Skip to content

  • LPDDR efficiency in a modular server form factor
    LPDDR efficiency in a modular server form factor
    LPDDR efficiency in a modular server form factor

    SAMSUNG SOCAMM2 LPDDR efficiency in a modular server form factor

Next-generation LPDDR server module
Next-generation LPDDR server module
Next-generation LPDDR server module

As data centers scale to support increasingly complex AI workloads, they require memory that delivers high performance without compromising power efficiency. SOCAMM21) is Samsung’s next-generation LPDDR5X-based server module that brings LPDDR-class power efficiency into a modular form factor purpose-built for AI infrastructure. By combining high bandwidth and low power in a compact, detachable design, SOCAMM2 helps data centers increase system density while improving cooling efficiency and overall total cost of ownership (TCO).

1) Small Outline Compression Attached Memory Module As data centers scale to support increasingly complex AI workloads, they require memory that delivers high performance without compromising power efficiency. SOCAMM21) is Samsung’s next-generation LPDDR5X-based server module that brings LPDDR-class power efficiency into a modular form factor purpose-built for AI infrastructure. By combining high bandwidth and low power in a compact, detachable design, SOCAMM2 helps data centers increase system density while improving cooling efficiency and overall total cost of ownership (TCO).

1) Small Outline Compression Attached Memory Module As data centers scale to support increasingly complex AI workloads, they require memory that delivers high performance without compromising power efficiency. SOCAMM21) is Samsung’s next-generation LPDDR5X-based server module that brings LPDDR-class power efficiency into a modular form factor purpose-built for AI infrastructure. By combining high bandwidth and low power in a compact, detachable design, SOCAMM2 helps data centers increase system density while improving cooling efficiency and overall total cost of ownership (TCO).

1) Small Outline Compression Attached Memory Module

Power efficiency for denser AI servers Power efficiency for denser AI servers Power efficiency for denser AI servers

Built on LPDDR5X, SOCAMM2 delivers significantly lower memory power consumption compared with conventional DDR-based server memory solutions2), offering more than 70% better power efficiency without compromising performance. This reduction in power consumption and heat simplifies thermal management, helping data centers maintain stable operating temperatures while easing cooling requirements — a critical consideration in high-density AI infrastructure. As a result, SOCAMM2 enables more AI servers to operate within the same power budget, improving overall energy efficiency and lowering TCO.


2) Compared with conventional DDR5 based RDIMM modules.

Built on LPDDR5X, SOCAMM2 delivers significantly lower memory power consumption compared with conventional DDR-based server memory solutions2), offering more than 70% better power efficiency without compromising performance. This reduction in power consumption and heat simplifies thermal management, helping data centers maintain stable operating temperatures while easing cooling requirements — a critical consideration in high-density AI infrastructure. As a result, SOCAMM2 enables more AI servers to operate within the same power budget, improving overall energy efficiency and lowering TCO.


2) Compared with conventional DDR5 based RDIMM modules.

Built on LPDDR5X, SOCAMM2 delivers significantly lower memory power consumption compared with conventional DDR-based server memory solutions2), offering more than 70% better power efficiency without compromising performance. This reduction in power consumption and heat simplifies thermal management, helping data centers maintain stable operating temperatures while easing cooling requirements — a critical consideration in high-density AI infrastructure. As a result, SOCAMM2 enables more AI servers to operate within the same power budget, improving overall energy efficiency and lowering TCO.


2) Compared with conventional DDR5 based RDIMM modules.

SAMSUNG SOCAMM2 More than 70% better power efficiency
Boosting memory bandwidth for AI workloads Boosting memory bandwidth for AI workloads Boosting memory bandwidth for AI workloads

Beyond its power-efficiency advantages, SOCAMM2 is also engineered to deliver the high memory bandwidth required by AI accelerators. Each module provides up to 153.6 GB/s of bandwidth — up to 2.6× higher than DDR-based server memory3) — helping reduce memory bottlenecks and keep GPUs more fully utilized. This additional throughput allows more AI tasks to run in parallel on the same server infrastructure, improving responsiveness for real-time and latency-sensitive inference services.


3) Compared with conventional DDR5 based RDIMM modules.

Beyond its power-efficiency advantages, SOCAMM2 is also engineered to deliver the high memory bandwidth required by AI accelerators. Each module provides up to 153.6 GB/s of bandwidth — up to 2.6× higher than DDR-based server memory3) — helping reduce memory bottlenecks and keep GPUs more fully utilized. This additional throughput allows more AI tasks to run in parallel on the same server infrastructure, improving responsiveness for real-time and latency-sensitive inference services.


3) Compared with conventional DDR5 based RDIMM modules.

