BRN Discussion Ongoing

Frangipani

Regular
AKD 1000 AKD 1500 AKD 2000 and moving ever so closer...AKD 3000...and 3 of which are 100% available now and setting the benchmark
which the company believe their IP will end up in everything, everywhere...(that I'd suggest is the target, NOT a proven fact to date).

Hi TECH,

just an attempt to sort out the terminology…

The following is how I understand BrainChip’s nomenclature, but I could be wrong, so everybody please feel free to brainstorm and chip in…

In early 2022, shortly after Sean Hehir had joined BrainChip as CEO, “AKD1000” was still used as an umbrella term to describe everything our company had for sale at the time (chip, IP, PCIe board), as evident by the following investor conference presentation slides:

7CF83133-83B4-4BD3-97DE-EA0CBC24FA44.jpeg



These days, however, a distinction appears to be made between the generational iterations of the Akida processor technology platform (categorised as akida, akida 2.0…) as opposed to the physical reference chips (so far AKD1000 and AKD1500).

The AKD1000 and AKD1500 SoCs are both silicon implementations of the Akida technology embodied in BrainChip’s 1st Generation Edge AI neuromorphic processor platform akida (technically speaking akida 1.0).

AKD1000 was implemented with TSMC at 28nm, whereas AKD1500 was taped out on GlobalFoundries’ MCU-friendly 22nm fully depleted silicon-on-insulator (FD-SOI) process, aka GF’s 22FDX technology. As reference chips, they were primarily meant to target prospective IP licencees (as a proof of concept), but at the same time the AKD1000 (on a PCIe board or inside a Dev Kit) benefitted individual developers (professional hardware engineers in companies or academic settings as well as advanced hobbyists) who were not interested in mass production of edge devices and the signing of an IP licence, but instead may have only required a single PCIe Board or Dev Kit for their projects or research (take note that it says on the BrainChip shop website in bold capital letters that development kits are not intended to be used for production purposes). Meanwhile, AKD1000 chips have also been integrated into the VVDN Edge AI Boxes, and the Unigen Cupcake Edge AI Server will soon be offered with a new configuration based on the AKD1500 (?) as an AI option.

Akida 2.0, the neural processing system’s enhanced 2nd Generation, was announced and introduced last year, but purely as an IP offering, productised in three different variations: akida-E, akida-S and akida-P, depending on where in the Edge AI spectrum (sensor edge < server edge) its prowess is required.

AKD2000, however, doesn’t exist - at least not yet.
The way I understand it, AKD2000 would be the name
of BrainChip’s (hypothetical) reference chip based on Akida Gen 2, which may or may not materialise.

Let’s recall what was said earlier this year:
In an interview with Jim McGregor from TIRIAS Research during CES 2024 (January 9-12), Todd Vierra replied the following to his interview partner’s comment “And correct me if I’m wrong, but this is the Akida 2?”
Todd: “This is actually all ran [sic] on Akida 1 hardware. Akida 2, erm - we are in the process of taping out and we’ll get that silicon back a little bit later, but these are all just Gen 1...” (from 9:26 min). His statement about an imminent tapeout seemed to confirm what (according to FF) Sean Hehir had told select shareholders in the November 2023 Sydney “secret meeting”.

Surprisingly, a mere seven weeks later, during the Virtual Investor Roadshow (February 27), neither Sean Hehir nor Tony Lewis mentioned anything at all about a tape-out in progress, but instead argued a second generation reference chip as proof of concept was unnecessary, while at the same time not totally excluding a potential future tape-out complementing their core IP business. However, our CTO made it clear that it is definitely not their intention to manufacture chips on a large scale:


From 43:17
Roger Manning’s question: “Given BrainChip’s business model is to largely sell its akida IP, will it be necessary to prove each new version of akida in silicon?”

Tony: “So I think the big question was, will this event-based paradigm yield, erm can it be done and will it yield some benefits to customers, and I think we achieved that by taping out our earlier generation of products, and so we’ve already achieved that. And it’s my belief that there is only marginal benefit in taping out the next generation, we already proved the main points of it. And clearly we don’t want to start to manufacture chips on a large scale. We’d be competing with our customers and that would really break our business model right now.

Sean: “Yeah, and to be clear, a lot of work has gone on with our ability to simulate workloads in Generation 2 as well. As Tony said, we certainly have reference chips in Generation 1, and the typical engagement course that we work with IP license prospects is we allow them to run models on there and/or simulate them in our simulation tools that we have for Generation 2, so Roger, I would say stay tuned, we may or may not, erm, but right now, there is no need for us to do that.”




Now the way I understood the “We’d be competing with our customers” comment is not for fear of treading on their customers’ or a specific customer’s toes (as other TSE users have interpreted it), but because they would shoot themselves in the foot by doing that: after all, it would be less profitable for our company when customers could just buy those chips off the shelves to utilise them in their in-house development (and hence save a lot of money, as there won’t be any follow-up costs for them) rather than having to pay an initial IP license fee and future royalties.

Does my interpretation make sense or am I overlooking something here?



And as for the next iteration of the Akida processor family, we don’t even know, yet, whether it will be called akida 3.0 or akida 2.X…

From min 47:41 min:

Sean Hehir: “… but round Generation 3, if you will: You know, I mentioned earlier in my slide that when I talked about our product planning and our execution cycle, if you will, you could see we are always planning on this product and we are always looking for improvements on our IP offering right there. Now whether we call it formally “Generation 3” or “2 something x” - but yes, we are in the middle of a planning cycle right now to make some changes and we’ll make announcements over time.”


So what we can say with confidence (provided we believe our CEO’s words) is that the advent of the Akida neural processor family’s 3rd generation is approaching… Whether there will ever be an AKD3000, though, is uncertain.

Regards
Frangipani


P.S.: I still can’t get my head around Todd Vierra’s statement, though. Would they really have backed out at the last minute, eg due to financial constraints?
Or was he possibly referring to a tapeout not of AKD2000 as a reference chip, but to a tape-out of a SoC by a specific customer that includes Akida 2.0 IP, along the lines of what DingoBorat said?

To me, it's saying there's an in the flesh, integrated circuit, customer custom SOC, with our IP in it...

