D
Deleted member 3781
Guest
Not even 20 
I see the big news coming from tomorrow any time now
I see the big news coming from tomorrow any time now
BRN Discussion Ongoing
- Thread starter TechGirl
- Start date
We all know that devices can communicate with each other, eg i see my phone messages on my lap top. Edge chips whether Neuromorphic or not in devices can communicate with each other if connected via some type of network not necessarily connected to the cloud. For Edge devices the network would mostly not be cloud connected (why would you bother).?
This where AKIDA shines with its on chip learning capabilities compared to other neuromorphic or traditional 'dumb' chips with no on chip learning capabilities.
Chips in autos, robots etc need to communicate. Robots in a team may need to send each other signals .
Its easy to see why Parsons see our on chip learning adaptability as so important. Great for space and commercial applications. Care would need to be taken in defense with networks even though will not be could connected.
Its also easy to how AKIDA well ahead of the pack.
I ask Chat GPT to summarise. Its common sense but saved me typing it.
This where AKIDA shines with its on chip learning capabilities compared to other neuromorphic or traditional 'dumb' chips with no on chip learning capabilities.
Chips in autos, robots etc need to communicate. Robots in a team may need to send each other signals .
Its easy to see why Parsons see our on chip learning adaptability as so important. Great for space and commercial applications. Care would need to be taken in defense with networks even though will not be could connected.
Its also easy to how AKIDA well ahead of the pack.
I ask Chat GPT to summarise. Its common sense but saved me typing it.
Comparison Table: With vs. Without OnâChip Learning
| Aspect | Akida with OnâChip Learning | Chips without OnâChip Learning |
|---|---|---|
| Adaptability | Can learn new patterns locally (e.g., new anomalies, new commands) and immediately use them in communication. | Limited to preâtrained models; cannot adapt on their own. Must be retrained in the cloud or offline and redeployed. |
| Communication Content | Sends newly learned events or adapted patterns to other devices (e.g., ânew anomaly detectedâ). | Sends only preâdefined outputs from its fixed model (e.g., âclass = 2â). |
| Collaboration | Devices can share discoveries: one node learns something new and alerts others, enabling systemâwide adaptation. | Devices can only share fixed detections; if one encounters a new pattern, others wonât recognize it until all are updated externally. |
| Autonomy | High â devices can selfâimprove and coordinate without cloud or human intervention. | Lower â devices depend on external retraining and updates to handle new situations. |
| Resilience | Ideal for remote or disconnected environments (spacecraft, autonomous vehicles, industrial plants) where cloud updates arenât possible. | Vulnerable in disconnected environments â performance degrades if new patterns appear that werenât in the original training. |
| Efficiency of Communication | Communicates only meaningful, newly learned events, reducing bandwidth. | Communicates fixed outputs, which may include irrelevant or incomplete information if the model is outdated. |
| System Intelligence | Creates a distributed, adaptive network â like a ânervous systemâ that learns and shares in real time. | Creates a static network â devices can coordinate, but only within the limits of their preâtrained knowledge. |
Last edited:
Interesting.
www.linkedin.com
#harishankarnarayanan | Pravin Karde | 18 comments
#HarishankarNarayanan | 18 comments on LinkedIn
âSony invested in Brainchipâ
Here might be why.
"
ie Sony cameras also comes with an FPGA chip, which can run akida2.
While akida may not yet be in the sensor chip ( a la Prophesee) we might be next to it, performing deeper inferencing on the entire image
thinklucid.com
Here might be why.
"
Sony IMX501: Vision and Intelligence in One Package
Parallel DSP Cores for Efficient AI Inference
The main engine for AI processing is the DSP subsystem core. This core is comprised of a High-Computation-Intensity (CI) DSP core, Standard-CI DSP core, and Tensor DMAs (TDMAs). The High-CI and Standard-CI DSPs work in parallel, executing neural network operations and moving the data directly into TDMA and then into L2 memory. Image and inference data are then transferred via the MIPI interface to the cameraâs FPGA."ie Sony cameras also comes with an FPGA chip, which can run akida2.
