Demands to Modernize
Today’s public sector infrastructure faces unprecedented pressure to modernize quickly while reducing costs and risk without disrupting services. Existing hardware is being pushed to its limits as five key forces converge:
- AI demand is outpacing available compute, power, and data center capacity
- Legacy infrastructure and technical debt are slowing modernization efforts
- HPC demand is growing rapidly, driven in large part by AI workloads
- Cybersecurity requirements, including zero trust architectures, are expanding in scope and priority
- Hybrid and multi-cloud strategies are now the dominant enterprise operating model
Where is the Public Sector Going?
The opportunities in the public sector for AMD business partners are considerable. IT modernization is accelerating alongside global IT growth, with spending increasingly concentrated on AI, hybrid cloud, cybersecurity, and data center refresh cycles, reflecting continued expansion across infrastructure and software markets.
Public sector organizations also represent a substantial share of global technology spending. Governments at the federal, state, and local levels continue increasing investment in digital infrastructure, cybersecurity, cloud platforms, and data systems to modernize services and support mission-critical operations.
This trajectory signals that governments will remain one of the world’s largest IT “verticals,” with 2026 priorities favoring platforms that deliver security-focused, energy-efficient, and accelerated compute infrastructure capable of supporting AI-enabled workloads across on-premises, sovereign, and hybrid cloud environments.
How AMD Supports the Public Sector
AMD products enable public sector organizations to modernize infrastructure, accelerate outcomes, and deliver efficient digital services through a portfolio grounded in high-performance CPUs, advanced accelerators, leadership-level energy efficiency, and robust integrated security features. Across the public sector, AMD technology provides the infrastructure and performance required to address growing compute demands while navigating increasingly constrained budgets.
AMD EPYC™ server CPUs provide exceptional performance per watt, high throughput, and significant improvements in total cost of ownership (TCO). Compared to common legacy systems still deployed in public sector data centers, AMD EPYC server CPUs deliver transformative gains in productivity, density, and efficiency.
AMD EPYC™ Server CPUs for On-Prem Infrastructure
For public sector organizations that require on‑premises infrastructure due to factors like mission, sovereignty, security, or regulatory requirements, AMD EPYC server CPUs offer exceptional performance and efficiency.
Customers can expect:
Higher core counts and faster cores, with up to 192 cores and high frequency models (up to 5 GHz) to handle workloads more efficiently
Up to 17% better instructions per clock (IPC) for enterprise and up to 37% higher IPC in AI and high-performance computing (HPC)1
Up to 1.6x the performance per core in virtualized infrastructure2
Beyond pure performance, AMD EPYC server CPUs bring new levels of efficiency to public sector needs compared to competitor options, including up to 88% fewer servers when upgrading, up to 70% less power usage, and up to 71% lower 3-year TCO.3
AMD EPYC Processors in the Cloud
2025 was a landmark year for AMD EPYC server CPUs in the cloud. What looks like a long list of cloud deployments is really a single story. Across OEMs, cloud service providers, and fast-growing AI clouds, the industry is demanding AMD for performance, efficiency, security features, and scalability.
Public sector customers who rely on cloud workflows can expect:
Expanded availability and a broader portfolio of AMD EPYC server CPU instance types across memory, storage, and HPC workloads
Low TCO through robust energy efficiency and high performance per watt
Enhanced security features with AMD Infinity Guard, strengthening confidential computing and protecting AI workloads4
For partners, this makes cloud instances enabled by AMD EPYC server CPUs a compelling recommendation for workloads that need cost-efficient scaling, particularly analytics, GIS systems, and high-uptime administrative workloads.
Energy efficiency is also an important requirement for public sector modernization as agencies balance new AI and analytics demands with fixed power, space, and cooling envelopes. By increasing performance per watt and consolidating more work into fewer systems, accelerated platforms can help reduce data center energy use and cooling requirements while maintaining (or improving) service levels. For public sector IT leaders, this can translate into measurable progress on carbon-reduction commitments and improved operational resilience when capacity is constrained. Coupled with modernization best practices such as right sizing, higher utilization, and placing workloads appropriately across on-premises and sovereign/hybrid cloud, platforms advanced by AMD solutions can help support both mission outcomes and sustainability goals.
Transforming the World
Hardware advanced by AMD is powering many of the world’s most demanding public sector AI and HPC applications. On the November 2025 TOP500 list, AMD enables 177 of the world’s 500 fastest supercomputers, including systems at the very top of the rankings. For agencies with sustainability mandates, the November 2025 Green500 rankings provide a benchmark for energy efficiency by measuring High Performance LINPACK (HPL) performance per watt; based on AMD analysis, 173 systems on the latest Green500 are advanced by AMD technology.5
In U.S. national labs, AMD is also advancing new large-scale AI infrastructure. Oak Ridge National Laboratory’s “Lux” AI system (announced in 2025) will use AMD Instinct™ MI355X GPUs and AMD EPYC server CPUs, along with AMD Pensando™ networking, to expand near-term AI capacity for large-scale AI training and distributed inference.