But then again, would he really say “…we’ll get that silicon back a little later”?
I suppose only if it was a joint development, such as possibly one with Socionext or Tata Elxsi (which could explain the different colour of the mysterious Custom SoC on that presentation slide)? As otherwise, wouldn’t such a SoC be taped out by the customer itself rather than by / in collaboration with BrainChip? I am a bit confused here. 🤔

It seems too early as a mere placeholder for the planned integration of Akida IP into the Frontgrade Gaisler SoC, which according to ESA’s Laurent Hili “We aim to tape out ideally before the end of the year, beginning of next year” (from 47:25 min shortly before the end of the mid -March BrainChip podcast Episode 31). After all, they could have labelled it “prospective Customer Custom SoC”, but the way it is presented on the slide, I agree with DingoBorat that it does appear to be an already existing physical implementation, indeed. Mmmmh… 🤔
 
Last edited:
  • Like
  • Love
  • Fire
Reactions: 32 users
View attachment 61438

I wonder why it’s in Australia ....
Nothing to see here as they have similar events which are normally just gaming

 
  • Like
Reactions: 3 users

Guzzi62

Regular
Hi TECH,

just an attempt to sort out the terminology…

The following is how I understand BrainChip’s nomenclature, but I could be wrong, so everybody please feel free to brainstorm and chip in…

In early 2022, shortly after Sean Hehir had joined BrainChip as CEO, “AKD1000” was still used as an umbrella term to describe everything our company had for sale at the time (chip, IP, PCIe board), as evident by the following investor conference presentation slides:

View attachment 61365


These days, however, a distinction appears to be made between the generational iterations of the Akida processor technology platform (categorised as akida, akida 2.0…) as opposed to the physical reference chips (so far AKD1000 and AKD1500).

The AKD1000 and AKD1500 SoCs are both silicon implementations of the Akida technology embodied in BrainChip’s 1st Generation Edge AI neuromorphic processor platform akida (technically speaking akida 1.0).

AKD1000 was implemented with TSMC at 28nm, whereas AKD1500 was taped out on GlobalFoundries’ MCU-friendly 22nm fully depleted silicon-on-insulator (FD-SOI) process, aka GF’s 22FDX technology. As reference chips, they were primarily meant to target prospective IP licencees (as a proof of concept), but at the same time the AKD1000 (on a PCIe board or inside a Dev Kit) benefitted individual developers (professional hardware engineers in companies or academic settings as well as advanced hobbyists) who were not interested in mass production of edge devices and the signing of an IP licence, but instead may have only required a single PCIe Board or Dev Kit for their projects or research (take note that it says on the BrainChip shop website in bold capital letters that development kits are not intended to be used for production purposes). Meanwhile, AKD1000 chips have also been integrated into the VVDN Edge AI Boxes, and the Unigen Cupcake Edge AI Server will soon be offered with a new configuration based on the AKD1500 (?) as an AI option.

Akida 2.0, the neural processing system’s enhanced 2nd Generation, was announced and introduced last year, but purely as an IP offering, productised in three different variations: akida-E, akida-S and akida-P, depending on where in the Edge AI spectrum (sensor edge < server edge) its prowess is required.

AKD2000, however, doesn’t exist - at least not yet.
The way I understand it, AKD2000 would be the name
of BrainChip’s (hypothetical) reference chip based on Akida Gen 2, which may or may not materialise.

Let’s recall what was said earlier this year:
In an interview with Jim McGregor from TIRIAS Research during CES 2024 (January 9-12), Todd Vierra replied the following to his interview partner’s comment “And correct me if I’m wrong, but this is the Akida 2?”
Todd: “This is actually all ran [sic] on Akida 1 hardware. Akida 2, erm - we are in the process of taping out and we’ll get that silicon back a little bit later, but these are all just Gen 1...” (from 9:26 min). His statement about an imminent tapeout seemed to confirm what (according to FF) Sean Hehir had told select shareholders in the November 2023 Sydney “secret meeting”.

Surprisingly, a mere seven weeks later, during the Virtual Investor Roadshow (February 27), neither Sean Hehir nor Tony Lewis mentioned anything at all about a tape-out in progress, but instead argued a second generation reference chip as proof of concept was unnecessary, while at the same time not totally excluding a potential future tape-out complementing their core IP business. However, our CTO made it clear that it is definitely not their intention to manufacture chips on a large scale:


From 43:17
Roger Manning’s question: “Given BrainChip’s business model is to largely sell its akida IP, will it be necessary to prove each new version of akida in silicon?”

Tony: “So I think the big question was, will this event-based paradigm yield, erm can it be done and will it yield some benefits to customers, and I think we achieved that by taping out our earlier generation of products, and so we’ve already achieved that. And it’s my belief that there is only marginal benefit in taping out the next generation, we already proved the main points of it. And clearly we don’t want to start to manufacture chips on a large scale. We’d be competing with our customers and that would really break our business model right now.

Sean: “Yeah, and to be clear, a lot of work has gone on with our ability to simulate workloads in Generation 2 as well. As Tony said, we certainly have reference chips in Generation 1, and the typical engagement course that we work with IP license prospects is we allow them to run models on there and/or simulate them in our simulation tools that we have for Generation 2, so Roger, I would say stay tuned, we may or may not, erm, but right now, there is no need for us to do that.”




Now the way I understood the “We’d be competing with our customers” comment is not for fear of treading on their customers’ or a specific customer’s toes (as other TSE users have interpreted it), but because they would shoot themselves in the foot by doing that: after all, it would be less profitable for our company when customers could just buy those chips off the shelves to utilise them in their in-house development (and hence save a lot of money, as there won’t be any follow-up costs for them) rather than having to pay an initial IP license fee and future royalties.

Does my interpretation make sense or am I overlooking something here?



And as for the next iteration of the Akida processor family, we don’t even know, yet, whether it will be called akida 3.0 or akida 2.X…

From min 47:41 min:

Sean Hehir: “… but round Generation 3, if you will: You know, I mentioned earlier in my slide that when I talked about our product planning and our execution cycle, if you will, you could see we are always planning on this product and we are always looking for improvements on our IP offering right there. Now whether we call it formally “Generation 3” or “2 something x” - but yes, we are in the middle of a planning cycle right now to make some changes and we’ll make announcements over time.”


