Feb. 26 at 8:32 PM
$NVDA When NVDA dropped CUDA, GPUs went from graphics engines to programmable compute beasts, the platform that made deep learning take off.
Neuromorphic chips (
$INTC Loihi 2,
$IBM Northpole/TrueNorth,
$BRCHF Akida, Spinncloud) are at that same pre‑CUDA stage imo. Lots of innovation, few standards. Each chip has its own neuron model, spike logic and toolchain.
Frameworks like Intels Lava, snnTorch, and Norse are bridging the gap giving the field the software glue it still lacks. See who’s building here if interested.. neuromorphiccore.ai/companies.
The missing piece is the killer workload that forces standardization.
My bet’s on low‑power edge AI, event‑based vision and real time adaptive systems. Axelera’s
$250M raise says plenty.
The CUDA moment for neuromorphic isn’t here yet, but you can feel it coming. Funny how few watched
$IONQ in single digits. Quantum and neuromorphic computing may be the tech pillars that thrive even as the software companies of the world struggle under Anthropics grip.
