Processing-in-Memory (PIM) Market Size & Share 2026-2035

Processing-in-Memory (PIM) Market Size

The global processing-in-memory (PIM) market was valued at USD 611.5 million in 2025. The market is expected to grow from USD 850 million in 2026 to USD 3 billion in 2031 & USD 8.5 billion in 2035, at a CAGR of 29.2% during the forecast period according to the latest report published by Global Market Insights Inc.

The growth of the market is attributed to the expansion of AI & generative AI workloads, rising deployment of HBM & advance memory architectures, growing adoption of compute express link (CXL) enabled memory pooling, increasing investment by governments in semiconductors & AI infrastructure, growing demand for energy-efficient computing, which aims at reducing data movement and enhancing system performance.

The processing-in-memory (PIM) market is significantly driven due to rapid adoption of AI/generative AI use cases for the proliferation of AI and generative AI (Genai) use cases. These AI and Genai use cases demand a significantly higher amount of memory bandwidth and low latency than traditional computing architectures. With AI models growing in their size and the intricacy of these models, memory bottlenecks have turned out to be an ardent problem that has escalated the need for PIM solutions to carry out computation at a location nearer to memory. For instance, the US department of commerce’s chips program office announced in July 2026 updated funding agreements under the chips and science act to build more US semiconductor manufacturing and supply of key components. The program expands US chip capacity, drives advanced memory, AI hardware, and next-generation computers, creating a rich PIM processing-in-memory ecosystem.^{[1]US department of commerce, nist.gov}

Additionally, the market is growing driven by increasing industry acceptance of CXL and next generation memory technologies that support memory sharing and heterogeneous computing for data center AI server. CXL allows to eliminate memory bottlenecks and increase the usage of resource for future PIM. For instance, European Commission announced investment of Euro 659 million in German state aid that will finance four first-of-a-kind semiconductor production sites. The project bolsters Europe’s semiconductor ecosystem within the EU Chips Act with investment in state-of-the-art chips, advanced memory technologies, and future AI chips designed to assist processing-in-memory (PIM) structures.^{[2]European commission, europa.eu}

The market increased steadily from USD 232.8 million in 2022 and reached USD 441.5 million in 2024, due to rise in application of artificial intelligence and high-performance computing workload where demand for high memory bandwidth and low latency increased, high capital investment in development of modern memory such as high bandwidth memory (HBM) and memory centric architectures boosted the market penetration during the period. Increase in use of AI accelerators for the deployment in data centers, increased implementation of compute express link enabled memory, and increased government spending in silicon manufacturing sector & artificial intelligence industry also contributed in penetration.

Processing-in-Memory (PIM) Market Trends

• High bandwidth memory (HBM) and novel memory architectures drive evolution of processing-in-memory for AI and high-performance computing. HBM and emerging architectures gain major traction since about 2022 as AI accelerators require higher memory bandwidth and lower latency. High deployment of HBM4 and following memory architectures is expected by 2033. This would drastically boost performance of AI training and inference, minimizing memory limitations.
• With increasing deployment of compute express link (CXL)-enabled memory expansion and pooling, memory-centric computing architecture is becoming the prevalent trend, which will continue till 2032. This accelerated deployment of CXL around 2023 with major hyperscale cloud vendors and semiconductor vendors will offer a better utilization of memory resources in the heterogeneous computing systems, and the proliferation of memory sharing, the expansion of the server memory footprint and further adoption of processing-in-memory technology will keep progressing.
• Low-power PIM chips gaining momentum, edge AI and the proliferation of smart IoT products have created new market demand for low-power processing-in-memory solutions, a growth driver that has become particularly prominent in 2021 due to the widespread adoption of smart cameras, automated systems, smart industrial equipment, and connected health devices. This growth trend in the PIM market will continue for a period up to 2031 due to increasing deployment of edge computing across the world, including in IoT devices, and will lead to a persistent demand for low-power, low-latency PIM chip solutions that can provide intelligent inference at the edge.
• Development of emerging non-volatile memory technologies continued commercial development of next-generation non-volatile memories like MRAM, ReRAM and emerging memories further propels commercial Processing-in-Memory. This process started in about 2022, when the global semiconductor market ramped up investments in memory-centric compute. The global processing-in-memory market will continue through 2034 driven by advancements in semiconductor processes and AI chips and thus offer more computational efficiency, data movement optimization, and bring broader commercial use cases of PIM to data centers, cloud data centers, and enterprise AI.

Processing-in-Memory (PIM) Market Analysis

Based on computing architecture, the processing-in-memory (PIM) market is divided into analog PIM, digital PIM and mixed-signal (AMS) PIM.

