Tothemoon24
Top 20
Again and again and again , where is the link , l can not think , is it possible there is no link ? Let me thinkAGAIN and again and again ... Where is the link?

Sorry boss I’ll endeavour to improve on my missing links

Again and again and again , where is the link , l can not think , is it possible there is no link ? Let me thinkAGAIN and again and again ... Where is the link?
It may be that FG are working on another function that requires Gen 2 and TENNs hence the provision for another licence.Global Foundries produced a batch of Akida 1500 22nm SoI chips. The 1500 is the Akida IP without the non-BRN ARM Cortex processor.
Maybe therse were tested with ESA/FG (Frontgrade) with a FG MCU.
So now it appears FG have taken a licence for Akida IP incorporated with an FG MCU in a So, and they like it so much they have an option to produce a second product line .
This is the 6 May announcement:
https://feeds.issuerdirect.com/news-release.html?newsid=4537416869883669 20240506
BrainChip and Frontgrade Gaisler to Augment Space-Grade Microprocessors with AI Capabilities
BrainChip Holdings Ltd (ASX: BRN, OTCQX: BRCHF, ADR: BCHPY), the world’s first commercial producer of ultra-low power, fully digital, event-based, neuromorphic AI IP, and Frontgrade Gaisler, a leading provider of space-grade system-on-chip solutions, announce their collaboration to explore the integration of BrainChip’s Akida TM neuromorphic processor into Frontgrade Gaisler’s next generation fault-tolerant, radiation-hardened microprocessors. This collaboration represents a significant milestone as it aims to introduce the first space-grade SoC worldwide with incorporated true artificial intelligence (AI) capabilities.
These next generation microprocessors would include BrainChip’s AI processing capabilities, thereby enabling a considerable step forward in the computing resources available for space-borne systems. In particular, the inclusion of the Akida neuromorphic processor aims to deliver compelling power efficiency and inference performance while maintaining compatibility with existing Convolutional Neural Networks (CNNs).
“Adding AI capabilities to our next generation microprocessors would set a new standard for modern space-grade computing devices,” said Sandi Habinc, General Manager at Frontgrade Gaisler. “The success of this joint effort could empower organizations to leverage AI technology to enhance mission efficiency, ultimately advancing the frontiers of what is possible in space.”
Now the other odd thing is that the licences are for Akida 1 IP, ie, not Akida 2 ... not even as an option ...
Clearly Akida 2 (with TENNs) comes at a significant premium. There is also the fact that Akida 2 has not been produced in silicon. If you're going into space, you'd probably want to see it working beforehand (see Hubble trouble).
So there's still room for FG to step up to Akida 2 at a later date.
What it does show is that Akida 1 performs at or above specification for FG's (and ESA's) purposes.
![]()
Frontgrade Gaisler Partners with BrainChip to Bring Akida AI Chips to Space - The Manufacture Data
BrainChip, the world’s first commercial producer of ultra-low power, fully digital, event-based neuromorphic AI, has reached an exciting milestone in itsthemanufacturedata.com
Frontgrade Gaisler Partners With BrainChip To Bring Akida AI Chips To Space
HOMEPRESS RELEASESFRONTGRADE GAISLER PARTNERS WITH BRAINCHIP TO BRING AKIDA AI CHIPS TO SPACE
- THE MANUFACTURE DATA
- DECEMBER 16, 2024
- PRESS RELEASES
![]()
BrainChip, the world’s first commercial producer of ultra-low power, fully digital, event-based neuromorphic AI, has reached an exciting milestone in its mission to revolutionize edge computing. The company recently announced a commercial licensing agreement with Frontgrade Gaisler, a leading provider of radiation-hardened microprocessors specifically designed for space applications. Under this agreement, Frontgrade Gaisler will integrate BrainChip’s Akida™ IP into their next-generation, fault-tolerant system-on-chip (SoC) solutions, enabling advanced hardware AI acceleration for space missions.
