Intel recently unveiled Hala Point, the world's largest neuromorphic system, the first such system to feature 1.15 billion neurons, powered by 1,152 Loihi 2 Chips. Hala Point will now be utilised by the Sandia National Laboratories for advanced brain-scale computing research and will optimise telecom infrastructure efficiency through Ericsson Research's application of Loihi 2 chips.

Neuromorphic technology offers an optimal solution to the challenges of Artificial Intelligence (AI) and sustainability, greatly improving efficiency and performance, thus broadening the reach of conventional deep learning models. Hala Point represents a significant enhancement of Intel's former large-scale research system, Pohoiki Springs, offering a neuron capacity more than ten times greater, and a performance up to twelve times higher.

Hala Point represents a landmark in neuromorphic systems, delivering efficiency and performance exceeding that of contemporary GPU and CPU architectures when handling real-time AI workloads. Emerging small-scale edge workloads have particularly benefited from the efficiency, speed and adaptability gains made possible by Loihi 2, according to recent results.

While Hala Point is primarily a research system, Intel envisages that lessons learned from its use will prompt the development of future commercial systems with practical breakthroughs in AI technology. Large Language Models (LLMs) could learn continuously from new data, greatly reducing the training burden of widespread AI deployment. Loihi-based systems could enable advancements in logistics, scheduling, scientific computing, engineering, and explainable AI, learning associations and solving optimisation problems up to 50 times faster and with 100 times less energy.

"The computing cost of today's AI models is rising at unsustainable rates. The industry needs fundamentally new approaches capable of scaling. For that reason, we developed Hala Point, which combines deep learning efficiency with novel brain-inspired learning and optimisation capabilities. We hope that research with Hala Point will advance the efficiency and adaptability of large-scale AI technology," stated Mike Davies, director of the Neuromorphic Computing Lab at Intel Labs.

Hala Point applies brain-inspired computing principles to its operation, achieving significant gains in energy consumption and performance. Its unique capabilities are seen as enabling continuous learning for future real-time AI applications, such as in logistics, smart city infrastructure management, large language models (LLMs), and AI agents. The system's neuron capacity, equivalent to that of an owl brain or a capuchin monkey's cortex, allows it to execute its full capacity of 1.15 billion neurons 20 times faster than a human brain and up to 200 times faster at lower capacity. However, Hala Point is not intended for neuroscience modelling.

"Working with Hala Point improves our Sandia team's capability to solve computational and scientific modelling problems. Conducting research with a system of this size will allow us to keep pace with AI’s evolution in fields ranging from commercial to defence to basic science," added Craig Vineyard, Hala Point team lead at Sandia National Laboratories.

This is the first of a new family of large-scale neuromorphic research systems that Intel plans to share with research collaborators. The development aims to help neuromorphic computing applications overcome the power and latency constraints that currently limit the real-time deployment of AI capabilities.