Ann: Weebit Nano and the Technion collaborate on ReRAM project, page-26

The AI market size for WBT doesnt matter if you can get to NVM product ubiquity. Any other moon to shoot for is a cop out for mine.

If your Non-Volatile Memory (NVM) can compute or be used for AI as well as for general purpose due to scalability and cost of manufacture then it matters little what the AI market is worth to WBT. That will be the technology that is used for NVM going forward.

We have an NVM with low manufacturing cost on standard processes in high yield. If it pushes down toward the comparative cost to Flash for a given level of size and can be used in various other scenarios as adjunct processing or synapse replication, then it becomes ubiquitous as the part you just grab off the shelf or model with in tCAD for most NVM scenarios.

You want NVM for anything, you use this. That should be the ultimate game here, anything that needs data retention, they use WBT reram, AI or otherwise.

Simples.