Wednesday, June 4th 2025

Micron Ships World's First 1γ (1-Gamma)-Based LPDDR5X

Micron Technology, Inc. (Nasdaq: MU), announced today that it is shipping qualification samples of the world's first 1γ (1-gamma) node-based low-power double data rate 5X (LPDDR5X) memory, designed to accelerate AI applications on flagship smartphones. Delivering the industry's fastest LPDDR5X speed grade of 10.7 gigabits per second (Gbps), combined with up to a 20% power savings, Micron LPDDR5X transforms smartphones with faster, smoother mobile experiences and longer battery life - even when executing data-intensive workloads such as AI-powered translation or image generation.

To meet the industry's increasing demand for compact solutions for next-generation smartphone designs, Micron's engineers have shrunk the LPDDR5X package size to offer the industry's thinnest package of 0.61 millimeters, making it 6% thinner compared to competitive offerings, and representing a 14% height reduction from the previous generation. The small form factor unlocks more possibilities for smartphone manufacturers to design ultrathin or foldable smartphones.
"Micron's 1-gamma node-based LPDDR5X memory is a game-changer for the mobile industry," said Mark Montierth, corporate vice president and general manager of Micron's Mobile and Client Business Unit. "This breakthrough technology delivers lightning-fast speeds and remarkable power efficiency - all within the industry's thinnest LPDDR5X package - paving the way for exciting new smartphone designs. This solution demonstrates our commitment to empowering the ecosystem to create extraordinary mobile experiences."

The company's 1γ-based LPDDR5X enables dramatic leaps in performance for mobile users by enabling faster AI insights. For example, Micron evaluated mobile AI response times from large language model Llama 2, based on 1γ LPDDR5X's 10.7 Gbps bandwidth compared to 1β (1-beta) LPDDR5X's 7.5 Gbps bandwidth, finding:
  • Responses are 30% faster when asking for location-based restaurant recommendations.
  • Results are more than 50% faster when translating a voice inquiry in English to text in Spanish to ask for directions.
  • Responses can be up to 25% faster when requesting car purchase recommendations based on vehicle type, affordability and certain infotainment and safety features.
Now ramping in Micron's mobile portfolio, Micron's 1γ-based LPDDR5X is the company's first mobile solution to leverage advanced EUV lithography - providing customers with early access to the latest performance and power efficiency advancements, based on the industry's most advanced memory node technology. This milestone builds on Micron's February sampling of 1γ-based DDR5 memory for next-generation CPUs in the data center and client segments. Micron's optimized 1γ DRAM node leverages CMOS advancements like next-generation high-K metal gate technology for improved transistor performance and incorporates leading-edge EUV lithography for enhanced bit density.

As energy-intensive mobile AI workloads are increasingly processed on-device rather than only in the cloud, low-power chips are crucial for devices like smartphones, tablets and laptops, which need to conserve power while performing AI computations.

Micron's 1γ-based LPDDR5X's significant 20% power savings will allow mobile users to enjoy their favorite AI applications, games and video content longer on a single charge. In addition, as AI intensifies the need for powerful, energy-efficient compute, data center servers, intelligent vehicles and AI PCs may also increasingly adopt LPDDR5X for its unique blend of optimized power efficiency and high performance.

Micron is currently sampling 1γ-based LPDDR5X 16 gigabyte (GB) products to select partners and will offer a wide range of capacities from 8 GB to 32 GB for use in 2026 flagship smartphones.
Source: Micron
Add your own comment

4 Comments on Micron Ships World's First 1γ (1-Gamma)-Based LPDDR5X

#1
kondamin
nice they got it thinner for the hbm stack, but why the bullshit AI metrics and not proper numbers.
Also 10nm++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Posted on Reply
#2
phints
kondaminnice they got it thinner for the hbm stack, but why the bullshit AI metrics and not proper numbers.
Also 10nm++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Right, because it's DRAM. This is a great advancement.
Posted on Reply
#3
Mr. Perfect
So what is this in actual nanometers? Is it really 10nm?
Posted on Reply
#4
kondamin
Mr. PerfectSo what is this in actual nanometers? Is it really 10nm?
Yes, if they managed a smaller node they wouldn't be calling it 1y
Posted on Reply
Jun 6th, 2025 11:00 CDT change timezone

New Forum Posts

Popular Reviews

TPU on YouTube

Controversial News Posts