So what we can say with confidence (provided we believe our CEO’s words) is that the advent of the Akida neural processor family’s 3rd generation is approaching… Whether there will ever be an AKD3000, though, is uncertain.

Regards
Frangipani


P.S.: I still can’t get my head around Todd Vierra’s statement, though. Would they really have backed out at the last minute, eg due to financial constraints?
Or was he possibly referring to a tapeout not of AKD2000 as a reference chip, but to a tape-out of a SoC by a specific customer that includes Akida 2.0 IP, along the lines of what DingoBorat said?



But then again, would he really say “…we’ll get that silicon back a little later”?
I suppose only if it was a joint development, such as possibly one with Socionext or Tata Elxsi (which could explain the different colour of the mysterious Custom SoC on that presentation slide)? As otherwise, wouldn’t such a SoC be taped out by the customer itself rather than by / in collaboration with BrainChip? I am a bit confused here. 🤔

It seems too early as a mere placeholder for the planned integration of Akida IP into the Frontgrade Gaisler SoC, which according to ESA’s Laurent Hili “We aim to tape out ideally before the end of the year, beginning of next year” (from 47:25 min shortly before the end of the mid -March BrainChip podcast Episode 31). After all, they could have labelled it “prospective Customer Custom SoC”, but the way it is presented on the slide, I agree with DingoBorat that it does appear to be an already existing physical implementation, indeed. Mmmmh… 🤔
From their newsletter (small part of it) sent to my email:

April has been buzzing with innovation and discovery here at BrainChip. We recently spent a whirlwind week at Embedded World 2024, where the team demonstrated Akida in real-life use cases and engaged in extensive conversations around AI, Edge processing, and Akida neuromorphic technology.

And it was about gen2 if I am correct so they must have some physical Gen2 chips around to build the models.

You can't really simulate using somebody's else's hardware, can you? Well I guess you can but not very professional and won't impress anybody.
 
  • Like
Reactions: 2 users

TECH

Regular
Hi TECH,

just an attempt to sort out the terminology…

The following is how I understand BrainChip’s nomenclature, but I could be wrong, so everybody please feel free to brainstorm and chip in…

In early 2022, shortly after Sean Hehir had joined BrainChip as CEO, “AKD1000” was still used as an umbrella term to describe everything our company had for sale at the time (chip, IP, PCIe board), as evident by the following investor conference presentation slides:

View attachment 61365


These days, however, a distinction appears to be made between the generational iterations of the Akida processor technology platform (categorised as akida, akida 2.0…) as opposed to the physical reference chips (so far AKD1000 and AKD1500).

The AKD1000 and AKD1500 SoCs are both silicon implementations of the Akida technology embodied in BrainChip’s 1st Generation Edge AI neuromorphic processor platform akida (technically speaking akida 1.0).

AKD1000 was implemented with TSMC at 28nm, whereas AKD1500 was taped out on GlobalFoundries’ MCU-friendly 22nm fully depleted silicon-on-insulator (FD-SOI) process, aka GF’s 22FDX technology. As reference chips, they were primarily meant to target prospective IP licencees (as a proof of concept), but at the same time the AKD1000 (on a PCIe board or inside a Dev Kit) benefitted individual developers (professional hardware engineers in companies or academic settings as well as advanced hobbyists) who were not interested in mass production of edge devices and the signing of an IP licence, but instead may have only required a single PCIe Board or Dev Kit for their projects or research (take note that it says on the BrainChip shop website in bold capital letters that development kits are not intended to be used for production purposes). Meanwhile, AKD1000 chips have also been integrated into the VVDN Edge AI Boxes, and the Unigen Cupcake Edge AI Server will soon be offered with a new configuration based on the AKD1500 (?) as an AI option.

Akida 2.0, the neural processing system’s enhanced 2nd Generation, was announced and introduced last year, but purely as an IP offering, productised in three different variations: akida-E, akida-S and akida-P, depending on where in the Edge AI spectrum (sensor edge < server edge) its prowess is required.

AKD2000, however, doesn’t exist - at least not yet.
The way I understand it, AKD2000 would be the name
of BrainChip’s (hypothetical) reference chip based on Akida Gen 2, which may or may not materialise.

Let’s recall what was said earlier this year:
In an interview with Jim McGregor from TIRIAS Research during CES 2024 (January 9-12), Todd Vierra replied the following to his interview partner’s comment “And correct me if I’m wrong, but this is the Akida 2?”
Todd: “This is actually all ran [sic] on Akida 1 hardware. Akida 2, erm - we are in the process of taping out and we’ll get that silicon back a little bit later, but these are all just Gen 1...” (from 9:26 min). His statement about an imminent tapeout seemed to confirm what (according to FF) Sean Hehir had told select shareholders in the November 2023 Sydney “secret meeting”.

Surprisingly, a mere seven weeks later, during the Virtual Investor Roadshow (February 27), neither Sean Hehir nor Tony Lewis mentioned anything at all about a tape-out in progress, but instead argued a second generation reference chip as proof of concept was unnecessary, while at the same time not totally excluding a potential future tape-out complementing their core IP business. However, our CTO made it clear that it is definitely not their intention to manufacture chips on a large scale:


From 43:17
Roger Manning’s question: “Given BrainChip’s business model is to largely sell its akida IP, will it be necessary to prove each new version of akida in silicon?”

Tony: “So I think the big question was, will this event-based paradigm yield, erm can it be done and will it yield some benefits to customers, and I think we achieved that by taping out our earlier generation of products, and so we’ve already achieved that. And it’s my belief that there is only marginal benefit in taping out the next generation, we already proved the main points of it. And clearly we don’t want to start to manufacture chips on a large scale. We’d be competing with our customers and that would really break our business model right now.

Sean: “Yeah, and to be clear, a lot of work has gone on with our ability to simulate workloads in Generation 2 as well. As Tony said, we certainly have reference chips in Generation 1, and the typical engagement course that we work with IP license prospects is we allow them to run models on there and/or simulate them in our simulation tools that we have for Generation 2, so Roger, I would say stay tuned, we may or may not, erm, but right now, there is no need for us to do that.”




