Ampere announced a 256-core AmpereOne server CPU and a Qualcomm inference partnership on May 16, 2024. The processor was presented as a roadmap design—not a retail launch—and public materials available through August 18, 2026, do not establish broad commercial availability of that specific 256-core SKU.
The announcement combined a high-density Arm server CPU with Qualcomm Cloud AI 100 accelerators for large-language-model inference. Since then, Ampere’s platform activity, SoftBank’s acquisition of the company, and Qualcomm’s separate Dragonfly server roadmap have changed the context without confirming that the announced 256-core part shipped.
![AMD EPYC 9004 [4th Gen] 9124 Hexadeca-core [16 Core] 3 GHz Processor](https://m.media-amazon.com/images/I/316ISTIk-UL._SL160_.jpg)
As an Amazon Associate I earn from qualifying purchases.
What Ampere announced on May 16, 2024
Ampere said the next AmpereOne expansion would scale to 256 cores, use 12 memory channels and be manufactured on TSMC’s 3-nanometer process. The company described it as an extension of the AmpereOne platform, whose preceding generation reached 192 cores, according to CRN.
The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Ampere also said the design was intended to use the same general air-cooled thermal approach as its 192-core AmpereOne. It claimed more than 40% higher performance than any CPU on the market at the time. That figure, like Ampere’s performance-per-watt and performance-per-rack comparisons with AMD EPYC Genoa and Bergamo, was a company claim rather than an independently verified result.
#1 Best Overall
- Media streaming
- Medium capacity data managementSpecifications
- No of CPU Cores: 32
- Base Clock: 2.4GHz
- Max Boost Clock: Up to 3.3GHz
The announcement also covered 12-channel platforms, OEM and ODM systems, FlexSKU and FlexSpeed functions, a NETINT video-processing collaboration, and an approach supporting UCIe and third-party intellectual property. These were platform and ecosystem disclosures, not evidence that every feature or processor SKU was immediately orderable.
Ampere’s announcement did not publish a retail price, a generally available ordering path, independent benchmark results or a definitive shipping date. CRN reported that Ampere expected the 256-core version in 2025, but later public materials reviewed through August 18, 2026, emphasize AmpereOne M platforms, systems partnerships and cloud deployments rather than confirming general availability of this exact processor.
What AmpereOne is
AmpereOne is Ampere’s own Arm-compatible server CPU family. Unlike earlier Ampere Altra processors, which used Arm Neoverse cores, AmpereOne uses Ampere-designed CPU cores. Arm compatibility means the processor can run software built for the 64-bit Arm architecture, but it does not guarantee that every x86 server application or appliance will run unchanged.
A 256-core count describes potential parallelism, not a fixed performance multiplier. Results depend on single-thread speed, clock frequency, memory bandwidth and latency, NUMA placement, vector and cryptographic extensions, compiler and library optimization, software compatibility, cooling limits and how effectively an application scales across hundreds of threads.
Rank #2
- Intel Core i5 2.50 GHz processor offers hyper-threading architecture that delivers high performance for demanding applications with improved onboard graphics and turbo boost
- The processor features Socket LGA-1700 socket for installation on the PCB
- Its 18 MB of L3 cache is good enough to carry routine data and process them in a flash giving you fast and smooth performance
- Built-in Intel UHD Graphics 730 controller for improved graphics and visual quality. Supports up to 4 monitors.
How the Qualcomm inference solution was supposed to work
The Qualcomm announcement concerned a joint system design, not a Qualcomm accelerator integrated into the Ampere CPU. Ampere CPUs—Altra or AmpereOne—would be paired with Qualcomm Cloud AI 100 or Cloud AI 100 Ultra inference accelerators. CRN described a Supermicro server configuration in an interview with Ampere executive Jeff Wittich.
Division of labor
- Ampere CPU: host operating-system work, request handling, orchestration, networking, input/output and tasks unsuitable for the accelerator.
- Qualcomm accelerator: neural-network inference operations for supported models and software frameworks.
- Server software: model loading, batching, quantization, scheduling and monitoring, all of which affect latency and throughput.
The target was large-language-model inference, including models that are inefficient to run on CPUs alone. This differs from a CPU-only deployment and from a GPU-centric system: supported operators, model formats, precision, memory capacity, serving frameworks and power limits determine the practical result.
CRN’s contemporaneous coverage is available at CRN.
Free tools Windows power users keep installed
One-click scans. No signup required.
Why 256 cores mattered in 2024
Server operators were pursuing denser compute to increase work per rack without automatically moving to exotic cooling. More cores can raise capacity for containers, microservices, web serving, media processing and other scale-out workloads. If the software scales well, fewer servers may be needed for the same parallel workload, reducing space, networking and power overhead.
Rank #3
Air cooling can simplify operations compared with liquid cooling, but “air-cooled” is not a universal drop-in guarantee. The complete server’s thermal design, chassis airflow, memory population and power envelope still matter. High core density can also expose software bottlenecks, synchronization costs and memory contention rather than remove them.
What changed after the announcement
AmpereOne M and systems activity
In May 2025, Ampere launched its Systems Builders program and announced additional 12-channel AmpereOne M platforms with Broadcom, Supermicro, Giga Computing, ASRock Rack, Jabil and Rebellions. Ampere said a Giga Computing platform was available for purchase and that a Jabil module was planned for sampling in the third quarter of 2025. Those statements establish platform activity, not availability of the announced 256-core SKU. See Ampere’s Systems Builders announcement.
Ampere’s March 2026 press listings include new AmpereOne and AmpereOne M cloud deployments in Europe, but the listings do not identify those deployments as 256-core products.
