2026– date –
-
Marvell: 2nm silicon photonics 3.2Tbps specs for ECOC 2026
remio AI USA Overview Marvell announced a demonstration at ECOC 2026 of groundbreaking 400 gigabits/lane PAM4 and 3.2 terabit optical modules, built on its industry-leading 2nm silicon photonics technology. The company has already proven... -
Racyics UCIe PHY IP: GlobalFoundries 22FDX silicon validation
Chiplet Marketplace Germany Overview Racyics announced the successful silicon validation of its Universal Chiplet Interconnect Express (UCIe) PHY IP for GlobalFoundries' 22FDX process. UCIe provides a common interface for efficiently con... -
Intel MicroLEDs: 10-meter optical interconnects for AI
Tom's Hardware USA Overview Intel has secured a U.S. patent for embedding MicroLEDs directly into chip packages. This innovative design facilitates optical signal transmission by integrating MicroLED-equipped semiconductor dies into a gl... -
Canon and NTT: Real-time remote sports viewing via IOWN APN
Canon Japan Overview Canon, in collaboration with NTT's IOWN All-Photonics Network (APN), successfully conducted the world's first real-time remote sports viewing demonstration by integrating its volumetric video technology and MREAL mix... -
Co-Packaged Optics explained: 400 Tbps data center capacity
AZoOptics Unknown Overview Co-packaged optics (CPO) is an emerging technology addressing energy challenges in data centers by integrating photonics directly with processors, enhancing performance and efficiency. Broadcom's 51.2-Tb Bailly... -
Yale University Nd Laser: Chip-scale MOPA system specs
Photonic Integrated Circuits News (PIC Magazine) USA Overview Researchers at Yale University have developed a chip-scale Neodymium-doped Yttrium Aluminum Garnet (Nd) laser amplifier system using a master oscillator power amplifier (MOPA)... -
Genesis chip: UT San Antonio neuromorphic continuous learning
UT San Antonio Today USA Overview The MATRIX AI consortium at UT San Antonio has unveiled the "Genesis" spike neuromorphic accelerator chip, which solves the long-standing AI problem of catastrophic forgetting, enabling continuous on-dev... -
AI Scientist: Autonomous yeast growth compound design and specs
ZME Science Sweden Overview Researchers at Chalmers University of Technology have developed an AI system capable of autonomously generating scientific hypotheses, designing and executing experiments with automated lab equipment, and inte... -
Li–P–S crystals: AI strategy identifies 10,000 stable candidates
arXiv USA Overview Researchers have developed an iterative strategy combining crystal-structure generative models with machine-learned force fields (MLFFs), successfully producing approximately 70,000 candidate structures for the Li–P–S ... -
ProcGNN explained: PyTorch physics-informed GNN for simulation
ProcGNN South Korea Overview A new paper, "ProcGNN," introduces a PyTorch-based physics-informed graph neural network for stream-level chemical process surrogate modeling. ProcGNN uses shared unit and stream representations to predict te...
