BRN Discussion Ongoing

[manny100]

I'm not sure if that's a good or a bad thing?..
As Qualcomm could "direct" the design and development direction..

Qualcomm to Bolster AI and IoT Capabilities with Edge Impulse Acquisition | Qualcomm

Enables more than 170,000 developers to create, deploy, and monitor AI models for a wide range of edge applications and hardware

www.qualcomm.com

"Qualcomm Technologies anticipates giving developers on Edge Impulse’s platform the ability to target Qualcomm Dragonwing™ processors, which feature superior on-device AI inferencing, computer vision, graphics and processing capabilities [Qualcomm Developer Blog]. Integration with Qualcomm® AI Hub enhances the optimization of AI models for Qualcomm Technologies’ platforms, leading to a peak of up to 4x increased inference performance, and reduced model size and memory footprint. Edge Impulse currently supports Dragonwing QCS6490 and QCS5430 processors and is scheduled to add support for additional Dragonwing processors for industrial and embedded IoT applications. Edge Impulse will maintain its current website (branded as “Edge Impulse, a Qualcomm company”) and remains dedicated to supporting developers and ecosystem partners"

It does say Edge Impulse "remains dedicated to supporting developers and ecosystem partners" but what influence, will they bear on encouragement or accessibility, of one over another?..

I presume Edge Impulse partnered with us because we make the Edge Impulse product better. In turn Dragonwing gets the neuromorphic AI it can only get from a 3rd party.
We all know the benefits of using AKIDA 1000.

 

[Bravo]

For what it's worth, although I’m sure it’s not much, I always felt Qualcomm had more to gain by siding with us. That said, I honestly wouldn’t know a semiconductor from a piece of toast.

My thoughts tend to focus on "efficiency", and I’m not convinced Qualcomm can match us when it comes to ultra-low power performance???

Qualcomm have been circling around Prophesee, but Prophesee seemingly found its perfect match with us. So, let’s just wait and see how this plays out.

From my completely non-technical perspective, I’ve always thought Qualcomm’s real challenge lay in areas like gesture recognition where an event-based, ultra-low power approach (i.e. Prophesee+BrainChip) might be the answer. They have always touted their "low power" capabilities, but I’ve had the feeling that we might present an opportunity to help them improve.

Time will tell.

• #91,959
• #91,746

 

[Bravo]

And then there is also this...

How will Qualcomm explain the following after the acquisition

• #74,292
• #79,733

 

[manny100]

Pitt Street reports are paid marketing.

They hold zero weight.

They actually back up their valuation:
Industry deals as a guide
In our report on BrainChip from August 2021, we valued the company using the Habana acquisition by Intel in 2019. At the time, Intel paid US$2BN for Habana, a pre-revenue Israel-based AI startup. This translated into A$1.50 per fully diluted BrainChip share. At the time, Habana was slightly ahead of BrainChip in terms of product launches with its Goya and Gaudi chip launches in 2018 and 2019. Today, BrainChip also has products on the market and has received license fees from multiple customers. In other words, we believe that US$2BN valuation for Habana is still a pretty good guide for BRN potential valuation at the moment.
Value of A$1.59 per share
Using the Habana valuation to value BrainChip, we arrive at a value of A$1.59 per fully diluted share (AUD/USD = 0.66). While this is substantially higher than today’s share price, BrainChip shares have been as high as A$2.34 in January 2022. Back then, the share price was driven by news around a major car manufacturer using Akida in one of its concept cars. We believe this illustrates how well the share price can potentially respond to good news.
We are now way, way ahead of Habana at the time it was taken over.
Also reports are their business. If they wrote dodgy reports, they would be out of business very quickly.
See page 19.

https://static1.squarespace.com/static/5af533a312b13fb602fe7d7b/t/66795c33b4685550a6583537/1719229495698/BrainChip+intiation+report+2024+25+6+2024.pdf

 

[Diogenese]

Well then, I suppose that makes Qualcomm our enablement partner.

