Advancements in semiconductor technology, GPU development, and data center infrastructure that power AI computing capabilities.
GSI Technology announced that its Gemini-I APU achieved GPU-level performance in AI workloads while consuming over 98% less energy, according to a Cornell University study presented at the Micro ’25 conference.
Connectbase announced plans to introduce GPU-as-a-Service (GPUaaS) within its Federated Ecosystem Fabric, enabling providers to sell and manage GPU capacity alongside connectivity services in a unified system.
NEXT Semiconductor Technologies has announced the NXLR1 LiDAR Processor Unit, a system-on-chip integrating 12 high-speed ADC channels with proprietary Full Waveform Capture processing for enhanced perception in autonomous vehicles.
Nebius has launched its AI Cloud platform in Israel, featuring one of the country's first public deployments of NVIDIA Blackwell GPUs, enabling local startups and researchers to access advanced compute resources.
Oracle and AMD announced an expanded partnership that will see Oracle Cloud Infrastructure deploy 50,000 AMD Instinct MI450 GPUs starting in Q3 2026. The new AI supercluster will use AMD’s Helios rack design and next-generation EPYC CPUs.
Arm Holdings has expanded its Flexible Access licensing program to include its Armv9 edge AI platform, offering startups and device makers low-cost access to chip design tools for on-device AI development.
Intel has announced a new data center GPU, code-named Crescent Island, optimized for AI inference tasks and designed for air-cooled enterprise servers. The GPU features 160GB of LPDDR5X memory and is based on the Xe3P microarchitecture, with customer sampling expected in the second half of 2026.
The global deep ultraviolet (DUV) lithography systems market is projected to grow from $15.44 billion in 2024 to $23.69 billion by 2031, according to a report by Valuates Reports. Growth is driven by AI, 5G, and semiconductor manufacturing demand.
XConn Technologies and MemVerge showcased a 100TiB Compute Express Link (CXL) memory pool integrated with NVIDIA’s Dynamo architecture at the 2025 OCP Global Summit, demonstrating significant performance gains for AI inference workloads.
The Open Compute Project Foundation and iMasons Climate Accord have released a new specification to standardize embodied carbon reporting across the data center industry, developed with contributions from Google, Meta, Microsoft, and Schneider Electric.
The SONiC Foundation announced new milestones at the OCP Global Summit 2025, including Nexthop AI advancing to Premier membership and multiple new AI infrastructure deployments across industries.
Axiado Corporation will demonstrate its autonomous AI agents embedded in silicon at the 2025 OCP Global Summit in San Jose, highlighting technologies for platform security, system efficiency, and infrastructure management.
Jabil has announced its new J-422G servers designed to handle high-performance AI, machine learning, and data center workloads. The servers, powered by sixth-generation Intel Xeon processors, will be generally available in November 2025.
ABB and NVIDIA have announced a collaboration to develop advanced AI data centers with high-efficiency power delivery systems. ABB will focus on scalable electrification solutions to support NVIDIA’s 800-volt architecture for faster, more efficient energy transfer.
Advantech has introduced the AOM-5721, a SMARC-based edge AI computer-on-module featuring Qualcomm’s Dragonwing QCS6490 SoC, designed for industrial, commercial, and medical applications.
Google announced a $15 billion investment to establish India's first AI hub and a gigawatt-scale data center campus in Visakhapatnam, Andhra Pradesh, in partnership with AdaniConneX and Airtel. The project will expand digital infrastructure, clean energy capacity, and subsea connectivity over the next five years.
ASUS unveiled its AI Factory and XA NB3I-E12 AI servers powered by NVIDIA HGX B300 systems at the 2025 OCP Global Summit, highlighting solutions for enterprise and cloud AI workloads.
GIGABYTE has announced that its AI TOP ATOM personal supercomputer, featuring the NVIDIA Grace Blackwell GB10 Superchip, will be available globally from October 15. The compact system delivers up to 1 PetaFLOP of FP4 AI performance and supports models with up to 200 billion parameters.
Wiwynn introduced a double-wide rack architecture at the OCP Global Summit 2025 to support larger, high-power AI accelerators with integrated HVDC power and advanced liquid cooling systems.
Compal Electronics presented new CXL-based memory pooling and liquid cooling server technologies at the 2025 OCP Global Summit, highlighting its collaboration with NVIDIA and focus on scalable AI data center infrastructure.
Weekly coverage of AI hardware developments including chips, GPUs, cloud platforms, and data center technology.