• The digital PIM segment led the market in 2025, holding a 61.4% share. The digital PIM systems contribute the largest to the PIM market because they feature greater scalability, reliability, and provide high accuracy. Also, it can be seamlessly integrated into conventional digital silicon fabrication techniques. Due to these characteristics, these systems are now found in various use cases such as cloud data centers, AI accelerators, and high-performance computing environments where reliability and CXL/HBM compatibility become significant for broad deployments.
• The analog PIM segment is anticipated to grow at a CAGR of 28.4% over the forecast period. This growth is mainly attributed to the rising need for edge and ultra-low power AI inferencing along with an increasing interest in energy-efficient neural network processing. Analog PIM can significantly lower the energy consumption and data movement, offering it the ability to be used in next generation of AI hardware, edge AI devices, self-driving cars, etc.

Based on memory technology the processing-in-memory (PIM) market is divided into DRAM-based PIM, SRAM-based PIM, MRAM-based PIM, RERAM-based PIM, flash-based PIM and others

• The DRAM-based PIM segment dominated the market in 2025 and was valued at USD 420 million due to growing adoption of dram-based PIM in AI accelerators, HPC, and hyperscale data centers can be attributed to the significant portion captured by the segment in the market. It offers high bandwidth and low latency with seamless integration of existing memory infrastructure and serves as the core for AI model training and inference of large AI models. Moreover, enhanced scalability, maturation, and high bandwidth memory (HBM) integration continue to boost the demand from cloud and enterprise computing.
• The MRAM-based PIM segment is expected to witness growth at a CAGR of 35.4% during the forecast period as companies are investing in developing non-volatile memory solutions for energy-efficient AI computation. The sector is driven by high-speed operation with low power consumption, non-volatility, and excellent endurance properties of MRAM. Its adoption for edge AI applications, automotive electronics, industrial automation, and IoT solutions is driving adoption. Technological advancements in spintronic memories and embedded AI processors are leading to the faster commercialization of mram-based PIM, thereby facilitating faster inference with drastically reduced energy consumption.

Based on application, the processing-in-memory (PIM) market is divided into AI & ML, HPC & scientific computing, data analytics & database processing, computer vision & image processing, NLP and others.

• The AI & ML segment dominated the market in 2025 and was valued at USD 318 million, due to the dramatic proliferation of generative AI, large language models (LLMs) and deep learning workloads. PIM technology slashes data movement between the processing unit and memory and provides significant improvements in inference latency, memory bandwidth and energy efficiency, allowing a greater number of AI accelerators to be deployed within hyperscale data centers, cloud services, and enterprise AI infrastructure.
• The NLP segment is expected to witness growth at a CAGR of 35.6% during the forecast period, fueled by the accelerated adoption of generative AI assistants, LLMs, smart chatbots, and live language translation solutions which demand energy-efficient and memory-rich computation at very low latencies. As transformer models used in various AI & ML use cases have become increasingly complex, PIM solutions can deliver major improvements in inference performance, scalability, and power consumption, fueling increased adoption by companies across sectors including healthcare, financial services, retail and enterprise.

North America Processing-in-Memory (PIM) Market

North America held a share of 41.7% of market in 2025.

• In North America, the market is growing due to rapid growth of artificial intelligence (AI), hyperscale cloud deployments, high-performance computing (HPC), as well as significant investments by leading technology firms in the adoption of AI accelerators, new memory technologies and advanced chip design techniques which require the processing-in-memory approach for improved memory bandwidth, lower latency, as well as energy efficiency.
• Domestic chip production is gaining momentum through initiatives like the CHIPS and science Act, while government and industry are scaling up investments in semiconductor facilities, AI capabilities, and advanced memory and packaging technologies. The region is projected to continue as a major center of innovation for artificial intelligence and high-performance computing, spurred by increasing deployments of cloud data centers, enterprise AI, and advanced AI models.

The U.S. processing-in-memory (PIM) market was valued at USD 89.5 million and USD 122.8 million in 2022 and 2023, respectively. The market size reached USD 234 million in 2025, growing from USD 169.2 million in 2024.

• The growth of the market in the U.S. is particularly strong due to increasing investments in the AI hardware infrastructure, leading semiconductor development and memory innovation. Government funded projects, such as the CHIPS and science act, alongside large investments from established semiconductor manufacturers, are powering the domestic manufacturing of state-of-the-art AI chipsets and memory devices which is driving the adoption of processing-in-memory solutions.
• For instance, across-agency programs funded by the U.S. national science foundation (NSF) are fueling ongoing research in areas such as energy-efficient computing, future semiconductor technology, AI hardware and computing architectures for the future. These projects target a broad range of research goals, from power efficiency and novel memory structures to novel hardware that will enable next generation artificial intelligence capabilities and will contribute to PIM’s research potential.^{[3]US NSF, nsf.gov}

Europe Processing-in-Memory (PIM) Market

Europe market accounted for USD 69.1 million in 2025 and is anticipated to show lucrative growth over the forecast period.

