At every turn we see artificial intelligence (AI) being discussed, the large language models (LLM) and the generative AI engines, be they ChatGPT, Co-Pilot, DeepSeek, and many others. What we don’t see a lot about is the status of AI chips and why they are so important to the U.S. Department of Defense, specifically the Defense Intelligence Agency (DIA). In October 2024, at the Department of Defense Information Systems Worldwide Conference (DODIIS 2024), Dr. Ramesh Menon, Chief Technology Officer and Chief AI Officer, made mention of an AI chip designed to mimic the brain. The project designing this chip is North Pole and is the epitome of a successful public-private partnership.
Clearance Jobs recently caught up with Dr. Menon to discuss North Pole and how this project, which saw its genesis in DARPA’s SyNAPSE (Systems of Neuromorphic Adaptive Plastic Scalable Electronics). DARPA tells us, “The program’s focus is to develop low-power neuromorphic computers that scale to the biological levels.” IBM, for its part, has been working together with DARPA, the U.S. Air Force Research Lab, and the DoD for about 15 years to create such a chip.
In September 2024, the breakthrough occurred and was shared at the IEEE HPEC by Dhamendra Modha of IBM. The chip was at the 12 nanometer (nm) scale. Menon noted how 4 nm has been achieved in the United States. The achievable goal of a 2nm fabricated chip will be a national security advantage, and according to Menon, it is within sight.
Why the North Pole 2nm AI chip matters
Menon noted how AI’s power has long been constrained by one thing: hardware limitations. Software can only take you so far when the chips running isn’t built for real-time intelligence and autonomous operations. That’s where IBM’s North Pole AI chip comes in a neuromorphic processor designed to mimic the efficiency of the human brain while operating at a fraction of the power consumption of traditional GPUs.
Unlike traditional chips, which move data back and forth between memory and processors (a process that wastes energy and slows computations), North Pole integrates memory and computation on the same chip. The result? Ultra-fast, energy-efficient AI processing is ideal for military, intelligence, and cybersecurity applications
Emphasized at DODIIS 2024 and again during our 1:1, the future will be AI-driven warfare and intelligence analysis demand more than just powerful algorithms; they demand AI hardware that can keep up. This shift, which currently only the United States has achieved, truly is a differentiator in national security. The implications? Faster intelligence processing, autonomous decision-making at the tactical edge, and a more secure national defense strategy.
The future is faster, smarter, and untethered
Faster, smarter, and untethered from cloud dependencies is the goal, according to Menon for the AI needed by the warfighter. Defense operations require low-latency AI processing that can function autonomously in contested environments. North Pole’s architecture enables just that high-speed, low-power AI inference without reliance on external computing infrastructure.
