Advancements in semiconductor technology, GPU development, and data center infrastructure that power AI computing capabilities.
A joint report by EPRI and Epoch AI reveals that AI model training could require over 4 GW of power by 2030, highlighting the growing energy demands of AI technologies.
Normal Computing has announced the tape-out of CN101, the first thermodynamic computing chip, aimed at enhancing AI and scientific computing efficiency.
Aitech has introduced the A230 Vortex AI GPGPU supercomputer, designed for military operations across various environments, as announced in a press release.
Nvidia and AMD have agreed to pay the US government 15% of their revenue from AI chip sales to China, following a new licensing agreement.
SK Hynix forecasts a 30% annual growth in the AI memory market until 2030, driven by increased demand for high-bandwidth memory (HBM) from tech giants like Amazon, Microsoft, and Google, according to Reuters.
SoftBank has acquired Foxconn's Ohio electric vehicle plant to advance its $500 billion Stargate AI infrastructure initiative, repurposing the site for data center use.
ChemTreat has launched CTSolutions D2C, a comprehensive water treatment and monitoring program for direct-to-chip cooling systems in AI-driven data centers, as announced in a press release.
Deutsche Telekom has raised its 2025 financial guidance and announced a collaboration with Nvidia and Brookfield to build an AI data center in Germany.
Oracle has announced the availability of its Globally Distributed Exadata Database on Exascale Infrastructure, designed to enhance performance and availability for mission-critical applications.
Blaize has introduced its AI Platform, designed to deliver multi-modal intelligence for mission-critical workloads, as announced in a press release. The platform aims to enhance AI deployment across various industries by reducing latency, cost, and complexity.
Tesla has decided to shut down its Dojo supercomputer project, redirecting its focus towards using Nvidia, AMD, and Samsung chips for its AI and data center needs.
The LF AI & Data Foundation has announced the Vortex Project, a new columnar storage format designed to improve data access for AI and analytics, with contributions from industry leaders like Microsoft and Snowflake.
AMD's Q2 2025 revenue reached $7.7 billion, a 32% year-over-year increase, despite a dip in AI datacenter revenue due to US export restrictions to China.
Chelsio Communications has announced the broad sampling availability of its T7 Data Processing Unit (DPU), designed for high-performance offloads in storage, security, and AI applications.
Vertiv has announced the global availability of Vertiv OneCore, a scalable prefabricated solution designed to accelerate high-density data center deployments, according to a press release.
NEO Semiconductor has introduced the world's first Extreme High Bandwidth Memory (X-HBM) architecture for AI chips, offering 16X more bandwidth or 10X higher density than traditional HBM, announced in a press release.
XConn Technologies and ScaleFlux have partnered to optimize interoperability between XConn's CXL 3.1 switch and ScaleFlux's MC500 memory controller, enhancing AI and cloud infrastructure.
The Open Compute Project Foundation has announced a new Universal D2D Transaction and Link-Layer specification to enhance silicon diversity in AI and HPC clusters.
Flex has introduced a new power shelf system for the NVIDIA GB300 NVL72 platform, enhancing power efficiency for AI data centers, announced in a press release.
A new AI data center near Cheyenne, Wyoming, will require more electricity than all state homes combined, highlighting the growing energy demands of AI infrastructure.
Weekly coverage of AI hardware developments including chips, GPUs, cloud platforms, and data center technology.
