand is tied to ERP modernization and AI, as the company, whose S/4HANA ERP is built on HANA, reported that its 2025 cloud revenue grew 23% to EUR 21.02 billion and that its cloud backlog reached a record EUR 77.29 billion, adding that Business AI appeared in two-thirds of fourth-quarter cloud orders. On the hardware side, energy is a growing concern: the International Energy Agency projects that data center electricity use will more than double from about 415 TWh in 2024 to around 945 TWh by 2030, which adds pressure to cut the energy spent moving data between memory and processors.
Market Dynamics
Growth is driven by the need to process data and serve AI models in real time. Payments, fraud screening, personalization, and AI applications must respond in milliseconds, which requires data to sit in memory rather than on disk, and in-memory stores now also hold embeddings, model features, and chat session data for generative AI tools. ERP modernization is a second driver, as SAP customers move from the legacy ERP Central Component, whose mainstream maintenance ends in 2027, to S/4HANA on HANA, lifting demand for database capacity, cloud infrastructure, and migration services. The memory wall in AI hardware is a third, as AI performance is now limited more by the bandwidth and energy of data movement than by compute, prompting memory makers and startups to move computation into or next to memory.Rising DRAM prices are the most immediate restraint, since in-memory systems depend on large memory footprints; Samsung was reported to have raised some memory chip prices by up to 60% in December 2025, and both Samsung and SK hynix reported more 2026 orders than they could supply. The 2024 Redis license change and the resulting Valkey fork created uncertainty over licensing, compatibility, and long-term support. Migrating legacy databases and ERP systems takes years of data migration, code redesign, and testing, and in-memory hardware is still immature, facing issues with precision, device variability, density, and software support. Further challenges include keeping data durable in volatile memory after Intel wound down its Optane persistent memory business, securing exposed in-memory stores, fitting processing-in-memory into programming models and standards, and a shortage of skilled architects and consultants.
Vector search and AI feature serving offer the largest opportunity, as retrieval-augmented generation, recommendation engines, and AI agents need fast access to embeddings and context; SAP reported that 90% of its 50 largest fourth-quarter 2025 deals included AI or its Business Data Cloud. Instant payment systems such as India's UPI, Brazil's Pix, and the U.S. FedNow service need transactions authorized and screened within seconds, and the push for low-power AI inference in devices and data centers opens room for compute-in-memory chips. Key trends include in-memory platforms adding vector search and AI features, wider use of open-source forks and fully managed services, Compute Express Link (CXL) for adding and sharing memory across servers, and PIM inside AI accelerators.
Segment Analysis
The report analyzes the market by component, deployment, hardware technology, application, organization size, industry vertical, and geography, helping stakeholders identify where value and growth are concentrated.By component, the market is divided into software, hardware, and services, each split into detailed product and service types, from databases and caches to PIM devices, CIM accelerators, and migration and managed services. Software is expected to hold the largest share in 2026, as in-memory databases, caches, and analytics platforms are already used widely across enterprise and cloud applications, from ERP and payments to web sessions, and are now being extended with vector search for AI. Hardware is expected to grow at the highest CAGR, as AI's memory bottleneck pulls PIM devices and CIM accelerators out of the lab and into first commercial shipments.
By deployment, the report covers cloud, on-premises, and hybrid models. Cloud deployment is expected to account for the largest share in 2026, supported by the rapid growth of managed in-memory database and caching services from the major cloud providers and by the shift of ERP to cloud subscriptions, as seen in SAP's Cloud ERP Suite growth. Hybrid deployment is expected to grow at the highest CAGR, as data sovereignty rules, including GDPR in Europe, and the need to link on-premises systems with cloud services lead organizations to run in-memory workloads across both environments.
