AMD surpassed its 30x25 energy efficiency goal and announced aggressive new 20x efficiency targets through 2027.
AMD has surpassed its 30x25 goal, which the company set in 2021 to improve the energy efficiency of AI-training and high-performance computing nodes by 30x from 2020 to 2025. This achievement marks the second major milestone in AMD's multi-decade effort to advance efficiency across computing platforms. The company first exceeded its 25x20 goal in 2020 by improving the energy efficiency of AMD mobile processors 25-fold in just six years. Building on that momentum, the 30x25 goal targeted AI and HPC workloads in accelerated nodes. As of mid-2025, AMD has gone beyond the target, achieving a 38x gain over the base system using a current configuration of four AMD Instinct MI355X GPUs and one AMD EPYC 5th Gen CPU, equating to a 97% reduction in energy for the same performance compared to systems from just five years ago. This gain represented more than a 2.5x acceleration over industry trends from the previous five years (2015-2020). The company achieved this through deep architectural innovations, aggressive optimization of performance-per-watt, and relentless engineering across its CPU and GPU product lines.
Building on this success, AMD is now setting an ambitious new target: a 20x improvement in rack-scale energy efficiency for AI training and inference by 2030, from a 2024 base year. This shift from node-level to rack-level efficiency focuses on the entire system, including CPUs, GPUs, memory, networking, storage and hardware-software co-design, based on AMD's latest designs and roadmap projections. AMD estimates this 20x improvement will exceed the industry improvement trend from 2018 to 2025 by almost 3x. Using training for a typical AI model in 2025 as a benchmark, these gains could enable significant energy savings and cost reductions for data center operators.
The projections underlying the 2030 target are grounded in a measurement methodology validated by energy-efficiency expert Dr. Jonathan Koomey. "By grounding the 2030 target in system-level metrics and transparent methodology, AMD is raising the bar for the industry," Dr. Koomey said. "The target gains in rack-scale efficiency will enable others across the ecosystem, from model developers to cloud providers, to scale AI compute more sustainably and cost-effectively."
AMD's 20x goal reflects hardware and system-level design improvements the company controls directly. However, AMD acknowledges that even greater delivered AI model efficiency gains will be possible—up to 5x over the goal period—as software developers discover smarter algorithms and continue innovating with lower-precision approaches at current rates. When those factors are included, overall energy efficiency for training a typical AI model could improve by as much as 100x by 2030. While AMD is not claiming that full multiplier in its own goal, the company aims to provide the hardware foundation that enables it and to support the open ecosystem and developer community through open standards, its open software approach with AMD ROCm, and close collaboration with partners to unlock those gains.