The neuromorphic semiconductor market is expanding rapidly, driven primarily by surging demand for ultra-low-power efficiency.
Sean Hehir, CEO of BrainChip, shared this optimistic outlook on the future of neuromorphic computing in a recent interview with Electronic Times.
BrainChip is widely recognized as a pioneer in the commercialization of neuromorphic chips, a technology that had long been confined to research laboratories. Neuromorphic semiconductors mimic the human brain's neural networks--specifically neurons and synapses. By integrating processing and memory into a single architecture, they eliminate the need for constant data transfer between separate processing and memory units. Like the human brain, these chips activate only when triggered by a stimulus or a change in data, allowing them to operate at less than one-tenth the power consumption of conventional chips.
“Traditional AI accelerators must keep the entire neural network active to process massive amounts of data, regardless of whether meaningful information is present,” Hehir explained. “In contrast, Akida--BrainChip's flagship neuromorphic architecture--is event-based. Similar to the human brain, it activates only when a meaningful change occurs.”
Hehir added that during “silent” periods when no new data is detected, the chip consumes virtually no power. Furthermore, by largely eliminating the data bottleneck between processor and memory, the architecture enables ultra-low-power operation.
Thanks to this exceptional power efficiency, neuromorphic chips are gaining traction as ideal hardware for Edge AI--where artificial intelligence is processed directly on the device. Target applications include power-constrained smart devices such as heads-up displays (HUDs) like Apple's Vision Pro, Meta's smart glasses, humanoid robots, smart security cameras, and advanced mobility platforms. According to BrainChip, Akida is already being adopted across a range of sectors, including wearables, automotive, and smart sensors.
Market research indicates that the global neuromorphic computing market is expected to grow from $28.5 million in 2024 to $1.325 billion by 2030, representing a compound annual growth rate (CAGR) of 89.7%. Reflecting this momentum, BrainChip's revenue is forecast to increase from $400,000 in 2024 to $1.9 million in 2025, according to the company.
BrainChip has already established a foothold in the South Korean market by licensing its intellectual property (IP) to EdgeAI, a leading domestic fabless semiconductor company. EdgeAI plans to fabricate a proprietary chip based on BrainChip's IP to power smart metering systems used in remote utility inspection.
Building on this momentum, Hehir emphasized his commitment to expanding partnerships with South Korean enterprises.
“South Korea possesses world-class semiconductor manufacturing capabilities and a robust hardware ecosystem,” Hehir said. “We plan to bolster our technical support to help promising Korean fabless companies integrate Akida IP into their systems-on-chip (SoCs), enhancing their global competitiveness. We are also exploring open-innovation initiatives to co-develop industry-specific reference designs for sectors such as consumer electronics and automotive electronics.”
