Avicena Partners with TSMC to Advance Photodetector Arrays for LightBundle Interconnects

Avicena is teaming up with TSMC to optimize photodetector arrays for the microLED-based LightBundle interconnect, aiming to power the next generation of Artificial Intelligence infrastructure with faster, denser, and more energy-efficient connections.

Avicena, a company based in Sunnyvale, California, announced a collaboration with Taiwan Semiconductor Manufacturing Company (TSMC) to optimize photodetector (PD) arrays that are integral to Avicena´s LightBundle microLED-based interconnect technology. LightBundle is designed to deliver over 1 terabit per second per millimeter shoreline density and enable ultra-high-density die-to-die (D2D) connections that can extend up to ten meters with best-in-class energy efficiency of sub-picojoule per bit.

This innovation dramatically surpasses the capabilities of legacy copper-based interconnects, both in reach and in reduced power consumption. The advanced photodetector arrays powered by TSMC´s process technology will allow data to move quickly and efficiently between multiple racks populated with high-performance GPUs, which are increasingly critical for large-scale Artificial Intelligence workloads.

The rapid progress of Artificial Intelligence models is fueling unprecedented demands on compute and memory infrastructure. Modern systems now require interconnects that offer greater bandwidth density, lower power consumption, and a longer reach to support both processor-to-processor (P2P) and processor-to-memory (P2M) communications. By leveraging TSMC’s manufacturing expertise, Avicena aims to scale up networks to handle growing clusters of GPUs, effectively addressing the connectivity bottlenecks that limit current Artificial Intelligence data center architectures.

73

Impact Score

AMD claims EPYC lead in agentic Artificial Intelligence workloads

AMD is using rack-level benchmarks to argue EPYC CPUs will remain central to agentic Artificial Intelligence infrastructure. The claims target Nvidia’s Vera platform and Intel’s Xeon lineup as data centers rebalance around CPU-heavy orchestration work.

Hades variant affects 23 PyPI package versions

The Mini Shai-Hulud Hades variant is targeting PyPI packages tied to bioinformatics and Artificial Intelligence themes. Socket researchers say the malware uses Python startup hooks and compiled extensions to run a JavaScript stealer.

DiffusionGemma rethinks text generation with diffusion

DiffusionGemma applies diffusion-style denoising to text, trading autoregressive token-by-token decoding for iterative canvas refinement. Its design combines encoder guidance, bidirectional denoising, scheduling, and entropy-based sampling.

NVIDIA shows RTX Spark platform at Computex 2026

NVIDIA presented RTX Spark in Taipei as a Windows on Arm platform spanning gaming, creator, and Artificial Intelligence workloads. Microsoft also detailed Windows 11 optimizations built specifically for the new NVIDIA silicon.

Contact Us

Got questions? Use the form to contact us.

Contact Form

Clicking next sends a verification code to your email. After verifying, you can enter your message.