Exactly, the Dory platform wasn't a failure - it helped them...

Exactly, the Dory platform wasn't a failure - it helped them identify a capability that wasn't indicated by the company before. That is - variable resistive programming, that could pave the way to analogue memory. That's surpassing the expectations for 4DS RERAM.

So in the immediate term, the company has a very fast, persistent memory that can help with current datacentre problems - that is, having to store loads of data in RAM, where it consumes a lot of energy and generates a lot of heat.

Over the longer term, the tech could become a AI memory option with the analogue option. You could also write a hexadecimal character (byte) on a single cell, rather than using 8 cells to achieve the same thing in binary, as long as they can write 256 identifiable levels of resistance onto a single cell.

That increases memory density by 8 times just with that without resorting to shrinking cell size. Also less energy again because you aren't writing and reading 8 cells, just one cell to get the same amount of data.

I hope they find a way to sell the binary capability of the cell, whilst retaining rights to the IP for the analogue/ variable resistance capability.

Then take the cash from the first deal to fund the bigger, more impactful (imo) variable resistance development program.