Home > Who Is Leading the DRAM Patent Race? Who Is Leading the DRAM Patent Race? Discover the top 30 DRAM patent owners AI INFRASTRUCTURE & SEP LICENSING DRAM PATENT TRENDS Who is leading? WHY THIS MATTERS TO YOU DOWNLOAD THE REPORT & RANKINGS Dynamic Random Access Memory (DRAM) is becoming one of the most strategic technologies in AI infrastructure Artificial intelligence is transforming the semiconductor industry, and memory has become one of its most strategic technologies.Based on our proprietary patent intelligence, our latest report identifies the organizations leading the global DRAM patent landscape, reveals the patent trends shaping next-generation AI infrastructure and explores why DRAM is emerging as one of the next major SEP licensing markets. The analysis behind the rankings 4 DRAM standards analyzed: DDR, LPDDR, GDDR and HBM 30 Top Ultimate Owners ranked 2 Complimentary ranking methodologies More than 7,500 Active, granted, patent families analyzed Why AI infrastructure matters for SEP licensing While public attention around AI focuses on large language models (LLMs) and AI accelerators, an equally important transformation is taking place across the infrastructure that powers them. Modern AI infrastructure depends on much more than GPUs alone. Training and deploying today’s AI models requires high-performance memory, chip interconnects, networking and storage technologies working together at scale. Historically, AI deployments relied heavily on proprietary ecosystems, with Nvidia being the most prominent example. Increasingly, however, hyperscalers and other technology companies are driving the adoption of open standards to improve interoperability and reduce vendor lock-in. As a result, standards are emerging across every layer of the AI infrastructure stack—from DRAM memory (DDR, LPDDR, GDDR and HBM) to chip interconnects (CXL, UALink and UCIe), networking (Ultra Ethernet and InfiniBand) and NAND storage. This transition is laying the foundation for one of the next major SEP licensing markets—with DRAM emerging as one of its most strategically important technologies. Layer Purpose Key Technologies / Components Examples of Key Standards Compute Performs AI training and inference workloads GPUs, CPUs, AI accelerators PCIe, OCP MGMT, ACPI Memory (DRAM) Provides high-bandwidth, low-latency memory DDR, LPDDR, GDDR, HBM DDR5, LPDDR5/5X, GDDR6/7, HBM2e, HBM3, HBM3E Interconnects (Chip / System) Connects chips and components within and across systems Die-to-die, chip-to-chip, board-to-board interconnects CXL, UALink, UCIe Networking Enables communication across servers and clusters Ethernet switches, NICs, fabrics Ultra Ethernet, InfiniBand, RoCEv2, NVMe-oF Storage (NAND) Provides persistent data storage NAND flash, SSDs, NVMe storage NVMe, NVMe-oF, ONFI, Toggle, Open Channel SSD Optics Enables high-speed optical communication Optical transceivers, modules, cables IEEE 802.3, OIF, Coherent Optics, OpenROADM Power & Cooling Delivers power and manages thermal performance PSUs, power distribution, liquid/air cooling OCP, Open19, ASHRAE, UEC Manufacturing & Packaging Enables semiconductor manufacturing and advanced packaging Foundry, advanced packaging, substrates, test SEMI Standards, EDA Interface Standards, JEDEC Packaging, ITF Discover who is shaping the future of DRAM innovation AI infrastructure is emerging as a new SEP licensing market All of the key standards underpinning AI infrastructure are covered by thousands of standard essential patents (SEPs), creating a substantial foundation for future SEP licensing activity. However, the industry has yet to establish a dominant SEP licensing model. Potential approaches range from component- and device-level licensing to licensing at the system, hyperscaler, and discussions about the royalty rate base to consider the chip level, the usage or even tokens. Given the scale of AI infrastructure deployments and the broad range of technologies involved, the resulting SEP licensing market could ultimately rival, or even exceed, the size of traditional cellular, wireless or video SEP licensing. The AI SEP ecosystem is beginning to take shape Recent market developments suggest that the SEP ecosystem for AI infrastructure is already beginning to take shape. In late 2025 and early 2026 Via Licensing Alliance (Via LA) launched patent calls for new DRAM and NAND licensing programs, while earlier this year Sisvel initiated a patent call for an Ultra Ethernet patent pool. These initiatives signal growing industry recognition that standardized AI infrastructure technologies will require scalable and efficient SEP licensing mechanisms. As adoption accelerates, patent pools are likely to play an increasingly important role in reducing transaction costs, mitigating litigation risk, and facilitating access to SEPs across the AI value chain. DRAM creates new licensing opportunities—and new legal challenges Unlike cellular standards, which evolved over decades, AI infrastructure standards are adopted at unprecedented speed. This may create huge legal challenges but at the same time a unique opportunity for licensors, implementers, and patent pools to build a more efficient and coordinated