While akida may not yet be in the sensor chip ( a la Prophesee) we might be next to it, performing deeper inferencing on the entire image
Sony IMX501 AI Sensor and Brain Builder for Edge AI Processing
AI at the edge: LUCIDâs Triton Smart camera uses Sonyâs IMX501 for efficient on-camera inference for smart manufacturing, logistics, and embedded vision systems
thinklucid.com
BrainShit
Regular
âSony invested in Brainchipâ
Here might be why.
"
Sony IMX501: Vision and Intelligence in One Package
Parallel DSP Cores for Efficient AI Inference
The main engine for AI processing is the DSP subsystem core. This core is comprised of a High-Computation-Intensity (CI) DSP core, Standard-CI DSP core, and Tensor DMAs (TDMAs). The High-CI and Standard-CI DSPs work in parallel, executing neural network operations and moving the data directly into TDMA and then into L2 memory. Image and inference data are then transferred via the MIPI interface to the cameraâs FPGA."
ie Sony cameras also comes with an FPGA chip, which can run akida2.
While akida may not yet be in the sensor chip ( a la Prophesee) we might be next to it, performing deeper inferencing on the entire image
Sony IMX501 AI Sensor and Brain Builder for Edge AI Processing
AI at the edge: LUCIDâs Triton Smart camera uses Sonyâs IMX501 for efficient on-camera inference for smart manufacturing, logistics, and embedded vision systems
Sony's IMX501 intelligent vision sensor contains its own proprietary parallel DSP cores for efficient AI inference and on-sensor processing, designed in-house, and uses a unique architecture to allow AI tasks to be run at the edge without external processors or cloud dependency.
Source: https://www.provideocoalition.com/s... Processor) dedicated to AI signal processing
Sure. There will be inferencing on two levels. On the pixels (can you make the image sharper) and on the image on the FPGA chip ( is it a duck?) and akida could come in the second category is my point.
I only say this because the sony adverts show a lot of image composition boxes..
Last edited:
BrainShit
Regular
Right, Akida could operate well in the second category (classifying objects in an image after the pixel-level processing is done) as an FPGA ... but I really doubt that will happen.
Because the models are used are specifically optimized for the memory and processing constraints of Sony IMX500/IMX501-equipped edge sensing devices... and Sonyâs Brain Builder tooling (AITRIOS) is used in the workflows to train and deploy models targeted.
In other words, I really belife they're using their own developments.
Bascom hunter.
The market is potentially huge.
" THE NAVY BENEFIT
The U.S. Navy can incorporate the technology developed into both modern and legacy weapons systems (via HOST and SOSA alignment), expect to outperform modern GPU systems running cutting edge algorithms; and require exceptionally low power usage. Moreover, this technology is scalable and, as mission needs evolve, super charging applications of interest to the Navy and the Department of Defense."
The point is that Bascom Hunter can integrate its card with AKIDA once testing is finalised straight into existing equipment.
Parsons should be able to do the same so our wait once the chips have been delivered may not be that long if they initially for drones.
Manned craft will be longer of course.
My bold above.
The market is potentially huge.
" THE NAVY BENEFIT
The U.S. Navy can incorporate the technology developed into both modern and legacy weapons systems (via HOST and SOSA alignment), expect to outperform modern GPU systems running cutting edge algorithms; and require exceptionally low power usage. Moreover, this technology is scalable and, as mission needs evolve, super charging applications of interest to the Navy and the Department of Defense."
The point is that Bascom Hunter can integrate its card with AKIDA once testing is finalised straight into existing equipment.
Parsons should be able to do the same so our wait once the chips have been delivered may not be that long if they initially for drones.
Manned craft will be longer of course.
My bold above.
Last edited:
Bascom Hunter Technologies developed an interference excision system (IES) solution, and its applications have gone far beyond the original intent. Unlike conventional radio frequency and digital interference mitigation or cancellation techniques, this technology provides previously unmatched levels of cancellation by using an electro-optic-based interference cancellation technique with more than 10,000 times the signal power range, all within a small package. This approach allows for the removal of strong interferers in the same channel as the signal of interest and allows for improved communications operations in a degraded or denied operating environment. This revolutionary effort paved the way for other interference excision devices and systems that target different types of interference for SATCOM systems. Additionally, it opens the door for new and advanced technologies and capabilities to improve resiliency to SATCOM systems beyond UHF.