Now the way I understood the “We’d be competing with our customers” comment is not for fear of treading on their customers’ or a specific customer’s toes (as other TSE users have interpreted it), but because they would shoot themselves in the foot by doing that: after all, it would be less profitable for our company when customers could just buy those chips off the shelves to utilise them in their in-house development (and hence save a lot of money, as there won’t be any follow-up costs for them) rather than having to pay an initial IP license fee and future royalties.

Does my interpretation make sense or am I overlooking something here?



And as for the next iteration of the Akida processor family, we don’t even know, yet, whether it will be called akida 3.0 or akida 2.X…

From min 47:41 min:

Sean Hehir: “… but round Generation 3, if you will: You know, I mentioned earlier in my slide that when I talked about our product planning and our execution cycle, if you will, you could see we are always planning on this product and we are always looking for improvements on our IP offering right there. Now whether we call it formally “Generation 3” or “2 something x” - but yes, we are in the middle of a planning cycle right now to make some changes and we’ll make announcements over time.”


So what we can say with confidence (provided we believe our CEO’s words) is that the advent of the Akida neural processor family’s 3rd generation is approaching… Whether there will ever be an AKD3000, though, is uncertain.

Regards
Frangipani


P.S.: I still can’t get my head around Todd Vierra’s statement, though. Would they really have backed out at the last minute, eg due to financial constraints?
Or was he possibly referring to a tapeout not of AKD2000 as a reference chip, but to a tape-out of a SoC by a specific customer that includes Akida 2.0 IP, along the lines of what DingoBorat said?



But then again, would he really say “…we’ll get that silicon back a little later”?
I suppose only if it was a joint development, such as possibly one with Socionext or Tata Elxsi (which could explain the different colour of the mysterious Custom SoC on that presentation slide)? As otherwise, wouldn’t such a SoC be taped out by the customer itself rather than by / in collaboration with BrainChip? I am a bit confused here. 🤔

It seems too early as a mere placeholder for the planned integration of Akida IP into the Frontgrade Gaisler SoC, which according to ESA’s Laurent Hili “We aim to tape out ideally before the end of the year, beginning of next year” (from 47:25 min shortly before the end of the mid -March BrainChip podcast Episode 31). After all, they could have labelled it “prospective Customer Custom SoC”, but the way it is presented on the slide, I agree with DingoBorat that it does appear to be an already existing physical implementation, indeed. Mmmmh… 🤔

Hi Frangi,

I will try to answer you tomorrow to the best of my understanding....regards Chris (Tech)
 
  • Like
Reactions: 7 users

Frangipani

Regular
From their newsletter (small part of it) sent to my email:

April has been buzzing with innovation and discovery here at BrainChip. We recently spent a whirlwind week at Embedded World 2024, where the team demonstrated Akida in real-life use cases and engaged in extensive conversations around AI, Edge processing, and Akida neuromorphic technology.

And it was about gen2 if I am correct so they must have some physical Gen2 chips around to build the models.

You can't really simulate using somebody's else's hardware, can you? Well I guess you can but not very professional and won't impress anybody.

I assume Akida 2.0 was merely presented in simulation and possibly per brochure.

According to the BrainChip website (https://brainchip.com/embeddedworld/), the demo partnerships at embedded world 2024 were with NVISO and VVDN and thus based on Gen 1:

2730C4BC-83E1-4F27-8CFC-29798678A885.jpeg


The BrainChip presentation area within the tinyML Foundation pavilion wasn’t exactly gigantic - strangely, we can’t even spot one of the VVDN Edge AI Boxes in the pictures…

4C91FEC8-A7AA-487C-8EEB-D593CAC0F8D1.jpeg
 
  • Like
  • Fire
  • Sad
Reactions: 8 users

Diogenese

Top 20
Hi TECH,

just an attempt to sort out the terminology…

The following is how I understand BrainChip’s nomenclature, but I could be wrong, so everybody please feel free to brainstorm and chip in…

In early 2022, shortly after Sean Hehir had joined BrainChip as CEO, “AKD1000” was still used as an umbrella term to describe everything our company had for sale at the time (chip, IP, PCIe board), as evident by the following investor conference presentation slides:

View attachment 61365


These days, however, a distinction appears to be made between the generational iterations of the Akida processor technology platform (categorised as akida, akida 2.0…) as opposed to the physical reference chips (so far AKD1000 and AKD1500).

The AKD1000 and AKD1500 SoCs are both silicon implementations of the Akida technology embodied in BrainChip’s 1st Generation Edge AI neuromorphic processor platform akida (technically speaking akida 1.0).

AKD1000 was implemented with TSMC at 28nm, whereas AKD1500 was taped out on GlobalFoundries’ MCU-friendly 22nm fully depleted silicon-on-insulator (FD-SOI) process, aka GF’s 22FDX technology. As reference chips, they were primarily meant to target prospective IP licencees (as a proof of concept), but at the same time the AKD1000 (on a PCIe board or inside a Dev Kit) benefitted individual developers (professional hardware engineers in companies or academic settings as well as advanced hobbyists) who were not interested in mass production of edge devices and the signing of an IP licence, but instead may have only required a single PCIe Board or Dev Kit for their projects or research (take note that it says on the BrainChip shop website in bold capital letters that development kits are not intended to be used for production purposes). Meanwhile, AKD1000 chips have also been integrated into the VVDN Edge AI Boxes, and the Unigen Cupcake Edge AI Server will soon be offered with a new configuration based on the AKD1500 (?) as an AI option.

Akida 2.0, the neural processing system’s enhanced 2nd Generation, was announced and introduced last year, but purely as an IP offering, productised in three different variations: akida-E, akida-S and akida-P, depending on where in the Edge AI spectrum (sensor edge < server edge) its prowess is required.

AKD2000, however, doesn’t exist - at least not yet.
The way I understand it, AKD2000 would be the name
of BrainChip’s (hypothetical) reference chip based on Akida Gen 2, which may or may not materialise.

