GPU– tag –
-
NTT’s IOWN AI Fund Makes Inaugural Investment in iPronics for AI Data Center Programmable Optical Networking
The Fast Mode Japan Overview NTT's IOWN AI Fund has made its first investment in iPronics, a developer of programmable optical networking technology for AI data centers. iPronics' silicon photonics-based platform dynamically reconfigures... -
GIGALIGHT Unveils 18W Ultra-Low Power 1.6T AI Interconnect at ECOC 2026, Signaling Independence from Liquid Cooling
GIGALIGHT Press Releases China Overview GIGALIGHT showcased a hybrid technology at ECOC 2026 for 1.6T AI interconnects, achieving a significantly reduced power consumption of 18W compared to mainstream 3nm full-DSP solutions. This innova... -
Silicon Photonics Dominates 1.6T Optical Transceivers, Targeted for Broad AI Data Center Deployment by 2026
AICP Light USA Overview Silicon photonics is rapidly becoming the leading technology for 1.6T transceiver optics, enabling higher per-lane data rates and improved power efficiency through CMOS-compatible manufacturing. The optical commun... -
AI Workloads Drive Data Centers to Liquid Cooling, NVIDIA Optimizes Next-Gen Servers for Liquid Immersion
Connect CRE USA Overview The explosion of AI workloads is forcing data centers to adopt liquid cooling solutions as traditional air cooling can no longer manage the intense heat. Major data center providers like Equinix and Digital Realt... -
NVIDIA Emerges as “Central Bank of AI,” Dominating Industry Capital Flows with Over $70 Billion in Investments
BigGo Finance USA Overview NVIDIA is positioned by The Economist as the "central bank of AI," commanding capital flows in the industry through investments exceeding $70 billion in startups and $300 billion in financial aid to customers o... -
LLM Test-Time Scaling: Candidate Generation Strategy Dramatically Increases Energy Consumption by 4.86x and Latency by 6.12x on A100 GPUs for Phi-3-mini and Qwen2.5-1.5B
Mobina Kashaniyan (referencing IEEE/ACM SC26 Workshop) Unknown Overview New research reveals that for LLM test-time scaling, the candidate generation strategy, not just the candidate count, critically impacts energy and performance. Eval... -
NVIDIA to Boost CoWoS Production Capacity by 3.5x to Over 120,000 Wafers/Month by End-2026, Securing HBM4 Supply
SemiconductorX USA Overview NVIDIA plans to scale its CoWoS (Chip-on-Wafer-on-Substrate) production capacity by 3.5x to 120,000–130,000 wafers per month by the end of 2026, up from 35,000, driven largely by its own demand. The company is... -
Semtech Unveils 224G Linear TIA Driver Portfolio for AI-Driven NPO/CPO Solutions
Converge Digest USA Overview Semtech announced a new portfolio of 224G linear Transimpedance Amplifier (TIA) drivers for Near-Packaged Optics (NPO) and Co-Packaged Optics (CPO) solutions, addressing the demands of AI applications. These ... -
Newphotonics Unveils 6.4Tbps Laser-Integrated NPO Engine for AI Networks, Compliant with Open CPX
Converge Digest Unknown Overview Newphotonics has unveiled a 6.4Tbps laser-integrated Near-Packaged Optics (NPO) engine designed for AI networks. This innovative optical engine aims to resolve data processing bottlenecks and significantl... -
FS Delivers 400G AI Compute Fabric Solution to Unleash Full Potential of NVIDIA B200/H200 GPU Clusters
FS Community Unknown Overview As AI models grow in complexity, GPU clusters demand high-bandwidth, low-latency communication between servers. FS is offering a 400G AI compute fabric solution designed to maximize the potential of NVIDIA B...
