2024 BrainChip Discussion, page-8178

So, we have 67 great organizations looking at us. Great.

Now what could be possibly going wrong that we have no revenue?

Potential customers are afraid of being stuck with obsolete chips if newer better ones come out.
Potential customers won't do the data analysis to go from CNN models to SNN models.
Potential customers want to go into AI with the major players only.
There are no programmers around to do the work..

The current BRN solution is to have Akida IP available as a cloud service running on FPGA chips.
That way potential customers and uni students can get their hands dirty and model their data using Akida.
If they like it they can go off and talk to Renesas or Megachips.
But is this still reasonable?
We've already designed 8 akida chips for potential customers, and they don't seem to have continued to a license.

So, why don't we release Akida-as-software running on those same FPGA chips ( eg PolarFire etc) on PCIE cards in our computers?
Maybe in this way we can bypass the killer step of needing the customer to go off and committ to making millions of chips. Once they have modeled their data successfully,they can start off making their products by buying COTS FPGA chips instead.

I'd like to ask Tony this but not sure I have this all right. Comments please anyone?