licensing ecosystem from the outset, potentially avoiding some of the fragmentation and complexity that characterized earlier generations of SEP licensing. Understanding where innovation is concentrated provides valuable insight into the future direction of AI infrastructure—and the organizations most likely to influence future standards and licensing. The four DRAM technologies shaping the AI infrastructure stack While DRAM has long been used across consumer electronics, communications equipment, automotive systems, and enterprise computing, its role has expanded significantly with the rise of AI data centers. The analysis focuses on four major DRAM technology families that underpin modern computing and increasingly support AI infrastructure at scale. DDR– Double Data Rate The dominant memory technology for PCs, servers, networking equipment and cloud infrastructure. LPDDR– Low Power Double Data Rate Low-power memory optimized for mobile, embedded and power-efficient computing environments. GDDR– Graphics Double Data Rate High-speed graphics memory designed for graphics-intensive workloads and AI accelerators. HBM– High Bandwidth Memory High-bandwidth memory built for advanced AI, high-performance computing and data-center workloads. Identifying DRAM Patents To identify patent families potentially relevant to DRAM standards, LexisNexis® Intellectual Property Solutions applied its AI-based LexisNexis® Classification methodology to detect patents associated with JEDEC DRAM standards, including DDR, LPDDR, GDDR, and HBM. The classification model is trained using both true positive and true negative patents. True positive examples include patents that have been independently claim-charted against DRAM standards and are considered known or likely SEPs. True negative examples include memory-related patents that are relevant to DRAM technology but are not essential to DRAM standards. This training approach enables the AI model to identify patents relevant to DRAM standards using an inclusive methodology that captures potential SEPs while also encompassing related non-SEPs. To continuously improve accuracy, random patent samples from the identified technology landscape are reviewed by subject matter experts. Patents confirmed to be essential to DRAM standards are added to the positive training set, while unrelated patents are incorporated into the negative training set. This iterative human-in-the-loop process refines the model and improves classification performance over time. The resulting dataset represents a landscape of DRAM “SEP candidates” also known as undeclared patents, patent families that are potentially essential to one or more JEDEC DRAM standards. SEP candidate patent families by DRAM standard and year The DRAM patent landscape continues to expand across all four technology families, reflecting sustained innovation in performance, efficiency and system integration. DDR remains the largest cumulative landscape. As HBM has become a key enabler of large-scale AI training and inference, providing the bandwidth required by increasingly powerful GPUs and AI chips, it is no surprise that it emerged as the fastest-growing technology. Starting from the smallest patent landscape among the major DRAM standards in 2016, HBM has grown to become the second-largest area of patenting activity in 2025 and 2026, surpassing both LPDDR and GDDR. This trend reflects the increasing importance of high-bandwidth memory in AI accelerators, high-performance computing, and next-generation data center infrastructure. How closely related are today’s DRAM standards? Although DDR, LPDDR, GDDR and HBM serve different markets, many of the underlying innovations are shared across multiple standards. Our analysis reveals where technologies converge, where they diverge and why HBM is emerging as a distinct area of innovation for AI infrastructure. The complete report includes the full overlap analysis together with the underlying data and strategic interpretation. DOWNLOAD THE FULL REPORT NOW Who is leading the DRAM patent race? To identify the leading ultimate owners of DRAM-related patent families, we ranked companies based on their active granted patent portfolios across the four major DRAM technology categories. The Top 30 ranking, accessible here, is based on active INPADOC (International Patent Documentation) patent families with at least one granted patent worldwide. Each company’s overall position is determined by its average ranking across the four DRAM standards. Top 5 ranks of DRAM Ultimate Owners by patent families and standard RankUltimate OwnerHQDDR Patent Families (active and granted)LPDDR Patent Families (active and granted)GDDR Patent Families (active and granted)HBM Patent Families (active and granted)1SK HynixKR11112SamsungKR22223MicronUS33334CXMTCN44455IdeahubKR8558 The ranking highlights a highly concentrated ownership structure. The three leading DRAM innovators, SK Hynix, Samsung, and Micron, rank top 3 across all four DRAM technology categories. Beyond the top five, the ranking includes a mix of DRAM manufacturers, semiconductor companies, system vendors, and a notable number of patent licensing entities. The rankings