Let’s recall what was said earlier this year:
In an interview with Jim McGregor from TIRIAS Research during CES 2024 (January 9-12), Todd Vierra replied the following to his interview partner’s comment “And correct me if I’m wrong, but this is the Akida 2?”
Todd: “This is actually all ran [sic] on Akida 1 hardware. Akida 2, erm - we are in the process of taping out and we’ll get that silicon back a little bit later, but these are all just Gen 1...” (from 9:26 min). His statement about an imminent tapeout seemed to confirm what (according to FF) Sean Hehir had told select shareholders in the November 2023 Sydney “secret meeting”.

Surprisingly, a mere seven weeks later, during the Virtual Investor Roadshow (February 27), neither Sean Hehir nor Tony Lewis mentioned anything at all about a tape-out in progress, but instead argued a second generation reference chip as proof of concept was unnecessary, while at the same time not totally excluding a potential future tape-out complementing their core IP business. However, our CTO made it clear that it is definitely not their intention to manufacture chips on a large scale:


From 43:17
Roger Manning’s question: “Given BrainChip’s business model is to largely sell its akida IP, will it be necessary to prove each new version of akida in silicon?”

Tony: “So I think the big question was, will this event-based paradigm yield, erm can it be done and will it yield some benefits to customers, and I think we achieved that by taping out our earlier generation of products, and so we’ve already achieved that. And it’s my belief that there is only marginal benefit in taping out the next generation, we already proved the main points of it. And clearly we don’t want to start to manufacture chips on a large scale. We’d be competing with our customers and that would really break our business model right now.

Sean: “Yeah, and to be clear, a lot of work has gone on with our ability to simulate workloads in Generation 2 as well. As Tony said, we certainly have reference chips in Generation 1, and the typical engagement course that we work with IP license prospects is we allow them to run models on there and/or simulate them in our simulation tools that we have for Generation 2, so Roger, I would say stay tuned, we may or may not, erm, but right now, there is no need for us to do that.”




Now the way I understood the “We’d be competing with our customers” comment is not for fear of treading on their customers’ or a specific customer’s toes (as other TSE users have interpreted it), but because they would shoot themselves in the foot by doing that: after all, it would be less profitable for our company when customers could just buy those chips off the shelves to utilise them in their in-house development (and hence save a lot of money, as there won’t be any follow-up costs for them) rather than having to pay an initial IP license fee and future royalties.

Does my interpretation make sense or am I overlooking something here?



And as for the next iteration of the Akida processor family, we don’t even know, yet, whether it will be called akida 3.0 or akida 2.X…

From min 47:41 min:

Sean Hehir: “… but round Generation 3, if you will: You know, I mentioned earlier in my slide that when I talked about our product planning and our execution cycle, if you will, you could see we are always planning on this product and we are always looking for improvements on our IP offering right there. Now whether we call it formally “Generation 3” or “2 something x” - but yes, we are in the middle of a planning cycle right now to make some changes and we’ll make announcements over time.”


So what we can say with confidence (provided we believe our CEO’s words) is that the advent of the Akida neural processor family’s 3rd generation is approaching… Whether there will ever be an AKD3000, though, is uncertain.

Regards
Frangipani


P.S.: I still can’t get my head around Todd Vierra’s statement, though. Would they really have backed out at the last minute, eg due to financial constraints?
Or was he possibly referring to a tapeout not of AKD2000 as a reference chip, but to a tape-out of a SoC by a specific customer that includes Akida 2.0 IP, along the lines of what DingoBorat said?



But then again, would he really say “…we’ll get that silicon back a little later”?
I suppose only if it was a joint development, such as possibly one with Socionext or Tata Elxsi (which could explain the different colour of the mysterious Custom SoC on that presentation slide)? As otherwise, wouldn’t such a SoC be taped out by the customer itself rather than by / in collaboration with BrainChip? I am a bit confused here. 🤔

It seems too early as a mere placeholder for the planned integration of Akida IP into the Frontgrade Gaisler SoC, which according to ESA’s Laurent Hili “We aim to tape out ideally before the end of the year, beginning of next year” (from 47:25 min shortly before the end of the mid -March BrainChip podcast Episode 31). After all, they could have labelled it “prospective Customer Custom SoC”, but the way it is presented on the slide, I agree with DingoBorat that it does appear to be an already existing physical implementation, indeed. Mmmmh… 🤔
A tapeout is an essential part of the IP, whether or not we were to progress to SoC.
 
  • Like
  • Thinking
  • Fire
Reactions: 16 users

Frangipani

Regular
View attachment 61438

I wonder why its in Australia ....

I wouldn’t read anything into it.

Looks like they moved it to Australia, after having had to cancel Nintendo Live 2024 Tokyo (planned for January 2024) at short notice for safety fears…



April 3: According to the Kyoto Shimbun, police have made an arrest followings [sic] threats tied to the cancelled Nintendo Live Tokyo 2024 event.

Nintendo Live 2024 Tokyo, which would have taken place this past January, was cancelled at the start of December. The suspect made “persistent threats” against employees as well as participants ahead of the Splatoon Koshien 2023 finals. Nintendo Live Tokyo 2024 was cancelled as a precautionary measure.

The man who has been arrested was not named. However, it’s been reported that he’s in his twenties and is a local civil servant from the Ibaraki Prefecture.

The suspect was arrested after sending death threats through Nintendo’s enquiry form on its website. Because of repeated threats sent through that form, he’s been accused of obstructing the company’s business. As of now it’s unclear what his motives were

April 24: Kyoto Shimbun has followed up with another report and the 27 year-old man has been formally charged. He sent messages to Nintendo between August 22 and November 29 of last year with 39 threats. The man said things such as “I’m going to make you regret releasing such a crappy game to the world” and threatened to “kill everyone involved.” Messages also warned Nintendo about events with fans.”
 
  • Sad
  • Like
  • Thinking
Reactions: 5 users

chapman89

Founding Member
Can somebody help me try understand something please.
For those that haven’t seen the post by Markus Schafer from Mercedes which was posted an hour or so ago, here it is below, followed by a screenshot of my comment and the Advanced UX Director of Mercedes to me.


Hello Beijing and Auto China 2024! Where better to present our Concept CLA Class to Chinese audiences for the first time.