• Europe market is expanding due to increasing investments in semiconductor manufacturing, artificial intelligence, and high-performance computing under the European Chips Act and digital Europe programme. These initiatives are strengthening Europe's semiconductor ecosystem while accelerating the adoption of advanced memory technologies and AI accelerators, creating favorable conditions for processing-in-memory deployment.
• The EU is supporting next-generation memory architectures in areas such as energy-efficient computing, cloud infrastructure and AI sovereignty. Ongoing partnerships among semiconductors companies, research bodies and EU-backed chip programs are paving the way for widespread processing-in-memory technology adoption.

Germany dominates the Europe market, showcasing strong growth potential.

• The processing-in-memory market is dominated by Germany due to the presence of Germany robust semiconductor manufacturing infrastructure, automotive electronics ecosystem, substantial investment in AI hardware and semiconductor R&D. Furthermore, Germany advantages include presence of significant chip makers and academic research institutions, in addition to the backing by governmental semiconductor roadmap that expedites the integration of modern memory hardware.
• For instance, German federal government approved microelectronics strategy to enhance Germany position on Europe leading semiconductor hub, microelectronics strategy emphasizes further research on semiconductors, leading-edge manufacturing processes, development of skilled workforce and cooperation within the European Chips Act, fostering innovative progress in sophisticated memory applications, processors for AI and processing-in-memory solutions.^{[4]Germany federal govt, bundes.de}

Asia Pacific Processing-in-Memory (PIM) Market

The Asia Pacific market is anticipated to grow at the highest CAGR of 30.7% during the forecast period.

• The Asia Pacific Market is witnessing rapid growth due to rising semiconductor ecosystems, rapid adoption of artificial intelligence, and the increase of investments in high-performance computing and data center infrastructure are paving the way for the Asia Pacific processing-in-memory market growth. South Korea, China, Taiwan, and Japan are currently leading the adoption of innovative memory technologies as well as the design of new AI accelerators and the development of processing-in-memory devices in the region.
• Increased government incentives for semiconductor self-sufficiency, AI breakthroughs, and enhanced chip fabrication technologies will continue to foster the rapid expansion of the processing-in-memory solutions market in Asia Pacific with significant investments pouring in from companies across HBM production, advanced packaging, and next generation computing technologies as well as for the proliferation of edge AI hardware.

India processing-in-memory (PIM) market is estimated to grow with a significant CAGR, in the Asia Pacific market.

• The government push for investments in semiconductors and AI infrastructure along with digital transformation activities has led India to become a prominent market for processing-in-memory (PIM). India semiconductor mission (ISM) and the semicon India program among others are bolstering India’s chip-making capabilities and are driving investments in high-performance chip design and manufacturing.
• For instance, the India semiconductor mission (ISM) is further strengthening India's semiconductor ecosystem by approving a number of fabrications, OSAT, compound semiconductor, and display manufacturing projects in several states. These projects will create India’s capabilities in advanced semiconductor manufacturing and packaging, setting up India's future in a wide range of memory intensive technologies like processing-in-memory.^{[5]Indian semiconductor mission, ism.gov}

Middle East and Africa Processing-in-Memory (PIM) Market

Saudi Arabia market to experience substantial growth in the Middle East and Africa.

• The growth of market is being fueled by the accelerating growth of AI, cloud computing and digital infrastructure driven by its vision 2030. Investments in advanced data centers, AI platforms and semiconductor technologies facilitated by the government-led initiative are laying a firm ground for the adoption of energy-efficient computing architectures like processing-in-memory.
• Furthermore, Saudi Arabia strategic push to establish itself as the regional hub of technological innovations is driving investments into AI, hyperscale cloud computing and digital innovation. Growing number of strategic partnerships with global semiconductor firms and tech vendors, and rising demand for AI-enabled applications across government, healthcare, banking and other smart city initiatives are bolstering the adoption of PIM solutions in Saudi Arabia.

Processing-in-Memory (PIM) Market Share

The market is led by players such as SK Hynix, Samsung Electronics, Syntiant, Mythic AI and d-Matrix. These five companies cumulatively accounted for 53% market share in 2025. Their prominent market positions are backed by advanced memory technology leadership, development of innovative AI accelerators, and in-memory computing architectures. Furthermore, their extensive investment in high bandwidth memory (HBM), analog/digital PIM solutions, and next-generation semiconductors assists them to tap into the growing needs of power-efficient AI applications and high-performance computing (HPC) environments.