SoftBank ownership
SoftBank completed its acquisition of Ampere on November 25, 2025. Ampere should therefore be described today as a SoftBank-owned company, not simply as an independent Oracle-backed chip designer. The ownership change alters the strategic backdrop around Arm-based data-center silicon, but there is no cited confirmation that SoftBank canceled, changed or shipped the 256-core processor. Ampere’s press archive records the acquisition and subsequent announcements: Ampere press archive.
Rank #4
- Intel dual CPU sockets: This C612 server chip motherboard is designed with dual CPU sockets, which can support Intel Core i7 5th/6th generation processors and Xeon E5 V3/V4 series processors on LGA 2011-3 socket. (Note: If only one CPU is installed, please install it in the right slot, and the graphics card needs to be installed in the bottom two slots.)
- DDR4 4-channel memory slot: The memory slot of the LGA 2011-3 motherboard is designed with four channels, which can install 8 memory. It supports effective frequencies of 2133/2400MHz, and the maximum capacity is 256GB. (Non-ECC memory is not compatible when using E5 V4 series processors)
- PCIe 3.0 protocol standard: Equipped with 4 PCIe 3.0 X16 graphics card slots (with steel case). The transfer rate can reach 15.754 GB/s using one graphics card, and the performance can be improved by at least 50% by using two graphics cards. Equipped with dual M.2 hard disk slots, it can achieve fast reading even if multiple programs are running
- Stable power supply: use 24+8+8pin standard power supply interface (need to use a dedicated power supply for dual server motherboards), 12 (CPU) + 4 (memory) + 1 (C612 chip) phase power supply. Precise modularization provides good heat dissipation and makes the program run more stably
- Strong expandability: The X99 motherboard is equipped with multiple expansion interfaces to ensure that the motherboard has more room for improvement. These include 4*USB 3.0 ports, 4*USB 2.0 ports, 10*SATA 3.0 ports, 4*3pin sys fan, 2*4pin CPU fan. Besides, dual network ports allow your computer to do more things
2026 CPU-based inference work
SoftBank and Ampere reported a February 2026 evaluation of CPU-based small-language-model and mixture-of-experts inference using Ampere-optimized llama.cpp. That work demonstrates software experimentation; it does not establish that the 256-core AmpereOne entered general availability. Details are in SoftBank’s release.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Qualcomm’s separate 2026 roadmap
In June 2026, Qualcomm announced a broader data-center portfolio: the Dragonfly C1000 server CPU, Dragonfly AI200, AI250 and AI300 inference accelerators, High Bandwidth Compute and related connectivity and rack-scale products. Qualcomm described the C1000 as a 250-plus-core chiplet design and said commercial availability was expected in 2028, alongside a multigeneration data-center CPU agreement with Meta.
This is separate from the 2024 Ampere-Qualcomm solution. The Dragonfly C1000 is Qualcomm’s own future server CPU; it is not the 256-core AmpereOne. The announcement is documented at Qualcomm.
What buyers should verify before committing
- Exact SKU status: Determine whether the 256-core processor is announced, sampled, orderable or shipping. Require a manufacturer part number and a current order channel.
- Qualified server: Confirm the OEM, chassis, firmware, BIOS, supported operating systems and memory-population rules.
- Memory behavior: Check capacity, DIMM type, bandwidth, latency and NUMA topology. Large models can be limited by memory rather than core count.
- Power and cooling: Evaluate CPU TDP, complete-node consumption, rack power and the server’s validated airflow design.
- Arm64 software: Inventory operating systems, containers, databases, JVMs, proprietary binaries, observability agents and security tools. Legacy x86 software may require recompilation, emulation or replacement.
- AI stack: For Cloud AI 100 systems, verify driver versions, model formats, supported operators, quantization, batching, serving runtimes and accelerator memory.
- Independent performance: Compare complete systems using the same model, precision, batch size, latency target, power boundary and software versions. Do not treat “40% faster” as a universal result.
- Total cost and portability: Include software migration, support, networking, memory, accelerator costs and the ability to move workloads among Ampere, x86 and cloud Arm environments.
Where the platform fits—and where it does not
| Workload or buyer | Likely fit | Main caution |
|---|---|---|
| Cloud-native Arm64 services and high-concurrency applications | High core density can improve throughput and rack utilization | Validate scaling, memory placement and software support |
| CPU-plus-accelerator LLM inference | Separates host orchestration from neural-network execution | Confirm Qualcomm software, model coverage and accelerator availability |
| Legacy x86 enterprise software | Possible only where Arm64 builds or emulation are supported | Vendor appliances and proprietary binaries may block migration |
| GPU-style AI training | Not the stated focus of the 2024 partnership | Do not assume an inference accelerator replaces a training platform |
| Immediate standardized hardware purchase | Established AmpereOne M or other OEM systems may be options | The specific 256-core SKU lacks a confirmed broad-availability signal in the cited materials |
Potential alternatives include AMD EPYC, Intel Xeon, AWS Graviton, Microsoft Azure Arm instances, Google Cloud Arm offerings, Nvidia inference systems, Qualcomm accelerator servers and custom hyperscaler silicon. Compare workload fit, software compatibility, memory, latency, power, total cost and availability—not headline core count.
Current answer: announced roadmap, not proven mass-market product
Ampere’s May 16, 2024 disclosure was strategically significant: it paired an ambitious 12-channel, 256-core Arm CPU roadmap with a concrete CPU-plus-Qualcomm-accelerator inference concept. As of August 18, 2026, however, the public materials cited here establish neither a broad retail launch nor a generally available cloud instance for that exact 256-core processor. Treat it as an announced roadmap part until Ampere or a qualified OEM supplies a verifiable SKU, system listing and ordering path.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