It was a huge pleasure to see it receive such a positive reaction in the world’s biggest single automotive market. This close-to-production insight into the vehicle family based on our Mercedes-Benz Modular Architecture (MMA) offers digital-savvy Chinese customers a feel for the hyper-personalised user experience based on our Mercedes-Benz Operating System (MB.OS).

Our software experts in China have been closely involved in the development of MB.OS from the very start and pay close attention to Chinese feedback. That’s why we brought with us to Auto China demos of some great new digital features and functions of the user interface in the Concept CLA Class.

We want MB.OS to provide our customers in China with an intelligent digital experience based on their own preferences as well as their favourite apps.

And of course, automated driving is another important domain, where the specifics of Chinese roads, traffic and driving styles have a clear impact on how we develop and adapt assistance systems such as our Automatic Lane Change (ALC) function for customers in China.

Furthermore, our MMA vehicles will be equipped with a comprehensive sensor set that feeds our neural network, which enables continuous learning capabilities. This will result in a steady stream of new functions, all provided over-the-air of course. One example will be our point-to-point urban navigation feature. This system can handle challenging urban driving scenarios and supports a fast city rollout due to its “map-less” nature.

But that’s not all I presented in Beijing. Stay “tuned” for more soon…

So below here is my comment on the post, followed by the reply to me by the Advanced UX Director at Mercedes, and take note, the post itself & my comment didn’t say anything about neuromorphic, so my question is, why the reply from the gentleman at Mercedes with a link to the “In The Loop” post over a year ago??

IMG_8102.jpeg
 
Last edited:
  • Like
  • Fire
  • Love
Reactions: 75 users
A tapeout is an essential part of the IP, whether or not we were to progress to SoC.
Quarterly Report out tomorrow? I hope so . With everyone working so hard for success , one would feel that we all deserve something in it to quietly celebrate. If there is good news in it , I will also happy for a Monday or Tuesday release.😊
 
  • Like
  • Thinking
Reactions: 5 users

Damo4

Regular
Can somebody help me try understand something please.
For those that haven’t seen the post by Markus Schafer from Mercedes which was posted an hour or so ago, here it is below, followed by a screenshot of my comment and the Advanced UX Director of Mercedes to me.


Hello Beijing and Auto China 2024! Where better to present our Concept CLA Class to Chinese audiences for the first time.

It was a huge pleasure to see it receive such a positive reaction in the world’s biggest single automotive market. This close-to-production insight into the vehicle family based on our Mercedes-Benz Modular Architecture (MMA) offers digital-savvy Chinese customers a feel for the hyper-personalised user experience based on our Mercedes-Benz Operating System (MB.OS).

Our software experts in China have been closely involved in the development of MB.OS from the very start and pay close attention to Chinese feedback. That’s why we brought with us to Auto China demos of some great new digital features and functions of the user interface in the Concept CLA Class.

We want MB.OS to provide our customers in China with an intelligent digital experience based on their own preferences as well as their favourite apps.

And of course, automated driving is another important domain, where the specifics of Chinese roads, traffic and driving styles have a clear impact on how we develop and adapt assistance systems such as our Automatic Lane Change (ALC) function for customers in China.

Furthermore, our MMA vehicles will be equipped with a comprehensive sensor set that feeds our neural network, which enables continuous learning capabilities. This will result in a steady stream of new functions, all provided over-the-air of course. One example will be our point-to-point urban navigation feature. This system can handle challenging urban driving scenarios and supports a fast city rollout due to its “map-less” nature.

But that’s not all I presented in Beijing. Stay “tuned” for more soon…

So below here is my comment on the post, followed by the reply to me by the Advanced UX Director at Mercedes, and take note, the post itself & my comment didn’t say anything about neuromorphic, so my question is, why the reply from the gentleman at Mercedes with a link to the “In The Loop” post over a year ago??

View attachment 61539

Another stellar post Chapman!
All but confirms it.

For those wanting to save a couple of seconds, see the below link for what Zane Amiralis posted:
https://www.linkedin.com/posts/mark...ity-7021055729805467648-WQHK?utm_source=share
 
  • Like
  • Fire
  • Love
Reactions: 30 users

Evermont

Stealth Mode
Can somebody help me try understand something please.
For those that haven’t seen the post by Markus Schafer from Mercedes which was posted an hour or so ago, here it is below, followed by a screenshot of my comment and the Advanced UX Director of Mercedes to me.


Hello Beijing and Auto China 2024! Where better to present our Concept CLA Class to Chinese audiences for the first time.

It was a huge pleasure to see it receive such a positive reaction in the world’s biggest single automotive market. This close-to-production insight into the vehicle family based on our Mercedes-Benz Modular Architecture (MMA) offers digital-savvy Chinese customers a feel for the hyper-personalised user experience based on our Mercedes-Benz Operating System (MB.OS).

Our software experts in China have been closely involved in the development of MB.OS from the very start and pay close attention to Chinese feedback. That’s why we brought with us to Auto China demos of some great new digital features and functions of the user interface in the Concept CLA Class.

We want MB.OS to provide our customers in China with an intelligent digital experience based on their own preferences as well as their favourite apps.

And of course, automated driving is another important domain, where the specifics of Chinese roads, traffic and driving styles have a clear impact on how we develop and adapt assistance systems such as our Automatic Lane Change (ALC) function for customers in China.

Furthermore, our MMA vehicles will be equipped with a comprehensive sensor set that feeds our neural network, which enables continuous learning capabilities. This will result in a steady stream of new functions, all provided over-the-air of course. One example will be our point-to-point urban navigation feature. This system can handle challenging urban driving scenarios and supports a fast city rollout due to its “map-less” nature.

But that’s not all I presented in Beijing. Stay “tuned” for more soon…

So below here is my comment on the post, followed by the reply to me by the Advanced UX Director at Mercedes, and take note, the post itself nor my comment said anything about neuromorphic, so my question is, why the reply from the gentleman at Mercedes with a link to the “In The Loop” post over a year ago??

View attachment 61539

Hi @chapman89,

Nothing I have read previously from MB suggests we are integrated within the MMA or Concept CLA at this time.

