Fullmoonfever
Top 20
Welcome.Thanks Fmf,
My short term memory is somewhat deficient.
Spikes not working at the mo....it is Fri
Welcome.Thanks Fmf,
My short term memory is somewhat deficient.
My reason for the question is the paragraph below that was written in the Akida 1500 tape out article:Hypothetical question. Could taping out Akida 1500 on Global Foudries be enough evidence that Akida works on GF technology for Oculi to have the confidence to proceed with developing their own chip incorporating Akida IP through GF?
I feel there is too many coincidences (at least in my head) to ignore.
I don't think it's just Oculi that were interested in seeing Akida performing on GlobalFoundries' 22FDX platform. It's a popular platform for efficient industrial and consumer applications and has had several years to mature and gain adoption by the industry.My reason for the question is the paragraph below that was written in the Akida 1500 tape out article:
"The tape-out was completed using GlobalFoundries’ 22nm fully depleted silicon-on-insulator (FD-SOI) technology and is being described as a milestone in validating BrainChip’s IP across different processes and foundries, providing its partners with varied global manufacturing options."
The Brainchip/Global Foudries tape out was announced on 29th Jan 2023.
Two days ago, Oculi, whom Brainchip is working with, announced that they are using Global Foundries to help them produce their latest chip (did the Akida 1500 tape out on Global Foudries technology give them what they needed to go ahead with their own GF chip?):
Oculi say this about their future chip:
"Oculi’s new vision is ideal for edge applications such as always-on gesture/face/people tracking and low-latency eye tracking, while alternative solutions are too slow, big, and power inefficient. GF is an excellent partner to enable us to quickly get our product to our customers."
Just found an article from yesterday which outlines Oculi and GF producing 3 chips but the first is using their 55LPx whereas AKD1500 is in 22FDX.My reason for the question is the paragraph below that was written in the Akida 1500 tape out article:
"The tape-out was completed using GlobalFoundries’ 22nm fully depleted silicon-on-insulator (FD-SOI) technology and is being described as a milestone in validating BrainChip’s IP across different processes and foundries, providing its partners with varied global manufacturing options."
The Brainchip/Global Foudries tape out was announced on 29th Jan 2023.
Two days ago, Oculi, whom Brainchip is working with, announced that they are using Global Foundries to help them produce their latest chip (did the Akida 1500 tape out on Global Foudries technology give them what they needed to go ahead with their own GF chip?):
Oculi say this about their future chip:
"Oculi’s new vision is ideal for edge applications such as always-on gesture/face/people tracking and low-latency eye tracking, while alternative solutions are too slow, big, and power inefficient. GF is an excellent partner to enable us to quickly get our product to our customers."
Another good reason vs 28nm CMOS.Just found an article from yesterday which outlines Oculi and GF producing 3 chips but the first is using their 55LPx whereas AKD1500 is in 22FDX.
Maybe @Diogenese can impart some wisdom as to the outcome here?
Get the feeling maybe no crossover as yet?
Maybe the AKD1500 will be testing etc for the stated next iterations of the SPU.
![]()
Smart vision startup Oculi turns to GF for next chips
Oculi, a fabless company seeking to add intelligence to image sensors at the pixel level, has partnered with chipmaker GlobalFoundries Inc.www.eenewseurope.com
Under the agreement GlobalFoundries will manufacture the SPU (sensing and processing unit) S12 software-defined vision sensor in its 55LPx manufacturing process. The chip is intended to be used for applications in smart devices and homes, industrial, IoT, automotive markets and wearables including AR/VR.
The SPU S12 will be the first of three sensors to follow on from the SPU S11 which Oculi offers on a series of PCB-based platforms.
For outline of GF chips.
![]()
Industrial IoT | GlobalFoundries
Bring forth Industry 4.0 with GF From agriculture to manufacturing, intelligent connectivity and AI are leading the way in the […]gf.com
Industrial MCUs using 22FDX® with eNVM
GlobalFoundries® (GF®) 22FDX® solutions, featuring low power, adaptive body bias, analog scaling and robust eNVM capabilities, enable integrated, area and power-optimized industrial MCUs. Solution options include eMRAM, low-cost charge trap technology (CTT) and CB-RAM embedded memory for faster TTM and wakeup times.
