Researchers are tackling key 6G challenges—from ultra-fast photonic AI for real-time RF processing to novel GaN amplifier behavior and ultra-compact phased-array packaging.
Jun. 27, 2025 –
Researchers are racing to deploy 6G with clear goals in mind: speeds exceeding one Tbps, sub-millisecond latencies, AI-native architectures, and seamless integration across the physical and digital worlds. Yet the technical barriers are immense. Spectrum scarcity, thermal constraints, RF front-end inefficiencies, and a need for miniaturized, high-gain antennas threaten to bottleneck progress.
This roundup highlights three university-led research efforts that target foundational problems in 6G design, focusing on the intersection of speed, efficiency, and scale. Whether by rethinking photonic processing, exploiting novel RF phenomena in semiconductors, or compressing entire beamforming systems into millimeter-scale modules, each team brings a new layer of feasibility to the 6G vision.
MIT Demonstrates Photonic AI Processor
MIT researchers have developed a photonic deep learning processor. Named MAFT-ONN (Multiplicative Analog Frequency Transform Optical Neural Network), it performs fully analog deep learning directly on raw radio frequency (RF) signals. As 6G systems edge toward terabit-per-second data rates, digital bottlenecks in RF front ends present a major hurdle. The MAFT-ONN system sidesteps this by combining frequency-domain encoding, photoelectric multiplication, and electro-optic nonlinearities to enable real-time, low-latency signal classification and spectrum analysis...