Pricing depends entirely on the nature of the license fee...

Pricing depends entirely on the nature of the license fee model.

I believe BRN would rarely sell a chip in isolation ("Here you go, here's your chips. Have fun with them.")

For satellite use, I would guess the application would need...

Custom silicon: in other words a radiation hardened ASIC (a new chip) that bundled Akida with other modules. BRN would allow TSMC to marry the Akida tape-out with the other modules to produce the test prototypes and final production runs.

Technical design assistance: BRN engineers would assist the customer to design software, hardware integration, and simulation methods needed for the various system modules to communicate within the ASIC IP bundle and be tested.

Given the low production volumes, compared to the value of the inputs (access to hardware IP and technical assistance) its highly probable BRN would have negotiated a MUCH higher unit price.

The general model for such licensing is based on pricing that takes into account the seller and buyer's valuation of...

1. Buyer's cost saving and faster speed to market of buying off the shelf IP versus developing the IP themselves (Make or Buy? is the decision).
2. Value of Seller's engineering inputs to assist the Buyer to get it all to work.
3. Technical value proposition of the seller's IP versus competitors.
4. Risk premium. The risk of the seller never seeing any license revenue because the customer fails to complete.
5. Discounted cash-flow of forecast license revenue stream compared to lump sum upfront (there would be a formula but in summary it would be a probability multiplied by a consensus best guess).

So, the price negotiation would also include the proportion between an up-front payment versus a license revenue stream.

We have no idea what the price of the IP is but we can assume that it's different for every deal.

IMHO