Pumped up, power-efficient performance
22FDX® offers best-in-class performance at the industry’s lowest operating voltage for bulk CMOS technologies (0.4 V) and 1 pA/µm for ultra-low standby leakage. It features an eMRAM NVM with >100x lower write power* that enables frequent, power-saving shutdowns to help designers extend battery life while boosting processing capability.
More function, in less space
22FDX® solutions enable designers to develop RF front-end modules (FEMs) with outstanding PA efficiency, LNA noise figure, and switch insertion loss benefits. These FEMs, baseband and eMRAM elements can be integrated into a single IIoT SoC that helps designers combine the features needed to meet goals, while significantly reducing overall area—and costs.
Design, made simple
The 22FDX® portfolio of silicon-proven, MCU-optimized IP, along with a broad range of services and solutions available through GF and the FDXcelerator™ partner program, can help designers reduce development time and have confidence in first-time-right results in hardware.
Industrial MCUs using 55LPX and 40LP with eNVM and 28SLP-ESF3
GlobalFoundries® (GF®) 55LPX and 40LP with eNVM and 28SLP-ESF3, built on bulk CMOS platforms, are optimized for range of power-performance and cost-sensitive MCU applications, including automotive, baseband SoCs, mobile multimedia, digital TVs/STBs, IoT and industrial. The solutions enable designers to leverage logic, analog, RF, ULP SRAM/logic combinations, high-K metal gate technology and high reliability, along with on-board memory (eNVM) for faster wakeup times, reduced system cost and improved security:
55LPx features high reliability (automotive-grade IP) and high-density, low-power SRAM
GF’s 40LP is the only 40 nm automotive grade 0 offering in the industry
28SLP is optimized for power, performance and die cost, when flexible mixed-technology options for RF and ultra-low power are required
ESF3 eFlash adds robust quality and reliability, offering zero failure rates in harsh temperature conditions
High voltage? No problem.
55LPx solutions are excellent fits for analog and power devices operating at 30 V and beyond; it is optimized for integrated analog, power and mixed-signal applications such as PMICs for mobile devices, audio amplifiers and applications requiring dense digital, analog and power elements.
I think we all need to have another beer or 2 and contemplate is dilemma SladiousHypothetical question. Could taping out Akida 1500 on Global Foudries be enough evidence that Akida works on GF technology for Oculi to have the confidence to proceed with developing their own chip incorporating Akida IP through GF?
I feel there is too many coincidences (at least in my head) to ignore.
Was looking for any connections of VVDN and Moschip (you know the joint presso that seems to have faded away at the mo) just for diff angle.![]()
VVDN strengthens its services portfolio with the addition of Automotive Engineering and Manufacturing Services for global markets
/PRNewswire/ -- VVDN Technologies, a global provider of embedded product engineering, manufacturing and digital services and solutions company, announced that...www.prnewswire.com
I am going to embarrass myself here. Can't Akida IP fit with any chip. I know that Akida is sensor agnostic. If we are proved on 22FDX would that be enough to suggest we will work on 55LPx (I ask this question having no idea what FDX or LPx is, nor what the difference is between 55 and 22......my guess is size. Makes sense to me that Oculi's chip is bigger. It needs the extra room for Akida....I'm clueless)? Who knows how long Oculi and Brainchip have been engaged, it could be a year or more.Just found an article from yesterday which outlines Oculi and GF producing 3 chips but the first is using their 55LPx whereas AKD1500 is in 22FDX.
Maybe @Diogenese can impart some wisdom as to the outcome here?
Get the feeling maybe no crossover as yet?
Maybe the AKD1500 will be testing etc for the stated next iterations of the SPU.
![]()
Smart vision startup Oculi turns to GF for next chips
Oculi, a fabless company seeking to add intelligence to image sensors at the pixel level, has partnered with chipmaker GlobalFoundries Inc.www.eenewseurope.com
Under the agreement GlobalFoundries will manufacture the SPU (sensing and processing unit) S12 software-defined vision sensor in its 55LPx manufacturing process. The chip is intended to be used for applications in smart devices and homes, industrial, IoT, automotive markets and wearables including AR/VR.
