Hi DB,
Going ubiquitous will take time.
If we are working with ARM, Intel, Nvidia, Tata ..., they have in-house NN expertise and may not need too much support, but providing the support they need will be critical to success.
Then there are companies like Valeo, Prophesee, etc, who have a fair understanding of NNs, but not to the same level as the first group, so they will need additional support. We have joint development projects with many of these.
After them, there are those who understand the advantages of NNs, but do not have any real in-house expertise, and these will require much more support.
BRN will no doubt be guarding against spreading itself too thinly, as it would be difficult to find sufficient staff with the requisite expertise. The BrainChip University program will in time go some way to addressing the shortfall and spreading the skills throughout industry.
I haven't seen any more recent employment numbers or the breakdown of tasks since we hit about 80.
A lot of the support work will be in providing appropriate data models for the customer projects. This may involve converting existing CNN models to SNN (CNN2SNN), and tweaking these models for the specific purpose, or it may involve producing new models which is a quite intensive undertaking.
When we achieve ubiquity, the customers will be producing their own SNN models.