An engineer who fine-tuned Llama 3.3 70B on Amazon Bedrock found that the pricing page lists training costs only for Llama 2, not for the newer model they actually used. Probing all 33 Bedrock regions…
AWS account architecture for MLflow model governance across multiple accounts requires either a hub-and-spoke topology using AWS Resource Access Manager (AWS RAM) or a hybrid isolation model for regul…
Qualcomm Technologies and Amazon have signed a multi-generation collaboration to deliver custom AI inference silicon at scale for AWS data centers, targeting compute, memory bandwidth, networking, and…
InMotion Cloud, backed by InMotion Hosting's 25 years of infrastructure experience, launched Managed Private Cloud, a fixed-price managed cloud platform built on OpenStack and bare-metal hardware, wit…
AWS published a guide for governing ML models across accounts using MLflow and Amazon SageMaker AI Model Registry sync, detailing hub-and-spoke and hybrid topologies for regulated environments. The po…
AWS announced an enhanced sync between managed MLflow on Amazon SageMaker AI and the SageMaker AI Model Registry, automatically carrying training metrics, evaluation results, lineage, and lifecycle st…
HPE Zerto, in collaboration with AWS, has built an agentic troubleshooting system powered by Amazon Bedrock to help operators manage hybrid and multi-cloud infrastructures. The system, deployed as a p…
DiDi Global's International Business Group partnered with AWS to build an intelligent contact center quality assurance system on Amazon Bedrock, replacing an opaque third-party solution. In production…
Qualcomm announced a multigeneration deal with Amazon Web Services (AWS) to develop custom silicon and optical connectivity solutions for large-scale AI data centers, marking its return to the data ce…
AWS is offering millions of students at 132 universities across 18 countries a free year of Kiro, its AI agent for software engineering, with 1,000 credits per month and no credit card required. The p…
A September 9, 2026 analysis of twelve AI accelerators from NVIDIA, AMD, Google, AWS, Cerebras, and Tenstorrent finds that memory per dollar, not FLOPS, is the primary purchasing criterion for AI comp…
Honeycomb's Nick Travaglini argues that the traditional three pillars of observability—metrics, logs, and traces—drive up AI observability costs, and recommends wide events as a single format to keep …
A developer outlines a foundational learning path for DevOps engineers that prioritizes Linux, networking, Git, Docker, and cloud concepts before tackling Kubernetes. The author argues that mastering …
NVIDIA invested $3.5 billion in MediaTek to advance NVLink Fusion, an interconnect technology that extends NVIDIA's proprietary NVLink protocol to custom AI chips from other vendors, enabling them to …
Dynoxide, a DynamoDB-compatible local database, has added vector search support in versions 1.0.0 and 1.1.0, including vector indexes, SearchVectors, and write paths. The developer behind Dynoxide doc…
Arm announced expansions to its physical AI ecosystem at Arm Everywhere China, including the expansion of Arm Total Design to encompass more than 80 companies and the introduction of the Robotics Capa…
A procurement census of 31 AI agent feature and delivery-surface combinations across Anthropic, OpenAI, Google, and AWS, read on September 9, 2026, finds that zero-data-retention (ZDR) eligibility var…
Anthropic's Claude Fable 5.1 became available on AWS last week, delivering frontier intelligence for coding, scientific research, and enterprise workflows, with data retention up to 30 days and human …
Tracarbon, a Python library that tracks device energy consumption and calculates carbon emissions, now supports monitoring GPU power and carbon output while running local large language models. It det…
Storj, founded in 2014 by Shawn Wilkinson, provides cloud object storage that claims to match major providers' security and performance at 90% lower cost and with reduced carbon emissions. The company…