The SPU S12 will be the first of three sensors to follow on from the SPU S11 which Oculi offers on a series of PCB-based platforms.
For outline of GF chips.
![]()
Industrial IoT | GlobalFoundries
Bring forth Industry 4.0 with GF From agriculture to manufacturing, intelligent connectivity and AI are leading the way in the […]gf.com
Industrial MCUs using 22FDX® with eNVM
GlobalFoundries® (GF®) 22FDX® solutions, featuring low power, adaptive body bias, analog scaling and robust eNVM capabilities, enable integrated, area and power-optimized industrial MCUs. Solution options include eMRAM, low-cost charge trap technology (CTT) and CB-RAM embedded memory for faster TTM and wakeup times.
Pumped up, power-efficient performance
22FDX® offers best-in-class performance at the industry’s lowest operating voltage for bulk CMOS technologies (0.4 V) and 1 pA/µm for ultra-low standby leakage. It features an eMRAM NVM with >100x lower write power* that enables frequent, power-saving shutdowns to help designers extend battery life while boosting processing capability.
More function, in less space
22FDX® solutions enable designers to develop RF front-end modules (FEMs) with outstanding PA efficiency, LNA noise figure, and switch insertion loss benefits. These FEMs, baseband and eMRAM elements can be integrated into a single IIoT SoC that helps designers combine the features needed to meet goals, while significantly reducing overall area—and costs.
Design, made simple
The 22FDX® portfolio of silicon-proven, MCU-optimized IP, along with a broad range of services and solutions available through GF and the FDXcelerator™ partner program, can help designers reduce development time and have confidence in first-time-right results in hardware.
Industrial MCUs using 55LPX and 40LP with eNVM and 28SLP-ESF3
GlobalFoundries® (GF®) 55LPX and 40LP with eNVM and 28SLP-ESF3, built on bulk CMOS platforms, are optimized for range of power-performance and cost-sensitive MCU applications, including automotive, baseband SoCs, mobile multimedia, digital TVs/STBs, IoT and industrial. The solutions enable designers to leverage logic, analog, RF, ULP SRAM/logic combinations, high-K metal gate technology and high reliability, along with on-board memory (eNVM) for faster wakeup times, reduced system cost and improved security:
55LPx features high reliability (automotive-grade IP) and high-density, low-power SRAM
GF’s 40LP is the only 40 nm automotive grade 0 offering in the industry
28SLP is optimized for power, performance and die cost, when flexible mixed-technology options for RF and ultra-low power are required
ESF3 eFlash adds robust quality and reliability, offering zero failure rates in harsh temperature conditions
High voltage? No problem.
55LPx solutions are excellent fits for analog and power devices operating at 30 V and beyond; it is optimized for integrated analog, power and mixed-signal applications such as PMICs for mobile devices, audio amplifiers and applications requiring dense digital, analog and power elements.
The short definition of Shorters=W@#€ers!!Hi Everyone
Just in regards to shorters, I'm trying to understand the way they work.
My understanding, which I could be completely wrong so please correct me.
They borrow a heap of shares, do they sit on them and then decide when is a good time for them to sell them, they then sell a massive amount so as to spook the market, triggering stop losses, thus this reduces the SP all the more and then when they think they've hit rock bottom they quickly buy back in so they can then return the borrowed shares? If this is the case then can we expect some frantic buying from them over the next couple of days/weeks, which should recover the SP?
Also thinking about the SP, I've been in since my baby girl was born, she's now 10, I see there are a lot of people here who are not happy about the SP and it seems that this is their only focus. I'm curious say if the SP had shot up to $1, would you sell or hold knowing what the potential is for BRN, I know I wouldn't sell, if you would still hold how is that any different to holding now. I appreciate some may need the money now which if that is the case I feel for you. I think people were expecting to see revenue towards the end of last year but the plan changed from a Chip supplier to selling IP which meant changes had to be made which took time to implement. I think that we could all agree that this will be far more beneficial for BRN in the long run.