Through ongoing innovation in AI hardware, advanced packaging, and memory-centric computing solutions, these companies maintain their strong competitiveness. Joint ventures and alliances with hyperscale cloud platforms, leading semiconductor foundries, and research institutions, in addition to their substantial capital investments in generative AI infrastructure, edge AI capabilities, and high-performance computing applications contribute to their success in garnering demand across data centers, enterprise AI, automotive and edge computing spaces.

Processing-in-Memory (PIM) Market Companies

Prominent players operating in the market are as mentioned below:

• Axelera AI
• d-Matrix
• EnCharge AI
• GSI Technology
• HOUMO.AI
• MemryX
• Micron Technology
• Mythic AI
• PIMIC
• Pimchip (Beijing Pimchip Technology)
• Samsung Electronics
• SK hynix
• Syntiant
• TetraMem
• Witmem (Zhicun Technology)
• XCENA

• SK hynix

SK hynix is a global innovator in advanced memory technology, particularly in the field of PIM. With its unique AI-optimized PIM technology embedded with high bandwidth memory (HBM) to optimize the highest levels of bandwidth, the lowest latency, and outstanding power efficiency in data centers and high-performance computing.

• Samsung Electronics

Samsung Electronics designs innovative PIM solutions which bring AI Processing into the DRAM and HBM Chip by embedding AI processing functionalities into DRAM and high bandwidth memory (HBM) and providing advanced PIM functionalities for boosting the performance of AI inference, minimizing power consumption, and alleviating data movement bottleneck.

• Syntiant

Syntiant is creating ultra-low power AI inference processors for the edge that run on memory efficient designs to provide real-time intelligence, at a fraction of power, powering a range of always-on AI applications like voice recognition, wearables, the IoT and automotive use cases, enabling a new era of memory-efficient computing at the edge.

• Mythic AI

Mythic AI aims to revolutionize artificial intelligence, building an analog processing-in-memory (PIM) technology to integrate artificial intelligence computation directly into the flash memory arrays to provide power efficiency and extremely low latency. Mythic AI designs its chips and systems for a variety of industries and use cases, including computer vision, robotics, industrial automation, and edge AI – applications for which low power consumption and performance are essential.

• d-Matrix

D-Matrix designs and builds digital in-memory computing platforms focused on generative AI and LLM inference. The company’s infrastructure reduces the transfer of data between processor and memory, yielding higher throughput and more power efficient infrastructure in the data centers of AI-first businesses.

Processing-in-Memory (PIM) Industry News

• In May 2026, Syntiant unveiled its new advanced neural-processing silicon device for always-on AI with ultra-low-power performance at the edge of networks to realize on-device AI inference for a new generation of IoT, automotive, and wearable solutions powered by a highly memory-efficient approach to computing.
• In March 2026, SK hynix has completed development and started mass production of their HBM4, a 12-layer AI semiconductor memory that realizes substantially improved performance. The new HBM4 solution supports new generation AI servers and PIM with a view to realize higher bandwidth, enhanced power efficiency, and faster memory accessibility while removing bottlenecks in data transfer.
• In February 2026, Samsung electronics announced its new AI memory portfolio including HBM4 and other advanced AI and HPC memory solutions designed to provide improved AI inference speed and lower power consumption, which was unveiled at the international solid-state circuits conference (ISSCC) 2026. The company’s lineup features PIM technology that combines memory and processing for enhanced efficiency.

The processing-in-memory (PIM) market research report includes in-depth coverage of the industry with estimates and forecast in terms of revenue (USD Million) from 2022 – 2035 for the following segments:

Click here to Buy Section of this Report

Market, By Memory Technology

• DRAM-based PIM
• SRAM-based PIM
• MRAM-based PIM
• ReRAM-based PIM
• Flash-based PIM
• Others

Market, By Computing Architecture

• Analog PIM
• Digital PIM
• Mixed-signal (AMS) PIM

Market, By Application

• AI & ML
• HPC & scientific computing
• Data analytics & database processing
• Computer vision & image processing
• NLP
• Others

Market, By End-User Industry

• IT & cloud services
• Consumer electronics
• Automotive
• Telecommunications
• Industrial
• Healthcare
• Aerospace & Defense
• Others

The above information is provided for the following regions and countries:

• North America
  • U.S.
  • Canada
• Europe
  • Germany
  • UK
  • France
  • Spain
  • Italy
• Asia Pacific
  • China
  • India
  • Japan
  • Australia
  • South Korea

• Latin America
  • Brazil
  • Mexico
  • Argentina

• Middle East and Africa
  • South Africa
  • Saudi Arabia
  • UAE