MB have referenced 'neural' previously in the context of Chat GPT for Hey Mercedes and the personality traits for the MBUX Virtual Assistant.

Still doesn't explain why you were pointed to an article which is solely neuromorphic related.

Cheers, I got nothing. :)
 
  • Like
  • Haha
Reactions: 6 users

Damo4

Regular
  • Like
  • Fire
Reactions: 12 users

Rach2512

Regular
Last edited:
  • Like
  • Love
  • Haha
Reactions: 4 users

Diogenese

Top 20
From the edge impulse April newsletter

Edge Impulse insightsnewsletter@edgeimpulse.com

26/4/24 01:13​


...

Webinar: From Prototype to Finished Product with BrainChip

Join Edge Impulse on May 6th for a demonstration of using the BrainChip AkidaTM neuromorphic processor to classify images for edge AI workloads. In this webinar, Low-Power Image Classification With the BrainChip Akida Edge AI Enablement Platform, we will cover X-ray image classification running on an edge device. You'll also learn machine learning model creation, deployment to the BrainChip Akida Enablement Platform, and the path from prototype to product.

https://events.edgeimpulse.com/brai...693&utm_content=304317693&utm_source=hs_email

...

New from the Edge Impulse Blog
Tech for Good: Using AI to Track Rhinos in Africa
Making Strides in Personalized Healthcare
Classifying Medical Imaging On-Device with Edge Impulse and BrainChip
Picking the Brain of AI: BrainChip on the State of Neuromorphic Processing
ML on the Edge with Zach Shelby Ep. 7 — EE Time's Sally Ward-Foxton on the Evolution of AI
Pump Up the Predictions: Prototyping Predictive Maintenance with RASynBoard

1714061096646.png

Launching May 6, 2024​

Join Edge Impulse as we demonstrate using the BrainChip AkidaTM neuromorphic processor to classify images for Edge AI workloads.

This webinar will dive into X-ray image classification running on an Edge device. We’ll cover machine learning model creation, deployment to the BrainChip Akida Enablement Platform, and the path from prototype to product.
 
  • Like
  • Fire
  • Love
Reactions: 66 users

Iseki

Regular
Renesas: You make the Akida2000 chip.
BRN: No! You make the Akida2000 chip.
Renesas: No, You make the Akida2000 chip.
BRN: No! You make the Akida2000 chip.
Renesas: No, You make the Akida2000 chip.
BRN: But you told us what changes you wanted in the chip!
Renesas: We never said we'd manufacture the thing.
BRN: You make the Akida2000 chip.
Renesas: We're selling our own.
BRN: You make the Akida2000 chip.
Renesas: No, You make the Akida2000 chip.
BRN: But we held off because we thought you were. We didn't want to tread on your toes.
Renesas: No. We have to protect our shareholders.
BRN: Who?
Renesas: Good bye.
 
  • Like
  • Haha
  • Love
Reactions: 9 users

Esq.111

Fascinatingly Intuitive.
Morning Chippers ,

Apparently a new release on Twitter.


Regards,
Esq.
 
  • Like
  • Love
  • Fire
Reactions: 44 users

Damo4

Regular
 
  • Like
  • Love
  • Fire
Reactions: 27 users

TECH

Regular
Hi TECH,

just an attempt to sort out the terminology…

The following is how I understand BrainChip’s nomenclature, but I could be wrong, so everybody please feel free to brainstorm and chip in…

In early 2022, shortly after Sean Hehir had joined BrainChip as CEO, “AKD1000” was still used as an umbrella term to describe everything our company had for sale at the time (chip, IP, PCIe board), as evident by the following investor conference presentation slides:

View attachment 61365


These days, however, a distinction appears to be made between the generational iterations of the Akida processor technology platform (categorised as akida, akida 2.0…) as opposed to the physical reference chips (so far AKD1000 and AKD1500).

The AKD1000 and AKD1500 SoCs are both silicon implementations of the Akida technology embodied in BrainChip’s 1st Generation Edge AI neuromorphic processor platform akida (technically speaking akida 1.0).

AKD1000 was implemented with TSMC at 28nm, whereas AKD1500 was taped out on GlobalFoundries’ MCU-friendly 22nm fully depleted silicon-on-insulator (FD-SOI) process, aka GF’s 22FDX technology. As reference chips, they were primarily meant to target prospective IP licencees (as a proof of concept), but at the same time the AKD1000 (on a PCIe board or inside a Dev Kit) benefitted individual developers (professional hardware engineers in companies or academic settings as well as advanced hobbyists) who were not interested in mass production of edge devices and the signing of an IP licence, but instead may have only required a single PCIe Board or Dev Kit for their projects or research (take note that it says on the BrainChip shop website in bold capital letters that development kits are not intended to be used for production purposes). Meanwhile, AKD1000 chips have also been integrated into the VVDN Edge AI Boxes, and the Unigen Cupcake Edge AI Server will soon be offered with a new configuration based on the AKD1500 (?) as an AI option.

Akida 2.0, the neural processing system’s enhanced 2nd Generation, was announced and introduced last year, but purely as an IP offering, productised in three different variations: akida-E, akida-S and akida-P, depending on where in the Edge AI spectrum (sensor edge < server edge) its prowess is required.

AKD2000, however, doesn’t exist - at least not yet.
The way I understand it, AKD2000 would be the name
of BrainChip’s (hypothetical) reference chip based on Akida Gen 2, which may or may not materialise.

Let’s recall what was said earlier this year:
In an interview with Jim McGregor from TIRIAS Research during CES 2024 (January 9-12), Todd Vierra replied the following to his interview partner’s comment “And correct me if I’m wrong, but this is the Akida 2?”
Todd: “This is actually all ran [sic] on Akida 1 hardware. Akida 2, erm - we are in the process of taping out and we’ll get that silicon back a little bit later, but these are all just Gen 1...” (from 9:26 min). His statement about an imminent tapeout seemed to confirm what (according to FF) Sean Hehir had told select shareholders in the November 2023 Sydney “secret meeting”.