So is this why BRN has ben hedging its bets with other "Enablement Partners"?

https://brainchip.com/partners/

 

[manny100]

Correct me if i am wrong but i believe Edge Impulse itself is not a neuromorphic system. It is a machine learning development platform. AKIDA 1000 gives Edge Impulse access to Neuromorphic capabilities.
The Qualcomm core is a traditional system. If it wants Neuromorphic capabilities it has to import it from a 3rd party.
Qualcomm may or may not want Edge Impulse for what it has developed using AKIDA 1000?
Its wait and see.

 

[JoMo68]

Edge Impulse to join Qualcomm Technologies! Edge Impulse is pleased to… | Edge Impulse | 20 comments

Edge Impulse to join Qualcomm Technologies! Edge Impulse is pleased to announce that we have entered into an agreement to be acquired by Qualcomm… | 20 comments on LinkedIn

www.linkedin.com

Wow !

View attachment 78940

I can’t help but feel this isn’t great for us… I hope we’re not squeezed out and replaced by Qualcomm tech. I hope I’m way off the mark though

 

[manny100]

I can’t help but feel this isn’t great for us… I hope we’re not squeezed out and replaced by Qualcomm tech. I hope I’m way off the mark though

Qualcomm tech is a traditional core system. If it wants Neuromorphic AI it has to import it from a 3rd party.

 

[Diogenese]

Correct me if i am wrong but i believe Edge Impulse itself is not a neuromorphic system. It is a machine learning development platform. AKIDA 1000 gives Edge Impulse access to Neuromorphic capabilities.
The Qualcomm core is a traditional system. If it wants Neuromorphic capabilities it has to import it from a 3rd party.
Qualcomm may or may not want Edge Impulse for what it has developed using AKIDA 1000?
Its wait and see.

Qualcomm have their in-house Hexagon NPU which grew out of a DSP:

https://www.qualcomm.com/products/technology/processors/hexagon
The star of the Qualcomm® AI Engine
The Hexagon NPU, designed from the ground up for accelerating AI inference at low power, features the industry’s most advanced NPU architecture—evolving along with the development of new AI use cases, models, and requirements.


... but it's pretty feeble as they share the AI load between the NPU, CPU, and GPU with increasing load.

What is the benefit of heterogeneous computing for AI?​

Heterogeneous computing for AI allows new AI instructions or hardware accelerators to be inserted into each processor. This processor diversity allows the architecture to continuously evolve for heterogenous computing features, in turn maximizing performance and power efficiency. This is beneficial as most other chipset vendors typically license several third-party processors, which cannot guarantee if they will fit or work under the same constraints.

The Qualcomm AI Engine, featured in our Snapdragon platforms and many of our other products, is at the core of our on-device AI and heterogenous computing advantage. With the CPU, GPU, NPU and Qualcomm Sensing Hub all working together and the result of many years of full-stack AI optimization, the Qualcomm AI Engine provides best-in-class on-device AI performance at extremely low power to support use cases today and in the future.

https://www.qualcomm.com/content/dam/qcomm-martech/dm-assets/images/uploads/npus.png

 

[DingoBorat]

I presume Edge Impulse partnered with us because we make the Edge Impulse product better. In turn Dragonwing gets the neuromorphic AI it can only get from a 3rd party.
We all know the benefits of using AKIDA 1000.

Dragonwing is already a developed and implemented suite of products.

Dragonwing | Qualcomm

Powered by leading-edge on-device AI, computing, and connectivity, Dragonwing products and solutions unlock potential, so today’s industries and businesses can scale to new heights.

www.qualcomm.com

Not saying they "can't" add some neuromorphic sauce later, but they don't "need" it now.

Plus Qualcomm is an American Company and many here, seem to want to exclude BrainChip from the US markets, until they have proven themselves in other smaller markets.

 

[Diogenese]

For what it's worth, although I’m sure it’s not much, I always felt Qualcomm had more to gain by siding with us. That said, I honestly wouldn’t know a semiconductor from a piece of toast.

My thoughts tend to focus on "efficiency", and I’m not convinced Qualcomm can match us when it comes to ultra-low power performance???

Qualcomm have been circling around Prophesee, but Prophesee seemingly found its perfect match with us. So, let’s just wait and see how this plays out.

From my completely non-technical perspective, I’ve always thought Qualcomm’s real challenge lay in areas like gesture recognition where an event-based, ultra-low power approach (i.e. Prophesee+BrainChip) might be the answer. They have always touted their "low power" capabilities, but I’ve had the feeling that we might present an opportunity to help them improve.

