OOOooh...
I just stumbled across this LinkedIn post about neuromorphic chips from Soumya Sagiri (Principal Gen AI Privacy Security Strategist & Engineering @ AWS).
IMO her post signals that neuromorphic hardware is the logical next step once current edge-to-cloud security and privacy workflows hit their practical limits.
She reframes neuromorphic computing as a security and privacy enabler, not just a power-efficiency experiment, but she notes that “neuromorphic hardware is still mostly in research and pilot stage today (e.g., Intel Loihi, IBM TrueNorth).”
I’m thinking I’ll email this to Tony Lewis tomorrow to see if he’d like to invite Soumya to try Akida Cloud, which allows hardware-free testing of BrainChip’s Akida 2 platform.
Soumya also says: “In the future, neuromorphic chips could act as a real-time security co-processor… flagging anomalies in milliseconds.” To me, that makes it pretty explicit - the future state of this architecture is neuromorphic, with mainstream adoption once hardware is production-ready.
And take a look at the comment below from Rudy Bakalov (Cloud/Cybersecurity/AI exec, AWS)!
Rudy says: “Neuromorphic chips will change the game once they move from pilot to production,” and he calls out on-device agentic AI as the first breakout use case, arguing neuromorphic creates a new security paradigm.
For clarity, on-device agentic AI = a self-contained intelligent agent that can sense its environment, decide, and take action in real time- right on the device.
Might as well ask Tony if he'd like to send Rudy an invite too while I'm at it...
View attachment 92346
View attachment 92348
View attachment 92347
View attachment 92351
Neuromorphic chips are processors that mimic neurons and synapses using event-driven signals. This is not only about power savings. It also opens a new way to build security and privacy controls directly into hardware. Neuromorphic hardware is still mostly in research and pilot stage today, for...
www.linkedin.com