# m9g

Published articles for m9g.

This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles.

## AWS launches Graviton5 server CPU in M9g and M9gd instances for agentic AI workloads

DevFeed: [AWS launches Graviton5 server CPU in M9g and M9gd instances for agentic AI workloads](<https://devfeed.tech/articles/aws-tunes-up-graviton5-for-agentic-ai-boosts-bang-for-the-buck-bigtime-47974.md>)

Original publisher: [Read original article](<https://www.nextplatform.com/compute/2026/06/11/aws-tunes-up-graviton5-for-agentic-ai-boosts-bang-for-the-buck-bigtime/5254538>)

Author: Timothy Prickett Morgan

Published: 2026-06-11T21:42:38Z

Content type: news

Language: en

Sources: [The Next Platform: In-depth coverage of high end computing](<https://devfeed.tech/sources/the-next-platform-in-depth-coverage-of-high-end-computing.md>)

Topics: [Amazon Web Services](<https://devfeed.tech/topics/aws.md>), [cpu](<https://devfeed.tech/topics/cpu.md>), [Arm](<https://devfeed.tech/topics/arm.md>), [ddr5](<https://devfeed.tech/topics/ddr5.md>), [NUMA](<https://devfeed.tech/topics/numa.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>)

Tags: [amazon](<https://devfeed.tech/tags/amazon.md>), [annapurna-labs](<https://devfeed.tech/tags/annapurna-labs.md>), [aws](<https://devfeed.tech/tags/aws.md>), [chip](<https://devfeed.tech/tags/chip.md>), [chiplets](<https://devfeed.tech/tags/chiplets.md>), [compute](<https://devfeed.tech/tags/compute.md>), [ddr5](<https://devfeed.tech/tags/ddr5.md>), [graviton4](<https://devfeed.tech/tags/graviton4.md>), [graviton5](<https://devfeed.tech/tags/graviton5.md>), [interconnect](<https://devfeed.tech/tags/interconnect.md>), [labs](<https://devfeed.tech/tags/labs.md>), [m9g](<https://devfeed.tech/tags/m9g.md>), [m9gd](<https://devfeed.tech/tags/m9gd.md>), [memory](<https://devfeed.tech/tags/memory.md>), [neoverse](<https://devfeed.tech/tags/neoverse.md>), [performance](<https://devfeed.tech/tags/performance.md>), [power](<https://devfeed.tech/tags/power.md>), [processor](<https://devfeed.tech/tags/processor.md>), [semiconductor](<https://devfeed.tech/tags/semiconductor.md>)

### AI overview

This article examines AWS Graviton5, an Arm server CPU shipping in new M9g and M9gd instances. It describes the four-chiplet design, 192 Neoverse V3 cores, DDR5 and PCIe 6.0 controllers, CXL 3.0 support, and the tradeoffs involving memory capacity, manufacturing cost, and power efficiency.

### Source excerpt

Back in December, the Annapurna Labs chip division of Amazon Web Services showed off a preview of it ...

## Graviton5 in Production at Honeycomb: Per-service Results From the m8g to m9g Migration

DevFeed: [Graviton5 in Production at Honeycomb: Per-service Results From the m8g to m9g Migration](<https://devfeed.tech/articles/graviton5-in-production-at-honeycomb-per-service-results-from-the-m8g-to-m9g-migration-77904.md>)

Original publisher: [Read original article](<https://www.honeycomb.io/blog/graviton5-honeycomb-per-service-results-m8g-m9g-migration>)

Author: Liz Fong-Jones

Published: 2026-06-10T13:00:00Z

Content type: article

Language: en

Sources: [Honeycomb Blog](<https://devfeed.tech/sources/honeycomb-blog.md>)

Topics: [NUMA](<https://devfeed.tech/topics/numa.md>), [Comparing SaaS, PaaS, and IaaS](<https://devfeed.tech/topics/comparing-saas-paas-and-iaas.md>)

Tags: [arm64](<https://devfeed.tech/tags/arm64.md>), [aws](<https://devfeed.tech/tags/aws.md>), [cost](<https://devfeed.tech/tags/cost.md>), [cpu](<https://devfeed.tech/tags/cpu.md>), [dogfooding](<https://devfeed.tech/tags/dogfooding.md>), [graviton5](<https://devfeed.tech/tags/graviton5.md>), [kubernetes](<https://devfeed.tech/tags/kubernetes.md>), [m9g](<https://devfeed.tech/tags/m9g.md>), [technical-deep-dives](<https://devfeed.tech/tags/technical-deep-dives.md>)

### AI overview

Honeycomb reports a 60-day production comparison of AWS Graviton4 m8g and Graviton5 m9g instances running identical workloads concurrently. The measured services used 11-26% less CPU on Graviton5; ingest P99 latency and sampling queue depth also improved, with no fleet regressions reported. The article explains how to evaluate per-service bottlenecks when considering a migration.

### Source excerpt

This is the fourth installment in the Graviton retrospective series we've been writing since 2021. The methodology is the same one I always reach for: hold the workload constant, run both generations on the same Kubernetes namespace concurrently, and let the per-pod numbers speak.