I wonder if there is any overlap between the cancellation of unwanted electric signals and unwanted acoustic (hearing aid) signals.
Is it just a matter of having a model adapted for wireless signals?
Wireless signals are certainly within TENNs wheelhouse.
Who amongst us doesn't see us succeeding?
That's the question.
This sector isn't all about us and our technology, I understand that, for us to even think of dictating how the edge ai market will play out over the next 3 years plus would be a totally arrogant view to hold, one which our company, I'd suggest, wouldn't even contemplate.
Unless we establish ourselves in this market on our on two feet with real growing revenue streams, we are definitely a takeover target by a small number of contenders, one's who we are all aware of..keep an open mind, this share market bullshit will come to an abrupt end soon enough, let's go Sean, it's time to giddy up!
That's the question.
This sector isn't all about us and our technology, I understand that, for us to even think of dictating how the edge ai market will play out over the next 3 years plus would be a totally arrogant view to hold, one which our company, I'd suggest, wouldn't even contemplate.
Unless we establish ourselves in this market on our on two feet with real growing revenue streams, we are definitely a takeover target by a small number of contenders, one's who we are all aware of..keep an open mind, this share market bullshit will come to an abrupt end soon enough, let's go Sean, it's time to giddy up!
D
Deleted member 3781
Guest
BrainShit
Regular
Who amongst us doesn't see us succeeding?
That's the question.
This sector isn't all about us and our technology, I understand that, for us to even think of dictating how the edge ai market will play out over the next 3 years plus would be a totally arrogant view to hold, one which our company, I'd suggest, wouldn't even contemplate.
Unless we establish ourselves in this market on our on two feet with real growing revenue streams, we are definitely a takeover target by a small number of contenders, one's who we are all aware of..keep an open mind, this share market bullshit will come to an abrupt end soon enough, let's go Sean, it's time to giddy up!
It is possible for BrainChip to be overtaken in general, but such an acquisition involves specific complexities due to national security concerns and government regulations... because BrainChip got some U.S. military contracts.
Source 1: https://www.governmentcontractslawb...g-government-contractors-and-their-suppliers/
Source 2: https://www.nationaldefensemagazine...ould-see-another-wave-of-mergers-acquisitions
Frangipani
Top 20
A couple of years before I became a BRN shareholder, I came across a company from Austria called g.tec medical engineering GmbH (https://www.gtec.at), founded in 1999 by Christoph Guger and GĂźnter Edlinger as a spin-out of TU Graz, and one of their products called mindBEAGLE, which uses Brain-Computer-Interface (BCI) technology for assessing patients suffering from disorders of consciousness or locked-in syndrome, can help with outcome prediction, and even provides very basic communication with some of them.
Over the past 25+ years, g.tec medical engineering have specialised âin developing high performance brain-computer interfaces and neurotechnologies for both invasive and non-invasive recordings in research and clinical settingsâ (https://www.gtec.at/about/) and are one of the leading companies, if not THE leading company, in this field world-wide.
In a November 2024 interview (https://www.gtec.at/2024/11/04/leading-the-bci-field/), Co-Founder and Co-CEO Christoph Guger, who has degrees from both Johns Hopkins University and TU Graz, shared the following about g.tec medical engineeringâs impressive journey:
âWe sell our platform to many Universities like Harvard, Stanford, Yale, and of course, Johns Hopkins and have expanded to 100 countries around the world.
Besides that, the systems are used for technology developments by major industrial players like BMW, Airbus, Meta, Apple, Amazon, and many more. About 10 years ago we started with the development of medical products that we sell to hospitals and rehabilitation clinics.
We established a franchise system that allows businessmen and therapists to use neurotechnology in their centers to treat patients. With our recoveriX system for the neurorehabilitation of stroke patients and patients with Multiple Sclerosis, we are already in many countries, and up to about 50,000 treatments were done.â
They have since also teamed up with Tobii to offer integrated EEG and eye-tracking technology.