In Europe, AMD was also selected to power France’s first exascale supercomputer. The “Alice Recoque” system—led by GENCI and operated by CEA, with Eviden as the system provider under the EuroHPC JU program—will use next-generation AMD EPYC server CPUs together with AMD Instinct™ MI430X accelerators to support large-scale scientific computing and sovereign AI workloads.
Together, these examples illustrate how AMD platforms can help organizations increase throughput, reduce infrastructure cost, and accelerate mission-relevant outcomes from real-time vision at the edge to national-scale supercomputing.
Support the Public Sector with AMD
Your public sector customers are facing unprecedented demands for increased performance, transformative new workloads in the form of AI and HPC, and growing restrictions on budgets and server footprints. AMD leadership across both CPU and GPU infrastructure provides an outstanding portfolio to win modernization projects, expand cloud and HPC engagements, deliver sovereign AI capability, and build long-term public sector accounts. The demand is growing globally, and AMD technologies give you a clear competitive advantage in addressing it.
Introduce the power of AMD technologies to your customers and help them evolve their infrastructure for groundbreaking results. Speak to your AMD representative to learn more or read about public sector solutions on amd.com.
AMD Arena
Enhance your AMD product knowledge with training on AMD Ryzen™ PRO, AMD EPYC™, AMD Instinct™, and more.
Subscribe
Get monthly updates on AMD’s latest products, training resources, and Meet the Experts webinars.
Related Articles
Related Training Courses
Related Webinars
Footnotes
- 9xx5-001: Based on AMD internal testing as of 9/10/2024, geomean performance improvement (IPC) at fixed-frequency.
- 5th Gen EPYC CPU Enterprise and Cloud Server Workloads generational IPC Uplift of 1.170x (geomean) using a select set of 36 workloads and is the geomean of estimated scores for total and all subsets of SPECrate®2017_int_base (geomean), estimated scores for total and all subsets of SPECrate®2017_fp_base (geomean), scores for Server Side Java multi instance max ops/sec, representative Cloud Server workloads (geomean), and representative Enterprise server workloads (geomean).
“Genoa” Config (all NPS1): EPYC 9654 BIOS TQZ1005D 12c12t (1c1t/CCD in 12+1), FF 3GHz, 12x DDR5-4800 (2Rx4 64GB), 32Gbps xGMI;
“Turin” config (all NPS1): EPYC 9V45 BIOS RVOT1000F 12c12t (1c1t/CCD in 12+1), FF 3GHz, 12x DDR5-6000 (2Rx4 64GB), 32Gbps xGMI
Utilizing Performance Determinism and the Performance governor on Ubuntu® 22.04 w/ 6.8.0-40-generic kernel OS for all workloads.
- 5th Gen EPYC generational ML/HPC Server Workloads IPC Uplift of 1.369x (geomean) using a select set of 24 workloads and is the geomean of representative ML Server Workloads (geomean), and representative HPC Server Workloads (geomean).
“Genoa” Config (all NPS1) “Genoa” config: EPYC 9654 BIOS TQZ1005D 12c12t (1c1t/CCD in 12+1), FF 3GHz, 12x DDR5-4800 (2Rx4 64GB), 32Gbps xGMI;
“Turin” config (all NPS1): EPYC 9V45 BIOS RVOT1000F 12c12t (1c1t/CCD in 12+1), FF 3GHz, 12x DDR5-6000 (2Rx4 64GB), 32Gbps xGMI
Utilizing Performance Determinism and the Performance governor on Ubuntu 22.04 w/ 6.8.0-40-generic kernel OS for all workloads except LAMMPS, HPCG, NAMD, OpenFOAM, Gromacs which utilize 24.04 w/ 6.8.0-40-generic kernel.
SPEC® and SPECrate® are registered trademarks for Standard Performance Evaluation Corporation.
- 9xx5-071: VMmark® 4.0.1 host/node FC SAN comparison based on “independently published” results as of 10/10/2024.
Configurations:
2 node, 2P AMD EPYC 9575F (128 total cores) powered server running VMware ESXi8.0 U3, 3.31 @ 4 tiles, https://www.infobellit.com/BlueBookSeries/VMmark4-FDR-1003
2 node, 2P AMD EPYC 9554 (128 total cores) powered server running VMware ESXi 8.0 U3, 2.64 @ 3 tiles, https://www.infobellit.com/BlueBookSeries/VMmark4-FDR-1002
2 node, 2P Intel Xeon Platinum 8592+ (128 total cores) powered server running VMware ESXi 8.0 U3, 2.06 @ 2.4 Tiles, https://www.infobellit.com/BlueBookSeries/VMmark4-FDR-1001
VMmark is a registered trademark of VMware in the US or other countries.