Happy to be corrected, I don't often contribute here so please don't shoot me!
Thank you to all the great posts and finds, they are very much appreciated.
Also sorry if a similar question has already been asked about shorters.
While companies may still choose to remain tight-lipped through NDAs in order to retain a competitive advantage in the present, I do think that the day will come when BrainChip's technology will become a lot more mainstream and companies will start to advertise its use in order to attract consumers. I like to call that the "Intel Inside" effect.
Don't forget NVISO and the benchmarks that they published recently compared to other competitors. The processing performance at lower clock speeds and lower memory requirements compared to the Jetson Nano and Tegra GPU are valuable selling points for them if they want to get attention from their potential customers.And why one of our happy customers did this so early.
I.e. before full implementation Jan 3rd 2022.
View attachment 29887
They have seen it, tested it implemented in operating systems and it blew them away, They wanted to be the first to profess their association with this new Ubiqitous technology. They know how far it will be adopted and can say they saw it first.
Follow the leader is a good strategy for going places.
Mercedes is the leadrer in the automotive industry and Brainchip is the leader in ultimate edge AI industry.
Onya Steve, always appreciate your contribution.I was cheeky & asked Prophesee which neuromorphic processor their metavision sensor will be paired with in Qualcomm's Snapdragon 8 Gen 2, however, as expected they didn't reply. I specifically asked if it's Akida or another processor. Nothing, all quiet & not even a thanks for my enquiry.
The thing about FD-SoI is that it is more power efficient than standard CMOS because the insulation base isolates the active transistors from potential leakage paths through the silicon base.I don't think it's just Oculi that were interested in seeing Akida performing on GlobalFoundries' 22FDX platform. It's a popular platform for efficient industrial and consumer applications and has had several years to mature and gain adoption by the industry.
I think this latest tape-out of AKD1500 and the choice to create the next batch of reference chips using the GlobalFoundries tech is likely the linchpin that will get signed IP contracts with some unnamed customers.
What I am hoping for at this point is that said customers have been working on prototypes of whatever products they are choosing to bring to market, and refactoring time from prototype to consumer good will be dramatically shortened, pending the successful testing of this generation of Akida.
While companies may still choose to remain tight-lipped through NDAs in order to retain a competitive advantage in the present, I do think that the day will come when BrainChip's technology will become a lot more mainstream and companies will start to advertise its use in order to attract consumers. I like to call that the "Intel Inside" effect.
Good research. I agree that there does not seem to be a direct tie with Oculi as one of the customers potentially waiting on the 22FDX platform.The thing about FD-SoI is that it is more power efficient than standard CMOS because the insulation base isolates the active transistors from potential leakage paths through the silicon base.
We already know that Akida is highly efficient and can run on a small battery for a long time. So adding extra battery life by eliminating leakage currents suggests a need for a very long time between battery changes. This in turn suggests remote or inaccessible applications.
As someone pointed out oculi are using GF55Lpx, where 55 refers to nm. In addition this is not the same as FD-SoI.
https://www.newelectronics.co.uk/content/news/oculi-forms-strategic-partnership-with-globalfoundries
The OCULI SPU S12 is the first of 3 product lines that Oculi says will disrupt and set a new standard in vision technology. It is based on GF’s 55LPx which is a feature-rich platform that supports Radio Frequency (RF), ultra-low power (ULP), embedded non-volatile memory (eNVM) and high voltage BCDLite options, making the platform a solution for System on Chip (SoC) integration to enable more functionality, less energy consumption and smaller form factor electronic applications.
https://gf.com/markets/home-industrial-iot/industrial-iot/
Industrial MCUs using 55LPX and 40LP with eNVM and 28SLP-ESF3
GlobalFoundries® (GF®) 55LPX and 40LP with eNVM and 28SLP-ESF3, built on bulk CMOS platforms, are optimized for range of power-performance and cost-sensitive MCU applications, including automotive, baseband SoCs, mobile multimedia, digital TVs/STBs, IoT and industrial. The solutions enable designers to leverage logic, analog, RF, ULP SRAM/logic combinations, high-K metal gate technology and high reliability, along with on-board memory (eNVM) for faster wakeup times, reduced system cost and improved security:
High voltage? No problem.