Beyond its power-efficiency advantages, SOCAMM2 is also engineered to deliver the high memory bandwidth required by AI accelerators. Each module provides up to 153.6 GB/s of bandwidth — up to 2.6× higher than DDR-based server memory3) — helping reduce memory bottlenecks and keep GPUs more fully utilized. This additional throughput allows more AI tasks to run in parallel on the same server infrastructure, improving responsiveness for real-time and latency-sensitive inference services.


3) Compared with conventional DDR5 based RDIMM modules.

SAMSUNG SOCAMM2 bandwidth up to 2.6× higher than DDR-based server memory
Improved cooling with a horizontal layout Improved cooling with a horizontal layout Improved cooling with a horizontal layout

SOCAMM2 is installed in a horizontal orientation, unlike conventional vertically mounted modules. This layout improves system-level space utilization around CPUs and AI accelerators, enabling more flexible heatsink sizing and placement as well as cleaner front-to-back airflow design. Combined with SOCAMM2’s low-power characteristics, this horizontal design enhances cooling efficiency and increases thermal headroom in high-density AI servers while remaining compatible with both air- and liquid-cooling systems.

SOCAMM2 is installed in a horizontal orientation, unlike conventional vertically mounted modules. This layout improves system-level space utilization around CPUs and AI accelerators, enabling more flexible heatsink sizing and placement as well as cleaner front-to-back airflow design. Combined with SOCAMM2’s low-power characteristics, this horizontal design enhances cooling efficiency and increases thermal headroom in high-density AI servers while remaining compatible with both air- and liquid-cooling systems.

SOCAMM2 is installed in a horizontal orientation, unlike conventional vertically mounted modules. This layout improves system-level space utilization around CPUs and AI accelerators, enabling more flexible heatsink sizing and placement as well as cleaner front-to-back airflow design. Combined with SOCAMM2’s low-power characteristics, this horizontal design enhances cooling efficiency and increases thermal headroom in high-density AI servers while remaining compatible with both air- and liquid-cooling systems.

SAMSUNG SOCAMM2 cooling efficiency
Modular by design, easier to manage Modular by design, easier to manage Modular by design, easier to manage

SOCAMM2 adopts a detachable modular design, unlike traditional soldered LPDDR implementations. This approach allows data centers to scale memory capacity and performance over time by upgrading or replacing modules without modifying the mainboard,  thereby simplifying lifecycle management and day-to-day system maintenance. By reducing the effort and disruption associated with memory service operations, SOCAMM2 helps minimize server downtime and extend the useful life of existing server platforms — bringing LPDDR-class efficiency into a compact, modular form factor designed to evolve alongside AI workloads.

SOCAMM2 adopts a detachable modular design, unlike traditional soldered LPDDR implementations. This approach allows data centers to scale memory capacity and performance over time by upgrading or replacing modules without modifying the mainboard,  thereby simplifying lifecycle management and day-to-day system maintenance. By reducing the effort and disruption associated with memory service operations, SOCAMM2 helps minimize server downtime and extend the useful life of existing server platforms — bringing LPDDR-class efficiency into a compact, modular form factor designed to evolve alongside AI workloads.

SOCAMM2 adopts a detachable modular design, unlike traditional soldered LPDDR implementations. This approach allows data centers to scale memory capacity and performance over time by upgrading or replacing modules without modifying the mainboard,  thereby simplifying lifecycle management and day-to-day system maintenance. By reducing the effort and disruption associated with memory service operations, SOCAMM2 helps minimize server downtime and extend the useful life of existing server platforms — bringing LPDDR-class efficiency into a compact, modular form factor designed to evolve alongside AI workloads.

SAMSUNG SOCAMM2 a detachable modular design

Applications for SOCAMM2

Explore more about SOCAMM2

* The contents of this page are provided for informational purposes only. No representation or warranty (whether express or implied) is made by Samsung or any of its officers, advisers, agents, or employees as to the accuracy, reasonableness or completeness of the information, statements, opinions, or matters contained in this page, and they are provided on an "AS-IS" basis. Samsung will not be responsible for any damages arising out of the use of, or otherwise relating to, the contents of this page. Nothing in this page grants you any license or rights in or to information, materials, or contents provided in this document, or any other intellectual property.
* The contents of this page may also include forward-looking statements. Forward-looking statements are not guarantees of future performance and that the actual developments of Samsung, the market, or the industry in which Samsung operates may differ materially from those made or suggested by the forward-looking statements contained in this page.
* All design, features and specifications represented herein may change without notice.
* Images shown here have been adjusted for demonstration purposes and may appear differently on the actual products.
* All data on products herein, including their performances, are based on internal testing using standard Samsung benchmarks under laboratory conditions.
* Test results do not guarantee future performance under such test conditions, and the actual throughput or performance that any user will experience may vary depending upon many factors.
* For further details on product specifications, please contact the sales representative of your region.

https://semiconductor.samsung.com/support/contact-info/global-network/