Leveraging silicon photonics for scalable and sustainable AI {hardware}

April 10, 2025

The GIST Editors' notes

This text has been reviewed in line with Science X's editorial course of and insurance policies. Editors have highlighted the next attributes whereas guaranteeing the content material's credibility:

fact-checked

proofread

Leveraging silicon photonics for scalable and sustainable AI {hardware}

The emergence of AI has profoundly remodeled quite a few industries. Pushed by deep studying expertise and Large Knowledge, AI requires vital processing energy for coaching its fashions. Whereas the present AI infrastructure depends on graphical processing items (GPUs), the substantial processing calls for and power bills related to its operation stay key challenges. Adopting a extra environment friendly and sustainable AI infrastructure paves the way in which for advancing AI growth sooner or later.

A latest research revealed within the IEEE Journal of Chosen Subjects in Quantum Electronics demonstrates a novel AI acceleration platform based mostly on photonic built-in circuits (PICs), which provide superior scalability and power effectivity in comparison with typical GPU-based architectures.

The research, led by Dr. Bassem Tossoun, a Senior Analysis Scientist at Hewlett Packard Labs, demonstrates how PICs leveraging III-V compound semiconductors can effectively execute AI workloads. Not like conventional AI {hardware} that depends on digital distributed neural networks (DNNs), photonic AI accelerators make the most of optical neural networks (ONNs), which function on the velocity of sunshine with minimal power loss.

"Whereas silicon photonics are straightforward to fabricate, they’re tough to scale for advanced built-in circuits. Our machine platform can be utilized because the constructing blocks for photonic accelerators with far larger power effectivity and scalability than the present state-of-the-art," explains Dr. Tossoun.

The workforce used a heterogeneous integration method to manufacture the {hardware}. This included using silicon photonics together with III-V compound semiconductors that functionally combine lasers and optical amplifiers to cut back optical losses and enhance scalability.

III-V semiconductors facilitate the creation of PICs with larger density and complexity. PICs using these semiconductors can run all operations required for supporting neural networks, making them prime candidates for next-generation AI accelerator {hardware}.

The fabrication began with silicon-on-insulator (SOI) wafers which have a 400 nm-thick silicon layer. Lithography and dry etching have been adopted by doping for steel oxide semiconductor capacitor (MOSCAP) gadgets and avalanche photodiodes (APDs).

Subsequent, selective development of silicon and germanium was carried out to type absorption, cost, and multiplication layers of the APD. III-V compound semiconductors (comparable to InP or GaAs) have been then built-in onto the silicon platform utilizing die-to-wafer bonding. A skinny gate oxide layer (Al₂O₃ or HfO₂) was added to enhance machine effectivity, and at last a thick dielectric layer was deposited for encapsulation and thermal stability.

"The heterogeneous III/V-on-SOI platform offers all important parts required to develop photonic and optoelectronic computing architectures for AI/ML acceleration. That is notably related for analog ML photonic accelerators, which use steady analog values for knowledge illustration," Dr. Tossoun notes.

This distinctive photonic platform can obtain wafer-scale integration of the entire varied gadgets required to construct an optical neural community on one single photonic chip, together with lively gadgets comparable to on-chip lasers and amplifiers, high-speed photodetectors, energy-efficient modulators, and nonvolatile part shifters. This permits the event of TONN-based accelerators with a footprint-energy effectivity that’s 2.9 × 10² occasions larger than different photonic platforms and 1.4 × 10² occasions larger than essentially the most superior digital electronics.

That is certainly a breakthrough expertise for AI/ML acceleration, lowering power prices, bettering computational effectivity, and enabling future AI-driven functions in varied fields. Going ahead, this expertise will allow datacenters to accommodate extra AI workloads and assist resolve a number of optimization issues.

The platform will probably be addressing computational and power challenges, paving the way in which for strong and sustainable AI accelerator {hardware} sooner or later.

Extra data: Bassem Tossoun et al, Massive-Scale Built-in Photonic Machine Platform for Vitality-Environment friendly AI/ML Accelerators, IEEE Journal of Chosen Subjects in Quantum Electronics (2025). DOI: 10.1109/JSTQE.2025.3527904

Offered by Institute of Electrical and Electronics Engineers Quotation: Leveraging silicon photonics for scalable and sustainable AI {hardware} (2025, April 10) retrieved 10 April 2025 from https://techxplore.com/information/2025-04-leveraging-silicon-photonics-scalable-sustainable.html This doc is topic to copyright. Other than any truthful dealing for the aim of personal research or analysis, no half could also be reproduced with out the written permission. The content material is offered for data functions solely.

Discover additional

Photonic chips increase computing velocity and effectivity to deal with rising demand shares