2021 BRN Discussion, page-6363

Lots of new faces here so maybe some old facts may be of use:

1. The AKD1000 produced as engineering samples last year had according to the CTO a target price of about $US15.00.

2. The AKD1000 released to TSMC for volume production has been upgraded so this target price may have changed.

3. The former CEO stated that they would have a pricing structure that was more favourable to larger volume purchases.

4. The very inexpensive pricing structure for the chip is based upon two ideas. The first is to ensure that there is absolutely no barrier to adoption by reason of price and to ensure that there is no possibility of a competitor with a lessor product competing. The second is to ensure that there is absolutely no profit to be had by attempting to bypass the patent protection surrounding AKD1000. If you can buy AKD1000 so cheaply why not buy the actual thing and then state that you are using it in your product. The costs of reverse engineering and then finding a foundry capable of producing AKD1000 should not make any commercial sense.

5. The Technology blogger Stephen Foskett was given an exclusive one on one briefing by Brainchip following which he stated the following: "Brainchip taking Ai everywhere". At $US15.00 this seems entirely feasible.

6. Vorago who have successfully completed their Phase 1 design contract for hardening AKD1000 in 28nm for NASA in their submission to NASA under Phase 1 stated in the other use cases for a hardened AKD1000 that their were billions of potential use cases beyond NASA.

7. Presently according to the NASA timetable the Phase 2 application by Vorago for the award of a contract to proceed to produce a hardened AKD1000 will be announced on 13 May, 2021. It has been noted by posters here that over the 12 months or more that NASA Phase 1 & 2 announcements have been made it is not uncommon for published dates to be extended.

8. If Vorago is the recipient of the Phase 2 award according to its web site it will forward the design to a foundry in much the same way as Socionext has forwarded the design to TSMC for the AKD1000 chip.

9. Despite the huge potential of the AKD1000 chip the actual primary business model of Brainchip is one based upon how ARM has achieved its success. This was first stated by Brainchip in 2015 and has been followed up to the present date. The ARM model is primarily based upon licencing of its Intellectual Property (IP). The former CEO of Brainchip described this model as highly lucrative with the margins to Brainchip being at least 97%. As has been announced Brainchip on 23.12.20 has signed an unconditional licence agreement with Renesas for the AKIDA technology IP.

10. Brainchip last year announced that Ford and Valeo had signed on as EAP customers.

11. Socionext announced jointly with Brainchip last year that it would be jointly marketing the AKD1000 chip with the Socionext SynQuacer and would also be assisting with the commercialisation of the full AKIDA family of products.

Hopefully this gives new players some starting points for doing their own research.

My opinion only DYOR
FF.

AKIDA Ballista