change when patent strength is taken into account The largest patent portfolio isn’t always the strongest. While patent family counts provide insight into the scale of a company’s DRAM patent portfolio, they do not capture differences in patent quality and competitive significance. To address this, the ranking was also analyzed using the LexisNexis Patent Asset Index, a metric designed to measure innovative strength rather than portfolio size alone. The result is a different perspective on technology leadership—one that highlights organizations whose patents have exceptional strategic value. Samsung moves into the top position, ahead of SK Hynix and Micron, indicating that its DRAM portfolio contains particularly high-value patent assets. SK Hynix and Micron remain strong in second and third place, demonstrating that the industry’s three largest memory suppliers not only own the largest portfolios but also control the most valuable DRAM patent assets. Top 5 ranks of DRAM Ultimate Owners by Patent Asset Index and standard RankUltimate OwnerHQDDR Patent Asset Index (active and granted)LPDDR Patent Asset Index (active and granted)GDDR Patent Asset Index (active and granted)HBM Patent Asset Index (active and granted)1SamsungKR11112SK HynixKR22233MicronUS33324CXMTCN44475IntelUS5555 Access the complete Top 30 ranking and report to: Explore which organizations beyond the top 5 dominate the market Discover how concentrated the DRAM patent landscape really is Digest the differences between quantity and quality See how organizations portfolios are split across the DRAM standards Learn where companies such as Intel, AMD, Qualcomm, Nvidia, Apple, MediaTek, TSMC, and Huawei rank among the world’s leading DRAM patent owners. Get the full report and ranking here The DRAM licensing ecosystem is already taking shape Notably, several of the top-ranked patent owners are patent assertion and licensing companies rather than active memory product manufacturers. Examples include Ideahub, Vector Capital, ND Lodestar Licensing Group. The strong presence of these organizations highlights the long history of patent transactions, portfolio acquisitions, and licensing activity in the DRAM sector. It also suggests that a significant portion of potentially relevant DRAM patents is already controlled by entities whose primary business model centers on intellectual property monetization rather than product sales. Download the full report to access the complete Top 15 Patent Assertion Entity (PAE) ranking and discover which organizations are shaping the DRAM licensing landscape. Why these rankings matter to you The DRAM patent landscape is becoming increasingly relevant for organizations across the AI infrastructure ecosystem. Understanding who owns the most significant patent portfolios can support licensing strategies, competitive intelligence, and investment decisions. Patent owners and licensors Benchmark portfolios, identify licensing opportunities, and evaluate competitive positioning. Implementers and semiconductor companies Understand licensing exposure and monitor technology leadership. Hyperscalers and AI infrastructure providers Gain insight into the technologies shaping next-generation AI infrastructure. Law firms and IP advisors Support licensing negotiations, litigation strategies, and portfolio valuation. Investors and market analysts Track innovation leadership and emerging market dynamics. Access the complete DRAM report and patent rankings The downloadable rankings expand on the insights presented here and provide a comprehensive view of the organizations shaping the future of DRAM innovation. Inside you’ll find: Complete Top 30 Ultimate Owner rankings by quantity Complete Top 30 Ultimate Owner rankings by Patent Asset Index The concentration of the DRAM landscape The organization’s portfolios split across the DRAM standards Top 15 DRAM Patent Assertion Entity (PAE) ranking Litigation activity in DRAM and the top 20 DRAM plaintiffs The detailed ranking methodology The tools and datasets underlying the analysis and rankings Whether you are benchmarking competitors, evaluating licensing opportunities, consider joining a patent pool, or monitoring AI infrastructure patent ownership, the DRAM Patent Rankings provide the context behind the data and the insight needed to make informed business decisions. Find out more about Wi-Fi and SEP analytics Bringing Clarity to the Wi-Fi Technology Landscape A standards and technology overview that includes Wi-Fi generations and points visitors toward exploring Wi-Fi 7 in LexisNexis® IPlytics™. Learn more LexisNexis® IPlytics™ Learn how LexisNexis® IPlytics™ supports analysis of SEPs, technical standards, and standards contributions, which aligns with the benchmarking and transparency needs highlighted in the Wi-Fi 7 story. Learn more Who Is Leading the 5G Patent Race? Benchmark your portfolio against SEP leaders in our 2026 5G report that reveals insights to support smarter licensing decisions Learn more Licensing and monetization with LexisNexis® PatentSight+™ Evaluate patent portfolio potential and identify patents that are most attractive for licensing using LexisNexis® PatentSight+™ . Learn more
