Research Note: Marvell Custom HBM for Cloud AI
Marvell recently announced a new custom high-bandwidth memory (HBM) compute architecture that addresses the scaling challenges of XPUs in AI workloads. The new architecture enables higher compute and memory density, reduced power consumption, and lower TCO for custom XPUs.
Extreme Networks Platform ONE: A New Approach to AI-Driven Network Management
Extreme Networks recently announced the launch of Extreme Platform ONE, its comprehensive management platform integrating networking, security, and AI to streamline enterprise infrastructure operations.
Research Note: Enfabrica ACF-S Millennium
First detailed at Hot Chips 2024, Enbrica recently announced that its ACF-S “Millennium” chip, which addresses the limitations of traditional networking hardware for AI and accelerated computing workloads, will be available to customers in calendar Q1 2025.
Research Note: Cohesity & Veritas Complete Merger
Cohesity completed its long-awaited merger with Veritas’ enterprise data protection business. The combined entity, valued at $7 billion, now serves 12,000 customers globally and generates $1.5 billion in ARR.
Research Note: Dell AI Products & Services Updates
Dell Technologies has made significant additions to its AI portfolio with its recent announcements at SC24 and Microsoft Ignite 2024 in November. The announcements span infrastructure, ecosystem partnerships, and professional services, targeting accelerated AI adoption, operational efficiency, and sustainability in enterprise environments.
Understanding AI Data Types: The Foundation of Efficient AI Models
AI Datatypes aren’t just a technical detail—it’s a critical factor that affects performance, accuracy, power efficiency, and even the feasibility of deploying AI models.
Understanding the datatypes used in AI isn’t just for hands-on practitioners, you often see published benchmarks and other performance numbers broken out by datatype (just look at an NVIDIA GPU data sheet). What’s it all mean?
Research Note: AWS Trainium2
Tranium is AWS’s machine learning accelerator, and this week at its re:Invent event in Las Vegas, it announced the second generation, the cleverly named Trainium2, purpose-built to enhance the training of large-scale AI models, including foundation models and large language models.
Quick Take: AWS re:Invent AI Services & Infrastructure Announcement Wrap-Up
AWS has made close to 50 announcements in the first 2 days of re:Invent. This blog takes a look at the most interesting AI services & infrastructure related ones.
Quick Take: AWS re:Invent Day 1
AWS unveiled a range of new features and services, reflecting its continued focus on innovation across generative AI, compute, and storage. These announcements include enhancements to Amazon Bedrock for improved testing and data integration, new capabilities for the generative AI assistant Amazon Q, high-performance storage-optimized EC2 instances, and advanced storage solutions like intelligent tiering and a dedicated data transfer terminal
Research Note: Hammerspace Global Data Platform v5.1 with Tier 0
Hammerspace recently announced the version 5.1 release of its Hammerspace Global Data Platform. The flagship feature of the release its new Tier 0 storage capability, which takes unused local NVMe storage on a GPU server and uses it as part of the global shared filesystem. This provides higher-performance storage for the GPU server than can be delivered from remote storage nodes – ideal for AI and GPU-centric workloads.