Surprisingly, a mere seven weeks later, during the Virtual Investor Roadshow (February 27), neither Sean Hehir nor Tony Lewis mentioned anything at all about a tape-out in progress, but instead argued a second generation reference chip as proof of concept was unnecessary, while at the same time not totally excluding a potential future tape-out complementing their core IP business. However, our CTO made it clear that it is definitely not their intention to manufacture chips on a large scale:


From 43:17
Roger Manning’s question: “Given BrainChip’s business model is to largely sell its akida IP, will it be necessary to prove each new version of akida in silicon?”

Tony: “So I think the big question was, will this event-based paradigm yield, erm can it be done and will it yield some benefits to customers, and I think we achieved that by taping out our earlier generation of products, and so we’ve already achieved that. And it’s my belief that there is only marginal benefit in taping out the next generation, we already proved the main points of it. And clearly we don’t want to start to manufacture chips on a large scale. We’d be competing with our customers and that would really break our business model right now.

Sean: “Yeah, and to be clear, a lot of work has gone on with our ability to simulate workloads in Generation 2 as well. As Tony said, we certainly have reference chips in Generation 1, and the typical engagement course that we work with IP license prospects is we allow them to run models on there and/or simulate them in our simulation tools that we have for Generation 2, so Roger, I would say stay tuned, we may or may not, erm, but right now, there is no need for us to do that.”




Now the way I understood the “We’d be competing with our customers” comment is not for fear of treading on their customers’ or a specific customer’s toes (as other TSE users have interpreted it), but because they would shoot themselves in the foot by doing that: after all, it would be less profitable for our company when customers could just buy those chips off the shelves to utilise them in their in-house development (and hence save a lot of money, as there won’t be any follow-up costs for them) rather than having to pay an initial IP license fee and future royalties.

Does my interpretation make sense or am I overlooking something here?



And as for the next iteration of the Akida processor family, we don’t even know, yet, whether it will be called akida 3.0 or akida 2.X…

From min 47:41 min:

Sean Hehir: “… but round Generation 3, if you will: You know, I mentioned earlier in my slide that when I talked about our product planning and our execution cycle, if you will, you could see we are always planning on this product and we are always looking for improvements on our IP offering right there. Now whether we call it formally “Generation 3” or “2 something x” - but yes, we are in the middle of a planning cycle right now to make some changes and we’ll make announcements over time.”


So what we can say with confidence (provided we believe our CEO’s words) is that the advent of the Akida neural processor family’s 3rd generation is approaching… Whether there will ever be an AKD3000, though, is uncertain.

Regards
Frangipani


P.S.: I still can’t get my head around Todd Vierra’s statement, though. Would they really have backed out at the last minute, eg due to financial constraints?
Or was he possibly referring to a tapeout not of AKD2000 as a reference chip, but to a tape-out of a SoC by a specific customer that includes Akida 2.0 IP, along the lines of what DingoBorat said?



But then again, would he really say “…we’ll get that silicon back a little later”?
I suppose only if it was a joint development, such as possibly one with Socionext or Tata Elxsi (which could explain the different colour of the mysterious Custom SoC on that presentation slide)? As otherwise, wouldn’t such a SoC be taped out by the customer itself rather than by / in collaboration with BrainChip? I am a bit confused here. 🤔

It seems too early as a mere placeholder for the planned integration of Akida IP into the Frontgrade Gaisler SoC, which according to ESA’s Laurent Hili “We aim to tape out ideally before the end of the year, beginning of next year” (from 47:25 min shortly before the end of the mid -March BrainChip podcast Episode 31). After all, they could have labelled it “prospective Customer Custom SoC”, but the way it is presented on the slide, I agree with DingoBorat that it does appear to be an already existing physical implementation, indeed. Mmmmh… 🤔


Hi Frangipani...

There is/was a lot to unpack in your post.

This morning I have gone through 3 years of emails trying to find an answer...when Sean came onboard, within 3 months of his appointment and after the Board approved his 5 year business plan things changed with regards us becoming a mass producer of chips.

We considered storing inventory in Japan at one point, then we decided to adopt the "placing inventory orders on the short holding" business
plan when we actually received clients orders ( I have forgotten the name) so in effect we were never holding large amounts of chips and having to outlay large sums upfront for production, with the hope that some client would buy them, if that makes sense.

The business plan appears to be very nimble, flexible and fluid, whether Todd was supposed to make the comment he did or not I don't know, but you are 100% correct it does become rather confusing.

AKD 1000 was going to be massed produced, then it became described as our reference chip, AKD 1500 I believe was produced for two reasons, one being as a reference chip, proving the technology was capable of functioning in a different format and could be manufactured out of different foundries, and secondly because a client requested it, that's my understanding.

AKD 2000 was always coming as our 2nd generation chip, but the company did seem to reaffirm a number of times that we are following the ARM business model in so many words and wanted to be known first and foremost as an IP provider only, with some flexibility if approached by other vendors/clients etc.


"We plan to progress to AGI in small steps. We are working on new architectures to add more intelligence to every step. The most important aspects are that we stay ahead of the competition and build products that appeal to a broader market to ensure that BrainChip is a commercial success. Each product we build will address a solution for a particular market segment. We do not focus on the numbers game of how many chips we will tape out to get to AGI over how many years; that is irrelevant and may well be more than seven chips if the market demands it. We may build specific IPs for select customers".

I don't know anymore than that currently, but I do believe one of our greatest assets is that we can turn on a dime if required and adapt very
quickly to a customers demands/needs, where much larger companies take a lot longer to turn around to change course to accommodate those
demands/needs.

Kind regards....Tech :)
 
  • Like
  • Fire
  • Thinking
Reactions: 33 users

mrgds

Regular
How good is Teslas latest FSD iteration? ................ this fu#@ing good.

 
  • Like
  • Wow
  • Fire
Reactions: 14 users
A nice overview of neuromorphic technology and brainchip has been noted in the mix.


I noticed the Snapchat logo alongside us, which was strange, but it turns out they acquired Grai Matter Labs last month :unsure:.

https://bits-chips.nl/artikel/grai-matter-labs-quietly-snapped-up-by-snap/




1714091816015.png
 
  • Like
  • Fire
  • Love
Reactions: 43 users
Top Bottom