By hardware technology, the market is divided into DRAM & HBM-based processing-in-memory, digital SRAM-based compute-in-memory, and analog non-volatile compute-in-memory. Digital SRAM-based compute-in-memory is expected to hold the largest share of the hardware market in 2026, as it works with standard CMOS manufacturing, avoids the precision and noise issues of analog designs, and has more mature commercial products, such as Axelera AI's edge AI chips. Analog non-volatile compute-in-memory is expected to grow at the highest CAGR, because it could deliver far more AI inference per watt than other designs, with companies such as EnCharge AI and Mythic and IBM Research advancing analog in-memory designs. Memory-based PIM from Samsung and SK hynix depends on progress in software and standards.
By application, the study covers real-time transaction processing, real-time analytics & business intelligence, caching & session management, fraud detection & risk management, AI/ML feature stores & vector search, edge AI inference, and ERP & supply chain. Real-time transaction processing is expected to lead in 2026, as banks, retailers, and telecom operators use in-memory databases for high-volume transactional systems where delays directly affect revenue and customer experience. AI/ML feature stores & vector search are expected to grow at the highest CAGR, driven by the fast rollout of generative AI and retrieval-augmented applications that need low-latency access to embeddings.
By organization size, the market is split into large enterprises and small & medium enterprises. Large enterprises run the most demanding in-memory workloads, including SAP S/4HANA ERP systems, trading and payment platforms, and large-scale analytics, and they lead most multi-year transformation programs. Small and medium enterprises are adopting in-memory databases and caches mainly as managed cloud services, which remove the need to run distributed clusters in-house and lower entry costs. The report compares adoption and deployment choices across the two groups.
By industry vertical, the report covers BFSI, retail & e-commerce, telecommunications & IT, manufacturing, healthcare & life sciences, government & defense, media, gaming & entertainment, and transportation & logistics. BFSI is expected to account for the largest share in 2026, as banks, insurers, and exchanges need instant payment handling, trading, risk control, and fraud checks, together with resilience rules such as the EU's Digital Operational Resilience Act. Retail & e-commerce is expected to grow at the highest CAGR, as online sellers use live personalization, stock tracking, and AI-based shopping features.
Regional Analysis
The report covers North America, Europe, Asia-Pacific, Latin America, and the Middle East & Africa, with analysis of 24 countries, the Nordic countries, and four rest-of-region groups. The analysis considers cloud and AI adoption, payment system development, ERP migration activity, memory manufacturing, and data sovereignty and financial regulation.North America is expected to hold the largest share of the market in 2026, as it hosts the leading cloud providers and in-memory platform vendors, has large financial services and technology sectors, and adopted AI early. Most major platform vendors, including Oracle, Microsoft, Amazon Web Services, Google, Redis, and Aerospike, are based in the U.S., as are CIM startups such as d-Matrix and EnCharge AI, and real-time payment services such as FedNow and demand for fraud detection support further investment. Europe holds a significant share as the home of SAP, with legacy ERP migrations ahead of 2027 driving demand in Germany, France, the U.K., and the Nordic countries.
Asia-Pacific is expected to grow at the highest CAGR during the forecast period, driven by fast digital adoption, huge instant payment flows, and its position as the center of global memory chip production. India's UPI handles billions of transactions a month, South Korea's Samsung and SK hynix are developing processing-in-memory products, while banks, online retailers, and telecom operators in China, Japan, Singapore, and Australia are expanding their use of in-memory platforms. In Latin America, Brazil's Pix system and digital banks in Mexico, Colombia, Chile, and Argentina support demand, while in the Middle East & Africa, AI and digital government investment in the UAE and Saudi Arabia, Israel's technology base, and mobile money in South Africa and Kenya drive growth.
