What a difference a year makes. The first tape out, in my mind was "sure hope it works out". That was another milestone Brainchip passed beautifully.
No concern this time, plus+ the portability/flexibility to do this now with Global Foundries is awesome.
The whole Spiking Neural Network trend, slow and steady taking down Von Neumann is also derisked 100%. The SNN BRN blast is at the very beginning.
The features focused on with AKD1500 have to be focused on "quicker to cash wins". Meaning, perhaps focus on the customers that are ready to hit the market faster. Some of the Brainchip staff mentioned that before, putting resources towards sooner rather than later wins.
Low power edge IoT might be faster to market over LSTM/Transformers, which is likely 2024 as
@Diogenese has formulated.
Whatever the staff is doing, I bet it is for faster to market wins.