Most of you will be familiar with the term Brain-Computer-Interface (BCI) - sometimes also called Brain-Machine-Interface (BMI) or Mind-Machine-Interface (MMI) - but may not be fully aware what it actually means.
In a December 2015 publication, Christoph Guger (thatâs, by the way, where the G in g.tec comes from - it stands for Guger Technologies) and two of his co-authors described a BCI as follows:
âA BCI is a device that reads voluntary changes in brain activity, then translates these signals into a message or command in real-time (âŚ) Most BCIs rely on the electroencephalogram (EEG). These signals (also called âbrainwavesâ) can be detected with electrodes on the surface of the head. Thus, these ânoninvasiveâ sensors can detect brain activity with very little preparation. Some BCIs are âinvasiveâ, meaning that they require neurosurgery to implant sensors. These BCIs can provide a much more detailed picture of brain activity, which can facilitate prosthetic applications or surgery for epilepsy and tumor removal.â
The implants used in clinical trials by Neuralink (founded in 2016 by Elon Musk and a team of eight scientists and engineers) are the most well-known examples of invasive BCIs. And while we BRN shareholders tend to roll our eyes when our companyâs silicon gets confused with Muskâs âbrain chipsâ, there is no doubt that BrainChipâs technology is also being evaluated in this field of BCIs.
In 2020, g.tec medical engineering introduced the BCI & Neurotechnology Spring School, a free ten-day virtual event - now held annually - which has become the worldâs largest neurotech event, orchestrated from a small town in Austria called Schiedlberg. Participants can access 140 hours of cutting-edge education and even earn 14 ECTS* credits and an official exam certificate at no cost.
*ECTS = European Credit Transfer and Accumulation System
I noticed that one of last yearâs 82,000 (!) participants was Temi Mohandespour, who used to work as a research scientist at BrainChipâs now closed Perth office from March 2021 until January 2025. She has since moved to Berlin and now works for Data4life, a non-profit organisation, whose mission is to digitalise health data for research (www.data4life.care/en/).
https://www.linkedin.com/posts/temi-mohandespour_here-is-a-big-thank-you-to-gtec-medical-activity-7193097495894208513-9euk?
Several of her colleagues at BrainChip gave her above âthank youâ post a
, including our CTO.
While I wasnât able to find out anything concrete about what Temi Mohandespour may have been working on relating to BCIs during her last nine months at BrainChip post-Spring School, I happened to discover the LinkedIn profile of someone else who worked not only on one, but on two BCI projects utilising Akida - although not as an employee of BrainChip:
https://www.linkedin.com/in/hammouamri-ilyass/
Ilyass Hammouamri, who recently defended his PhD thesis at the UniversitĂŠ de Toulouse (https://doctorat.univ-toulouse.fr/as/ed/cv.pl?mat=140961&site=EDT)
and whose PhD supervisor was TimothĂŠe Masquelier (one of the four co-inventors of the JAST patent that BrainChip first licensed and later acquired), was a part-time research engineer at Neurobus between September 2024 and April 2025.
It was during that time - still under Gregor Lenz as CTO - that he âdeveloped a Proof of Concept solution for motor imagery classification from a Dry EEG Headset using a BrainChip Akida neuromorphic chip for robotic arm controlâ.
âMotor imagery (MI) is a mental process in which a subject vividly imagines performing a movement without any actual physical execution. MI is widely used in BCI systems to enable control of external devices, such as a cursor on a screen or a robotic arm, through brain activity.â
https://docs.medusabci.com/kernel/1.4/tutorials.php (by the Biomedical Engineering Group at the University of Valladolid, Spain)
I wonder whether this project may have been the continuation of the BMI* project that Neurobusâs first employee, Ljubica CimeĹĄa, had developed in collaboration with Airbus, which also used EEG signals for robotic control:
*The terms Brain-Computer-Interface (BCI) and Brain-Machine-Interface (BMI) are often used interchangeably.
https://www.linkedin.com/in/cimesa-ljubica/
But his part-time contract job with Neurobus was not the first time Ilyass Hammouamri had been involved in BCI research using Akida: During his time at CNRS (Centre national de la recherche scientifique) CerCo (Centre de Recherche Cerveau et Cognition) in Toulouse, where he was as a PhD candidate in TimothĂŠe Masquelierâs NeuroAI lab from to September 2021 to February 2025, he âworked on a joint project between different labs and BrainChip: Decoding speech from ECoG brain signalsâ.
