Mt09 , for the second time , would you like to present your best presentation as to why Akida ip is not in the Arm Ethos U85 ?
Looking at your recent contributions , I will understand if you decide to decline again .
Arm Ethos U85 and BrainChip Akida 2.0 comparisons
- Thread starter Doz
- Start date
Mt09 , for the second time , would you like to present your best presentation as to why Akida ip is not in the Arm Ethos U85 ?
Looking at your recent contributions , I will understand if you decide to decline again .
Namaz
Emerged
It depends on what is more important in your project: performance and flexibility (Ethos U85) or energy efficiency and specialized real-time computing (Akida 2.0).
Arm Ethos U85 is ideal for use in applications that require high performance and versatility.
BrainChip Akida 2.0 is the perfect choice for applications where energy efficiency and real-time performance are critical, such as IoT, medical devices, and power-constrained systems.
Arm Ethos U85 is ideal for use in applications that require high performance and versatility.
BrainChip Akida 2.0 is the perfect choice for applications where energy efficiency and real-time performance are critical, such as IoT, medical devices, and power-constrained systems.
Hi Namaz ,
You will find on pages 1 ,2 and 3 , such things as Arm‘s Paul Williamson’s claim that the Ethos U85 will operate in milliwatts and Rene claims that it is the lowest power transformer on the market , fast inference , SOTA performance , IoT , medical , edge , endpoint and much , much more .
Are you in a position to tell me one difference between Arm’s Ethos U85 and BrainChip’s Akida ?
Last edited:
Interesting change in the wording , from Graph Neural Networks AND Spiking Neural Networks , to
Graph Neural Networks OR Spiking Neural Networks .
This to me indicates that Akida with TeNNs / Pleiades / aTENNuate can perform what current tasks GNN’s can perform but with the additional benefits as previously highlighted .
Arpan Pal moving from GNN to SNN for EEG processing may support this theory .
Graph Neural Networks OR Spiking Neural Networks .
This to me indicates that Akida with TeNNs / Pleiades / aTENNuate can perform what current tasks GNN’s can perform but with the additional benefits as previously highlighted .
Arpan Pal moving from GNN to SNN for EEG processing may support this theory .
Last edited:
Hi wasMAD ,
First , I don’t follow NVU but I see that they plan to tape out a sample chip in quarter 4 2025 , with expectations of full production tape out Q1 2026 .
This fondly takes me back to 2020 . As such if I was interested in NVU , I would study Brainchip’s road in between 2020 and Q1 2026 when we are expecting an ASIC tape out of the AKD 2.0 by a customer . You may discover that the industry is very protective and only those who respect the ecosystem are welcome . You may discover that just providing a chip with 30 GOP’s is not what most customers want , or the fast change of technology requires NVU to reassess . NVU may need to raise lots of capital in this 5 or 6 year period , employ additional sales teams , PhD’s and replace those who leave . Is providing a chip enough , will potential customers want a full stack to go with the chip . Then there is that old saying , no one ever got the sack for buying IBM .
MAD , as any Brainchip shareholder will tell you , the 6 year road is tough , even tougher if you don’t realise when the share price is over valued . In my opinion , if NVU are thinking that they are going to disrupt the market in 2026 , boy have I got some news for you …..
MAD , remember I don’t follow NVU and I am not an advisor , so please DYOR and seek professional advice as required .
Doz .
Aizip Works with SoftBank Corp. to Launch Customized Small Language Model Solutions for Privacy-Critical Enterprise Applications - Aizip
Cupertino, CA, December 10, 2024 – Aizip, Inc. in partnership with SoftBank Corp., announced the release of customized Small Language Model (SLM) and Retrieval Augmented Generation (RAG) solutions for enterprise applications. The system operates locally on mobile devices or on-premises servers...
aizip.ai
Attachments
Yubei Chen, Co-Founder of Aizip Inc – Interview Series
Yubei Chen is co-founder of Aizip inc., a company that builds the world’s smallest and most efficient AI models. He is also an assistant professor in the ECE Department at University of California, Davis. Chen’s research is at the intersection of computational neuroscience and deep unsupervised...
Did someone mention something about engineering samples ?
Ensemble E8 DevKit (DK-E8) | Alif Semiconductor
Get started with the Alif Semiconductor Ensemble DevKit, which brings out all signals on the device to easily accessible pins for power and performance profiling, prototyping and more.
alifsemi.com
Looking back at the Brainchip / AiZip / Plumerai / Arm demo at Embedded world 2023 as exposed by Sally Ward Foxton .
Note how this demo has been given the Renesas code RA8D1 .
The existing Renesas RA8D1 Block Plan does not currently have an NPU .
Note : The current Renesas RA8P1 Block Plan does include a NPU - The old Arm Ethos U55 .
This statement made by Arm’s Paul Williamson on the 26/02/2025 , claims that the Arm Ethos U85 was in the Renesas RA8E microcontroller back in October 2023 . ( Note : Well before Arm launched the Ethos U85 on the 09/04/2024 .
Note : Once again the Block Plan does not include a NPU for the Renesas RA8E1 .
The below comment was made in regards to the Renesas RA8P1 , whilst the RA8E ( claimed Ethos U85 ) and RA8D1 ( Embedded World demo Akida ) devices remain without NPU .
www.rdworldonline.com
Interesting that Arm can produce a new Chip designed for targeting 50% of the mobile phone industry in quick time , but a little NPU is still in development . I wonder what Arm’s Roadmap looks like ?
.
Note how this demo has been given the Renesas code RA8D1 .
The existing Renesas RA8D1 Block Plan does not currently have an NPU .
Note : The current Renesas RA8P1 Block Plan does include a NPU - The old Arm Ethos U55 .
This statement made by Arm’s Paul Williamson on the 26/02/2025 , claims that the Arm Ethos U85 was in the Renesas RA8E microcontroller back in October 2023 . ( Note : Well before Arm launched the Ethos U85 on the 09/04/2024 .
Note : Once again the Block Plan does not include a NPU for the Renesas RA8E1 .
The below comment was made in regards to the Renesas RA8P1 , whilst the RA8E ( claimed Ethos U85 ) and RA8D1 ( Embedded World demo Akida ) devices remain without NPU .
First MCU combines gigahertz CPU, 35× faster AI engine and MRAM in single device - Research & Development World
Japanese semiconductor firm Renesas Electronics has unveiled the RA8P1, positioning it as a breakthrough microcontroller that integrates gigahertz-class processing, dedicated AI acceleration hardware and magnetoresistive RAM (MRAM) on a single 22-nanometer chip. The chip includes graphics...
www.rdworldonline.com
Interesting that Arm can produce a new Chip designed for targeting 50% of the mobile phone industry in quick time , but a little NPU is still in development . I wonder what Arm’s Roadmap looks like ?
.
Can’t wait to see if the new Arm mobile chip designs include the super resolution feature as previously discussed on this thread .
What would work out cheaper for Renesas . Either utilising their subscription and existing Ethos license with Arm and then saving additional money by combining a Cortex or multiple Cortex’s with the Ethos U85 , or continue to take out single use licenses with Brainchip ?
I think the combo deal myself .
I think the combo deal myself .
Similar threads
