4DS - Anything but Charting, page-20262

A little solace for suffering long term shareholders. There has been much conjecture as to the upper value of the IP of 4DS regardless of the performance of the mbit chip achieved. Some have said this value is capped at $150m with no further upside.

As such, I reached out through a trustworthy contact to a respected commentator / expert in the memory space with a direct question.

What would the IP of a mbit chip with the following performance characteristics be worth to a memory maker (who then faced an additional 2 years development / costs from the mbit chip to a commercial product).

The chip performance I stated was:
- Yield of > 97% or as per DRAM
- Speed 5X faster than DRAM (i.e. we were comparable to DRAM before, and now we are stated to be up to an order of magnitude faster. I also note in the recent interview Wilbert stated we are aiming to replace and augment DRAM, which wouldn’t happen unless we are at least 2-3X faster than DRAM, due to the time / cost / risk involved. Further, the compromise Wilbert mentioned was between retention and endurance, not speed.)
- Memory cell size of 5 nm (speculation of 2-3nm as this may be the lowest cell size supported at imec, but 5nm is known and more probable.)
- Retention of 2 days (i.e. 48 hours, as per Wilbert interview)
- Power saving of 10^3 less than DRAM (in the 2017 Sydney presentation, we were told its 1 million times less power, but the switching current has since been increased so assuming 10^3)
- Endurance of 10^10 cycles (i.e. giving a useful life of 19 years if cycled every 60 milliseconds as per currently with DRAM. The last pronounced endurance is only 10^9 – but I am hoping it is now 10^10. This is critical, as 10^9 is 1.9 years of life only which is too soon to replace memory in data centers, and probably why even a small decrease in endurance becomes a material announcement.)

The response from the expert was:
- These types of metrics are impossible.
- IF it existed, it would replace all other memory (I don’t think he was speaking of NAND here)
- IF it existed and was patentable, the IP would be worth billions.

The problem is, experts see 4DS as a competitor to other technologies like Weebit. This type of solution would have an IP value of $150m or so, but this is not the market the company is going for. If you believe the company that is.

I know there is some consternation too about no longer defining / chasing ‘storage class memory’. In my opinion, in publicly stating we are trying to replace DRAM, this is even better than achieving ‘near DRAM’ speeds in a new market.

Anyway, as wally used to say – shit, or get off the pot. I have put additional capital to work.

Just thought some of the other longs would find the above expert’s response useful.

I also note Wilberts experience and note that he wouldn't waste his time with a startup like ours & make public comments if the solution was not solving a next gen problem.