Time will tell.

• #91,959
• #91,746

Akida models are unique. Many of them have been built with Edge Impulse.

Let's hope the BRN/Edge Impulse NDA is watertight.

PS: Semiconductors are more like chips than toast.

 

[Diogenese]

Akida models are unique. Many of them have been built with Edge Impulse.

Let's hope the BRN/Edge Impulse NDA is watertight.

PS: Semiconductors are more like chips than toast.

I'd missed the outcome of the ARM/qualcomm/Nuvia licence breach litigation:

https://www.xda-developers.com/qualcomm-vs-arm-lawsuit-finished/

Spoiler: Qualcomm won.

 

[manny100]

Qualcomm have their in-house Hexagon NPU which grew out of a DSP:

https://www.qualcomm.com/products/technology/processors/hexagon
The star of the Qualcomm® AI Engine
The Hexagon NPU, designed from the ground up for accelerating AI inference at low power, features the industry’s most advanced NPU architecture—evolving along with the development of new AI use cases, models, and requirements.


... but it's pretty feeble as they share the AI load between the NPU, CPU, and GPU with increasing load.

What is the benefit of heterogeneous computing for AI?​

Heterogeneous computing for AI allows new AI instructions or hardware accelerators to be inserted into each processor. This processor diversity allows the architecture to continuously evolve for heterogenous computing features, in turn maximizing performance and power efficiency. This is beneficial as most other chipset vendors typically license several third-party processors, which cannot guarantee if they will fit or work under the same constraints.

The Qualcomm AI Engine, featured in our Snapdragon platforms and many of our other products, is at the core of our on-device AI and heterogenous computing advantage. With the CPU, GPU, NPU and Qualcomm Sensing Hub all working together and the result of many years of full-stack AI optimization, the Qualcomm AI Engine provides best-in-class on-device AI performance at extremely low power to support use cases today and in the future.

https://www.qualcomm.com/content/dam/qcomm-martech/dm-assets/images/uploads/npus.png

View attachment 78946

Thanks, just checked Qualcomm's description of dragonwing.
It says "powered by leading-edge on- device, AI, computing and connectivity"
It does mention Neuromorphic AI at all but 'leading-edge on-device confuses things a little.

Dragonwing | Qualcomm

Powered by leading-edge on-device AI, computing, and connectivity, Dragonwing products and solutions unlock potential, so today’s industries and businesses can scale to new heights.

www.qualcomm.com

 

[Bravo]

Just checking out some of the "likes" again on Tony Lewis' recent LinkedIn post.

M Anthony Lewis on LinkedIn: #ai #edgeai #machinelearning #braininspiredcomputing #brainchip

Embedded World in Nuremberg, Germany awaits! BrainChip's moment to shine! Wednesday of next week, I will be presenting at Embedded World in Nuremberg…

www.linkedin.com

 

[manny100]

Thanks, just checked Qualcomm's description of dragonwing.
It says "powered by leading-edge on- device, AI, computing and connectivity"
It does mention Neuromorphic AI at all but 'leading-edge on-device confuses things a little.

Dragonwing | Qualcomm

Powered by leading-edge on-device AI, computing, and connectivity, Dragonwing products and solutions unlock potential, so today’s industries and businesses can scale to new heights.

www.qualcomm.com

Qualcomm Ann.

Qualcomm to Bolster AI and IoT Capabilities with Edge Impulse Acquisition | Qualcomm

Enables more than 170,000 developers to create, deploy, and monitor AI models for a wide range of edge applications and hardware

www.qualcomm.com

 

[Beebo]

Maybe, just maybe, redomiciling to the US is also intended to facilitate potential M&A. Companies tend to ‘clean house’ years in advance of strategic maneuvers.

Would ARM be interested in BRN?
Would AVGO be interested in BRN?
Would QCOM be interested in BRN?

 

[manny100]

I forgot that Edge Impulse has integrated AKIDA into its platform and according to the BRN news release in Jan'25 is a key partner in our Edge Box.
Edge Impulse in our Edge Box in Jan'25 and taken over by Qualcomm in March'25.. . hmmmm.