Competitive Landscape
The report reviews key growth strategies, competitive benchmarking, the competitive dashboard, which groups companies into industry leaders, market differentiators, vanguards, and emerging companies, and a 2025 market share/rank analysis. Company profiles are grouped into enterprise software and cloud providers, in-memory database and data grid specialists, memory manufacturers, and compute-in-memory hardware companies, and cover business overview, financials, product portfolio, strategic developments, and SWOT analysis.The market brings together major business software and database suppliers, hyperscale clouds selling hosted in-memory services, specialist database and data grid firms, open-source communities, memory makers developing processing-in-memory, and startups building compute-in-memory AI chips. Competition centers on latency and performance, scalability and resilience, AI and vector search features, licensing openness, cloud integration, and total cost of ownership, and in hardware on performance per watt and toolchain maturity. Leading companies are building vector and AI features into their platforms, shifting to subscriptions, answering open-source forks with new licenses and hosted offerings, supporting CXL and memory tiering, and funding in-memory chip development. Recent moves include the launch of Valkey with backing from Amazon Web Services, Google Cloud, and Oracle, the AGPL option added with Redis 8 in 2025, and EnCharge AI's 2025 funding round.
Key companies profiled in the report include SAP SE (Germany), Oracle Corporation (U.S.), Microsoft Corporation (U.S.), Amazon Web Services, Inc. (U.S.), Google LLC (U.S.), IBM Corporation (U.S.), Redis Ltd. (U.S.), Aerospike, Inc. (U.S.), SingleStore, Inc. (U.S.), GridGain Systems, Inc. (U.S.), Hazelcast, Inc. (U.S.), GigaSpaces Technologies Ltd. (Israel), Samsung Electronics Co., Ltd. (South Korea), SK hynix Inc. (South Korea), d-Matrix Corporation (U.S.), EnCharge AI, Inc. (U.S.), Axelera AI B.V. (Netherlands), and Mythic, Inc. (U.S.).
How This Report Helps
- Provides reliable market size estimates and long-term forecasts for the global in-memory computing market through 2036.
- Explains how real-time data, AI workloads, and ERP modernization are expanding demand for in-memory platforms.
- Assesses the impact of rising DRAM prices, open-source licensing changes, and migration complexity.
- Identifies high-growth opportunities across component, deployment, hardware technology, application, organization size, industry vertical, and region.
- Tracks emerging technologies such as processing-in-memory, compute-in-memory accelerators, vector search, and CXL memory pooling.
- Benchmarks software vendors, cloud providers, memory makers, and hardware startups on offerings and strategy.
- Supports product, partnership, investment, and market entry decisions for vendors, cloud providers, chip makers, integrators, and investors.
Key Questions Answered
- How large is the global in-memory computing market today, and how will it grow through 2036?
- How are AI, vector search, and real-time payments changing demand for in-memory databases and caches?
- What are the main drivers, restraints, opportunities, challenges, and trends shaping the market?
- How will rising DRAM and HBM prices affect the cost of in-memory deployments?
- How has the Redis license change and the Valkey fork reshaped the open-source and managed service landscape?
- Which component, deployment, hardware technology, application, and industry vertical segments will grow the fastest?
- When will processing-in-memory and compute-in-memory hardware reach broad commercial adoption?
- Which regions and countries offer the most attractive opportunities?
- Who are the leading companies in the market, and how do software, cloud, memory, and hardware players compare?
Table of Contents
1. Introduction
2. Research Methodology
4. Market Overview
5. In-Memory Computing Market, by Component
6. In-Memory Computing Market, by Deployment
7. In-Memory Computing Market, by Hardware Technology
8. In-Memory Computing Market, by Application
9. In-Memory Computing Market, by Organization Size
10. In-Memory Computing Market, by Industry Vertical
11. In-Memory Computing Market, by Geography
12. Competitive Landscape
13. Company Profiles
14. Appendix
LIST OF TABLES
LIST OF FIGURES
Companies Mentioned
- SAP SE
- Oracle Corporation
- Microsoft Corporation
- Amazon Web Services, Inc.
- Google LLC
- IBM Corporation
- Redis Ltd.
- Aerospike, Inc.
- SingleStore, Inc.
- GridGain Systems, Inc.
- Hazelcast, Inc.
- GigaSpaces Technologies Ltd.
- Samsung Electronics Co., Ltd.
- SK hynix Inc.
- d-Matrix Corporation
- EnCharge AI, Inc.
- Axelera AI B.V.
- Mythic, Inc.