Which means there must have been at least one more lab involved in that project, possibly more.
ECoG stands for electrocorticography. In contrast to EEG, it involves recording electrical activity directly from the surface of the brain und thus requires a craniotomy.
en.wikipedia.org
Here is a good illustration I found online, which happens to be from a video by g.tec medical engineering:
I have no idea whether or not any of g.tec medical engineeringâs products (such as wearable EEG headsets, biosignal amplifiers) were actually used for either of the two BCI projects that Ilyass Hammouamri was involved in.
What I can tell you, though, is that they list Airbus under âHappy Customersâ alongside quite a few other interesting names (https://www.gtec.at/).
Over the past 25+ years, g.tec medical engineering have specialised âin developing high performance brain-computer interfaces and neurotechnologies for both invasive and non-invasive recordings in research and clinical settingsâ (https://www.gtec.at/about/) and are one of the leading companies, if not THE leading company, in this field world-wide.
In a November 2024 interview (https://www.gtec.at/2024/11/04/leading-the-bci-field/), Co-Founder and Co-CEO Christoph Guger, who has degrees from both Johns Hopkins University and TU Graz, shared the following about g.tec medical engineeringâs impressive journey:
âWe sell our platform to many Universities like Harvard, Stanford, Yale, and of course, Johns Hopkins and have expanded to 100 countries around the world.
Besides that, the systems are used for technology developments by major industrial players like BMW, Airbus, Meta, Apple, Amazon, and many more. About 10 years ago we started with the development of medical products that we sell to hospitals and rehabilitation clinics.
We established a franchise system that allows businessmen and therapists to use neurotechnology in their centers to treat patients. With our recoveriX system for the neurorehabilitation of stroke patients and patients with Multiple Sclerosis, we are already in many countries, and up to about 50,000 treatments were done.â
They have since also teamed up with Tobii to offer integrated EEG and eye-tracking technology.
Most of you will be familiar with the term Brain-Computer-Interface (BCI) - sometimes also called Brain-Machine-Interface (BMI) or Mind-Machine-Interface (MMI) - but may not be fully aware what it actually means.
In a December 2015 publication, Christoph Guger (thatâs, by the way, where the G in g.tec comes from - it stands for Guger Technologies) and two of his co-authors described a BCI as follows:
âA BCI is a device that reads voluntary changes in brain activity, then translates these signals into a message or command in real-time (âŚ) Most BCIs rely on the electroencephalogram (EEG). These signals (also called âbrainwavesâ) can be detected with electrodes on the surface of the head. Thus, these ânoninvasiveâ sensors can detect brain activity with very little preparation. Some BCIs are âinvasiveâ, meaning that they require neurosurgery to implant sensors. These BCIs can provide a much more detailed picture of brain activity, which can facilitate prosthetic applications or surgery for epilepsy and tumor removal.â
The implants used in clinical trials by Neuralink (founded in 2016 by Elon Musk and a team of eight scientists and engineers) are the most well-known examples of invasive BCIs. And while we BRN shareholders tend to roll our eyes when our companyâs silicon gets confused with Muskâs âbrain chipsâ, there is no doubt that BrainChipâs technology is also being evaluated in this field of BCIs.
In 2020, g.tec medical engineering introduced the BCI & Neurotechnology Spring School, a free ten-day virtual event - now held annually - which has become the worldâs largest neurotech event, orchestrated from a small town in Austria called Schiedlberg. Participants can access 140 hours of cutting-edge education and even earn 14 ECTS* credits and an official exam certificate at no cost.