This groundbreaking collaboration signifies a major leap forward in the deployment of artificial intelligence in space environments, as it paves the way for AI chips to operate reliably in the harsh conditions of outer space. The Akida neuromorphic computing platform was selected for its unique ability to process real-time data streams with resilience, autonomy, and precision, all while maintaining minimal power consumption and occupying a small physical footprint.
Driving Neuromorphic AI to the Furthest Edge
The European Space Agency (ESA) has been at the forefront of efforts to bring neuromorphic computing to space, highlighting its transformative potential in applications such as computer vision, navigation, and fault detection. Laurent Hili, Microelectronics and Data Handling Engineer at ESA, explained why neuromorphic computing is poised to revolutionize space technology.
While many fields undoubtedly benefit from neuromorphic computing, one critical application area is computer vision, said Hili. “Current FPGA or GPU technologies often fall short when deployed at the edge in space missions due to their significant power, mass, and volume constraints. Neuromorphic technology, as demonstrated through several ongoing projects with satellite manufacturers and IP/component suppliers like Frontgrade Gaisler and BrainChip, addresses these challenges head-on.”
This sentiment underscores the importance of the Akida platform’s cutting-edge capabilities. Neuromorphic computing mimics the human brain by processing information through event-based, spike-driven architectures, which are inherently more power-efficient than conventional AI systems. Such efficiency is particularly valuable in space missions, where every watt of power, gram of mass, and cubic centimeter of volume is at a premium.
Enhancing Space-Borne Computers
BrainChip’s Akida neuromorphic technology will significantly enhance the performance of space-based, on-board computers. Its ability to deliver superior power efficiency and inference performance while maintaining compatibility with existing Convolutional Neural Networks (CNNs) makes it a game-changer for space applications. By leveraging the Akida IP, Frontgrade Gaisler aims to deliver AI-enabled space processors that can handle complex tasks such as anomaly detection, image recognition, and data analysis in real time without the need for continuous communication with Earth-based systems.
For the last 20 years, Geisler’s space-grade microprocessors have been successfully deployed to every planet in our solar system through missions led by ESA and other major space agencies, said Sandi Harbin, General Manager at Front grade Geisler, a Front grade company. Our evaluation of Brain Chip’s Aida IP revealed that it perfectly complements our roadmap for future space processors. Licensing this technology is a natural next step in advancing our systems with cutting-edge neuromorphic AI capabilities.
The collaboration between BrainChip and Frontgrade Gaisler demonstrates the growing demand for space-optimized AI solutions. Neuromorphic technology’s ability to process data in a more human-like and efficient manner aligns perfectly with the stringent requirements of space exploration, where autonomy and reliability are non-negotiable.
Realizing a Vision for Space AI
Sean Hehir, CEO of BrainChip, emphasized the significance of this partnership in advancing the boundaries of AI in space. “This collaboration with Frontgrade Gaisler to license Akida IP for space-based SoCs is a critical milestone in addressing the market demand for AI in space deployments,” said Hehir. “What was once thought unattainable—bringing advanced AI to space—is now becoming a reality, and we are proud to play a pivotal role in this transformation. Our trusted relationship with Frontgrade Gaisler enables us to push the limits of what’s possible in space computing.”
The integration of BrainChip’s Akida IP into Frontgrade Gaisler’s SoCs represents a union of two innovative approaches: BrainChip’s focus on event-based neuromorphic AI and Frontgrade Gaisler’s expertise in developing radiation-hardened, fault-tolerant microprocessors. Together, the companies are set to overcome some of the most challenging obstacles in space technology, such as the harsh radiation environment and the need for ultra-efficient computing solutions.
As space missions become increasingly complex and ambitious, the demand for autonomous, intelligent systems capable of making split-second decisions without relying on Earth-based control is growing rapidly. The Akida platform’s ability to support on-board AI inference while conserving resources makes it a vital component of the next generation of space exploration technologies.
This partnership highlights the importance of collaboration in advancing the frontiers of technology. By combining their strengths, BrainChip and Frontgrade Gaisler are enabling AI to extend its reach far beyond Earth, opening up new possibilities for innovation and discovery in the vast expanse of space.