- 55LPx features high reliability (automotive-grade IP) and high-density, low-power SRAM
- GF’s 40LP is the only 40 nm automotive grade 0 offering in the industry
- 28SLP is optimized for power, performance and die cost, when flexible mixed-technology options for RF and ultra-low power are required
- ESF3 eFlash adds robust quality and reliability, offering zero failure rates in harsh temperature conditions
55LPx solutions are excellent fits for analog and power devices operating at 30 V and beyond; it is optimized for integrated analog, power and mixed-signal applications such as PMICs for mobile devices, audio amplifiers and applications requiring dense digital, analog and power elements.
Clearly, oculi will need to work with a NN and Anil has said BrainChip is working with oculi, and Akida is probably the best match, but the fact that oculi are working with GF at LPx55nm does not establish the connexion with Akida FD-SoI at 22 nm . Further more, they are built on different wafers so could not be a single SoC. FD-SoI has a layer of insulation on the silicon wafer and the transistors are built on top of the insulation.
That is not to say that oculi and BrainChip will not collaborate with GF in a future project.
Interesting you raise good points on remote and inaccessible, abeit in reference to power.The thing about FD-SoI is that it is more power efficient than standard CMOS because the insulation base isolates the active transistors from potential leakage paths through the silicon base.
We already know that Akida is highly efficient and can run on a small battery for a long time. So adding extra battery life by eliminating leakage currents suggests a need for a very long time between battery changes. This in turn suggests remote or inaccessible applications.
As someone pointed out oculi are using GF55Lpx, where 55 refers to nm. In addition this is not the same as FD-SoI.
https://www.newelectronics.co.uk/content/news/oculi-forms-strategic-partnership-with-globalfoundries
The OCULI SPU S12 is the first of 3 product lines that Oculi says will disrupt and set a new standard in vision technology. It is based on GF’s 55LPx which is a feature-rich platform that supports Radio Frequency (RF), ultra-low power (ULP), embedded non-volatile memory (eNVM) and high voltage BCDLite options, making the platform a solution for System on Chip (SoC) integration to enable more functionality, less energy consumption and smaller form factor electronic applications.
https://gf.com/markets/home-industrial-iot/industrial-iot/
Industrial MCUs using 55LPX and 40LP with eNVM and 28SLP-ESF3
GlobalFoundries® (GF®) 55LPX and 40LP with eNVM and 28SLP-ESF3, built on bulk CMOS platforms, are optimized for range of power-performance and cost-sensitive MCU applications, including automotive, baseband SoCs, mobile multimedia, digital TVs/STBs, IoT and industrial. The solutions enable designers to leverage logic, analog, RF, ULP SRAM/logic combinations, high-K metal gate technology and high reliability, along with on-board memory (eNVM) for faster wakeup times, reduced system cost and improved security:
High voltage? No problem.
- 55LPx features high reliability (automotive-grade IP) and high-density, low-power SRAM
- GF’s 40LP is the only 40 nm automotive grade 0 offering in the industry
- 28SLP is optimized for power, performance and die cost, when flexible mixed-technology options for RF and ultra-low power are required
- ESF3 eFlash adds robust quality and reliability, offering zero failure rates in harsh temperature conditions
55LPx solutions are excellent fits for analog and power devices operating at 30 V and beyond; it is optimized for integrated analog, power and mixed-signal applications such as PMICs for mobile devices, audio amplifiers and applications requiring dense digital, analog and power elements.
Clearly, oculi will need to work with a NN and Anil has said BrainChip is working with oculi, and Akida is probably the best match, but the fact that oculi are working with GF at LPx55nm does not establish the connexion with Akida FD-SoI at 22 nm . Further more, they are built on different wafers so could not be a single SoC. FD-SoI has a layer of insulation on the silicon wafer and the transistors are built on top of the insulation.
That is not to say that oculi and BrainChip will not collaborate with GF in a future project.