*ECTS = European Credit Transfer and Accumulation System
I noticed that one of last yearâs 82,000 (!) participants was Temi Mohandespour, who used to work as a research scientist at BrainChipâs now closed Perth office from March 2021 until January 2025. She has since moved to Berlin and now works for Data4life, a non-profit organisation, whose mission is to digitalise health data for research (www.data4life.care/en/).
https://www.linkedin.com/posts/temi-mohandespour_here-is-a-big-thank-you-to-gtec-medical-activity-7193097495894208513-9euk?
Several of her colleagues at BrainChip gave her above âthank youâ post a
, including our CTO.While I wasnât able to find out anything concrete about what Temi Mohandespour may have been working on relating to BCIs during her last nine months at BrainChip post-Spring School, I happened to discover the LinkedIn profile of someone else who worked not only on one, but on two BCI projects utilising Akida - although not as an employee of BrainChip:
https://www.linkedin.com/in/hammouamri-ilyass/
Ilyass Hammouamri, who recently defended his PhD thesis at the UniversitĂŠ de Toulouse (https://doctorat.univ-toulouse.fr/as/ed/cv.pl?mat=140961&site=EDT)
and whose PhD supervisor was TimothĂŠe Masquelier (one of the four co-inventors of the JAST patent that BrainChip first licensed and later acquired), was a part-time research engineer at Neurobus between September 2024 and April 2025.
It was during that time - still under Gregor Lenz as CTO - that he âdeveloped a Proof of Concept solution for motor imagery classification from a Dry EEG Headset using a BrainChip Akida neuromorphic chip for robotic arm controlâ.
âMotor imagery (MI) is a mental process in which a subject vividly imagines performing a movement without any actual physical execution. MI is widely used in BCI systems to enable control of external devices, such as a cursor on a screen or a robotic arm, through brain activity.â
https://docs.medusabci.com/kernel/1.4/tutorials.php (by the Biomedical Engineering Group at the University of Valladolid, Spain)
I wonder whether this project may have been the continuation of the BMI* project that Neurobusâs first employee, Ljubica CimeĹĄa, had developed in collaboration with Airbus, which also used EEG signals for robotic control:
*The terms Brain-Computer-Interface (BCI) and Brain-Machine-Interface (BMI) are often used interchangeably.
https://www.linkedin.com/in/cimesa-ljubica/
But his part-time contract job with Neurobus was not the first time Ilyass Hammouamri had been involved in BCI research using Akida: During his time at CNRS (Centre national de la recherche scientifique) CerCo (Centre de Recherche Cerveau et Cognition) in Toulouse, where he was as a PhD candidate in TimothĂŠe Masquelierâs NeuroAI lab from to September 2021 to February 2025, he âworked on a joint project between different labs and BrainChip: Decoding speech from ECoG brain signalsâ.
Which means there must have been at least one more lab involved in that project, possibly more.
ECoG stands for electrocorticography. In contrast to EEG, it involves recording electrical activity directly from the surface of the brain und thus requires a craniotomy.
Electrocorticography - Wikipedia
Here is a good illustration I found online, which happens to be from a video by g.tec medical engineering:
I have no idea whether or not any of g.tec medical engineeringâs products (such as wearable EEG headsets, biosignal amplifiers) were actually used for either of the two BCI projects that Ilyass Hammouamri was involved in.
What I can tell you, though, is that they list Airbus under âHappy Customersâ alongside quite a few other interesting names (https://www.gtec.at/).
stockduck
Regular
BrainChip Unveils Breakthrough AKD1500 Edge AI Co-Processor at Embedded World North America - BrainChip Investor Portal
BrainChip Unveils Breakthrough AKD1500 Edge AI Co-Processor at Embedded World North America Laguna Hills, Calif. â November 4th, 2025 â BrainChip Holdings Ltd (ASX: BRN, OTCQX: BRCHF, ADR: BCHPY), a global leader in ultra-low power, fully digital, event-based neuromorphic AI, today announced...
investor.brainchip.com
"...800 giga operations per second (GOPS) while operating under 300 milliwattsâsetting a new benchmark for edge AI efficiency."
That sounds incredible massiv power to me.....
