Student Team Designs Energy-Efficient AI Architecture
The MemoryFlow AI project reduces power usage by processing data closer to memory in a 3D-stacked configuration.
Updated on Oct. 22, 2026 in Semiconductors
Live Poll
Do you believe university students should prioritize sustainable technology innovation to address global energy challenges?
Swinburne University of Technology Sarawak students won the grand prize at the SEMI TECH Zoomers Bootcamp 2.0 held in Singapore from September 21 to 24, 2026. The team was recognized for their MemoryFlow AI project, an architecture designed to improve chip energy efficiency.
Why it matters
The project addresses a persistent bottleneck in semiconductor design: the energy cost of moving data between memory and processing units. By demonstrating a research-stage approach to minimizing this traffic, the team highlights current efforts to optimize power consumption in AI-focused hardware.
The MemoryFlow AI project uses 3D-stacked memory, reusable compute chiplets, and compute-near-memory processing. These techniques aim to reduce overall energy consumption by minimizing the distance data must travel between storage and processing units.
The players
Felix Thian Zhi Xiang
A third-year Electrical and Electronic Engineering student at Swinburne University of Technology Sarawak who led the prize-winning team.
Swinburne University of Technology Sarawak
A branch campus focused on research in engineering and technology, located in Kuching.
The details
The MemoryFlow AI architecture relies on compute-near-memory processing, a technique that places logic circuits adjacent to memory storage to reduce data transit. By utilizing 3D-stacked memory—a design where memory layers are layered vertically—the architecture increases bandwidth density. The project further incorporates reusable compute chiplets, which are modular integrated circuit components designed for assembly into larger systems to improve manufacturing efficiency.
Timeline
May 2026: Initial selection bootcamp held in Kuala Lumpur.
September 21-24, 2026: SEMI TECH Zoomers Bootcamp 2.0 held in Singapore.
The Tech Race
The SEMI TECH Zoomers Bootcamp 2.0 acts as a talent incubator for the global semiconductor industry, focusing on next-generation efficiency. The event benchmarks student designs against the ongoing industry shift toward specialized AI-centric architectures.
This project represents an academic research effort into improving AI hardware power efficiency rather than a consumer-facing product launch. The design remains in the conceptual stage, with no immediate timeline for integration into commercial silicon chips.
The takeaway
The MemoryFlow AI project signals an industry trend toward reducing the energy footprint of AI processing by moving compute closer to memory. Readers should watch for future technical whitepapers or public benchmarks from the participants to determine if this architecture can scale beyond academic competition.
Further reading
Explore the latest developments in semiconductor engineering by visiting the Semiconductors section.
Source note: This article includes information reported by New Sarawak Tribune.
Live Poll
Do you believe university students should prioritize sustainable technology innovation to address global energy challenges?