- Based on 80,000 Integer Performance, all performance benefits listed are estimates. See endnotes SP5TCO-055 and SP5TCO-056.
- AMD Infinity Guard features vary by EPYC™ Processor generations and/or series. Infinity Guard security features must be enabled by server OEMs and/or Cloud Service Providers to operate. Check with your OEM or provider to confirm support of these features. GD-183A
- Based on the November 2025 TOP500 and Green500 lists (top500.org). “177 systems” reflects the number of TOP500 systems powered by AMD processors. “173 systems” reflects AMD internal analysis of the November 2025 Green500 list identifying systems powered by AMD EPYC processors. TOP500 and Green500 rankings are based on published High Performance LINPACK (HPL) benchmark results; Green500 ranks systems by performance per watt.
- 9xx5-001: Based on AMD internal testing as of 9/10/2024, geomean performance improvement (IPC) at fixed-frequency.
- 5th Gen EPYC CPU Enterprise and Cloud Server Workloads generational IPC Uplift of 1.170x (geomean) using a select set of 36 workloads and is the geomean of estimated scores for total and all subsets of SPECrate®2017_int_base (geomean), estimated scores for total and all subsets of SPECrate®2017_fp_base (geomean), scores for Server Side Java multi instance max ops/sec, representative Cloud Server workloads (geomean), and representative Enterprise server workloads (geomean).
“Genoa” Config (all NPS1): EPYC 9654 BIOS TQZ1005D 12c12t (1c1t/CCD in 12+1), FF 3GHz, 12x DDR5-4800 (2Rx4 64GB), 32Gbps xGMI;
“Turin” config (all NPS1): EPYC 9V45 BIOS RVOT1000F 12c12t (1c1t/CCD in 12+1), FF 3GHz, 12x DDR5-6000 (2Rx4 64GB), 32Gbps xGMI
Utilizing Performance Determinism and the Performance governor on Ubuntu® 22.04 w/ 6.8.0-40-generic kernel OS for all workloads.
- 5th Gen EPYC generational ML/HPC Server Workloads IPC Uplift of 1.369x (geomean) using a select set of 24 workloads and is the geomean of representative ML Server Workloads (geomean), and representative HPC Server Workloads (geomean).
“Genoa” Config (all NPS1) “Genoa” config: EPYC 9654 BIOS TQZ1005D 12c12t (1c1t/CCD in 12+1), FF 3GHz, 12x DDR5-4800 (2Rx4 64GB), 32Gbps xGMI;
“Turin” config (all NPS1): EPYC 9V45 BIOS RVOT1000F 12c12t (1c1t/CCD in 12+1), FF 3GHz, 12x DDR5-6000 (2Rx4 64GB), 32Gbps xGMI
Utilizing Performance Determinism and the Performance governor on Ubuntu 22.04 w/ 6.8.0-40-generic kernel OS for all workloads except LAMMPS, HPCG, NAMD, OpenFOAM, Gromacs which utilize 24.04 w/ 6.8.0-40-generic kernel.
SPEC® and SPECrate® are registered trademarks for Standard Performance Evaluation Corporation.
- 9xx5-071: VMmark® 4.0.1 host/node FC SAN comparison based on “independently published” results as of 10/10/2024.
Configurations:
2 node, 2P AMD EPYC 9575F (128 total cores) powered server running VMware ESXi8.0 U3, 3.31 @ 4 tiles, https://www.infobellit.com/BlueBookSeries/VMmark4-FDR-1003
2 node, 2P AMD EPYC 9554 (128 total cores) powered server running VMware ESXi 8.0 U3, 2.64 @ 3 tiles, https://www.infobellit.com/BlueBookSeries/VMmark4-FDR-1002
2 node, 2P Intel Xeon Platinum 8592+ (128 total cores) powered server running VMware ESXi 8.0 U3, 2.06 @ 2.4 Tiles, https://www.infobellit.com/BlueBookSeries/VMmark4-FDR-1001
VMmark is a registered trademark of VMware in the US or other countries.
- Based on 80,000 Integer Performance, all performance benefits listed are estimates. See endnotes SP5TCO-055 and SP5TCO-056.
- AMD Infinity Guard features vary by EPYC™ Processor generations and/or series. Infinity Guard security features must be enabled by server OEMs and/or Cloud Service Providers to operate. Check with your OEM or provider to confirm support of these features. GD-183A
- Based on the November 2025 TOP500 and Green500 lists (top500.org). “177 systems” reflects the number of TOP500 systems powered by AMD processors. “173 systems” reflects AMD internal analysis of the November 2025 Green500 list identifying systems powered by AMD EPYC processors. TOP500 and Green500 rankings are based on published High Performance LINPACK (HPL) benchmark results; Green500 ranks systems by performance per watt.