Tothemoon24
Top 20
BrainChip Unveils Breakthrough AKD1500 Edge AI Co-Processor at Embedded World North America
ShareLAGUNA HILLS, Calif.--(BUSINESS WIRE)--BrainChip Holdings Ltd (ASX: BRN, OTCQX: BRCHF, ADR: BCHPY), a global leader in ultra-low power, fully digital, event-based neuromorphic AI, today announced the launch of its AKD1500, a neuromorphic Edge AI accelerator co-processor chip, at Embedded World North America.
Designed to deliver exceptional performance with minimal power consumption, the AKD1500 achieves 800 giga operations per second (GOPS) while operating under 300 milliwattsâsetting a new benchmark for edge AI efficiency. This makes AKD1500 ideal for deployment in battery powered wearables, smart sensors, and heat-constrained environments where battery life and thermal limits are critical.
The AKD1500 integrates seamlessly with x86, ARM, and RISC-V host processing platforms via PCIe or Serial interfaces, enabling rapid adoption across a wide range of applications. The AKD1500 co-processor approach is ideal for a wide range of environments and industries, upgrading multi-processor SoCs within defense, industrial and enterprise settings, and upgrading embedded microcontrollers for AI solutions in healthcare, wearables, and consumer electronics without a complete system redesign. The AKD1500 product has been delivered and designed into several end solutions in AI enabled sensing for medical and defense related applications, including Parsons, Bascom Hunter and Onsor Technologies.
âThe AKD1500 is a catalyst for the next wave of intelligent AIoT devices,â said Sean Hehir, CEO of BrainChip. âWeâre empowering developers to break free from cloud dependency and bring adaptive learning directly to the edge in a compact, cost-effective package. This technology will make AI truly ubiquitous in smart factories, homes, and wearable devices.â
âBrainChipâs AKD1500 on our 22FDXÂŽ platform delivers outstanding compute and memory efficiency,â said Anand Rangarajan, Director of AI & IOT Compute at GlobalFoundries. âEmbedded developers are constantly innovating to get the right level of AI to fit within performance, power and area constraints. Using BrainChipâs neuromorphic architecture combined with GlobalFoundriesâ 22FDXÂŽ process technology, the AKD1500 offers an excellent performance, power and cost envelope that fits into edge devices. Weâre proud to support BrainChipâs end-to-end embedded AI solutions using GlobalFoundries silicon.â
AKD1500 is supported by BrainChipâs MetaTF⢠software development tools environment, enabling machine learning engineers to easily convert, quantize, compile and deploy models on Akida using standard TensorFlow/KERAS formats which dramatically reduces development time and cost while expanding accessibility for AI developers. BrainChipâs event-based Akida⢠neuromorphic architecture also enables the AKD1500 to provide on-chip learning, a critical differentiator from conventional AI accelerators that rely solely on cloud-based training.
AKD1500 samples are available today with volume production scheduled for Q3â26.
BrainChipâs Chief Development Officer, Jonathan Tapson, will present âThe Impact of GenAI Workloads on Compute-in-Memory Architecturesâ at Embedded World North America on November 4th.
For more information: join us for a demo of AKD1500 at Booth 3080; visit the BrainChip developer site for free tutorials, tools, and models for neuromorphic computing; and check out the Embedded World North America site.
smoothsailing18
Regular
It now has been 1 year since brainchip launched their Pico.
I would think we should have some news soon as that was at the request of a client i believe
.
Some news on this would be good.
I would think we should have some news soon as that was at the request of a client i believe
.Some news on this would be good.
GStocks123
Regular
perceptron
Regular
That is impressive performance. No wonder the potential and current applications are varied.BrainChip Unveils Breakthrough AKD1500 Edge AI Co-Processor at Embedded World North America - BrainChip Investor Portal
BrainChip Unveils Breakthrough AKD1500 Edge AI Co-Processor at Embedded World North America Laguna Hills, Calif. â November 4th, 2025 â BrainChip Holdings Ltd (ASX: BRN, OTCQX: BRCHF, ADR: BCHPY), a global leader in ultra-low power, fully digital, event-based neuromorphic AI, today announced...
"...800 giga operations per second (GOPS) while operating under 300 milliwattsâsetting a new benchmark for edge AI efficiency."
That sounds incredible massiv power to me.....
View attachment 92755
Similar threads