Yeah ... but ...It may be that FG are working on another function that requires Gen 2 and TENNs hence the provision for another licence.
As space craft evolve there will be many different functions that require AKIDA. The size of the royalty indicates the importance of AKIDA to this program.
... well I s'pose Peter's got a bit more time on his hands now ... how does he go on a ladder?What I liked about the FG announcement was the word augment.
View attachment 74291
View attachment 74290
I feel we can act like a tree than a climber. Brainchip have a merit of a tree, then why to remain like a parasite.Too many what ifs. We don’t want to have to depend on another company to maybe start generating large revenue. We want an established company that already generates high revenue. The AI industry is currently booming and making more money than ever so this small amount with such a minor company doesn’t get me that excited. Hopefully something with a bigger name arrives soon
Contract!!! Well done Sean. Now can you lock in about 20 more please!!
Sorry, what do you mean? No offence intended Galaxy, just tongue in cheek.I have stayed away off this site due to its bias, but anyone here that seriously think there will be an announcement prior to Christmas on the iword of who is running this company Is delirious. Why would anyone sell shares in their company if they knew in the next 4 weeks that a deal would fall. Hehir has constantly strung shareholders along and continues to do it.You all had your chance to vote him out last AGM but no like lemmings you followed his words, Remember them IMMINENT…… what have we got FKN not a word since. Lemmings that what I called you bunch.IWork for one of the largest worldwide companies and today I had to resort to send a email to one of our General Managers as they were looking at integrating AI into our company. I forwarded him Brainchips website link In the hope of achieving something. It only takes 200 signatures from shareholders to instigate a extraordinary AGM. If Christmas hits and nothing announced.I suggest as a shareholder you must think TIME TO GO HEHIR. You missed your chance last AGM don’t make the same mistake. Don’t wait till there is no money left in the bank account get rid of them.
Yes, hasn't aged terribly well has it?Sorry, what do you mean? No offence intended Galaxy, just tongue in cheek.
This deal is a real positive as is the recent contract simply because they validated the product and represent another step in the BRN journey to aspirational leadership in the Edge industry over the next decade.I stand by my statement, if the happy clappers think that a $150k up deal is going to change shareholders angst we’ll suggest you reconsider. How many shareholders have lost 80% plus of their shareholder value, A majority is the answer, cause this bullshit NDA path management have gone down. If Brainchip would have come out and quantified that the 10% net sales royalty Brn estimate would bring in estimated 20-30 million year ending 2026 I would doubt myself. But no they leave us guessing tell us nothing, All this absolute crap about shorters getting burnt today,is that crap they are laughing all the way to the bank they are still in control, Shares up 1 CENT, Why because the shorters know if they get in the shit management will sell them 50 millions shares just like in the last capital raise. Yes the sofisticated investers were the shorters. Management sold us all out to the shorters to stay afloat. So no the next investor votes won’t be any different than the last, votes will go against them and in a substantial way. The shorters aim is to get rid of management, choke the supply of money, keep the share price low at any cost, to upset shareholders. I cant for the life of me can’t see anything changing, WOW 150K deal, ALL THE HAPPY CLAPPERS HERE have stated wait for the first IP deal wait for the first IP Deal well we waited and what we were still all refreshing over share accounts all day 100 times 1 FUCKEN CENT. Pantene that
Hello fellow brners,
Happy silly season soon,
Been waiting like everyone else but things are slowly being uncovered and it was very nice to see this announcement this morning re Frontgrade Gaisler inc a nice mention in the Herald Sun.
BrainChip Holdings (ASX:BRN) has announced that Frontgrade Gaisler, a provider of radiation-hardened microprocessors for space, has licensed its Akida 1.0 Neuromorphic AI IP for integration into space-grade, fault-tolerant chips. BrainChip will receive royalties from Frontgrade.
I also happy to come across an interesting article on Risc-V with Brainchip mentioned. (this probably has surfaced here on tse before) but given the injection of funds by Jeff Bezos to Tenstorrent in early December, I thought it might be worth a revisit of it especially the part where it says Synergies between Risc-v, Tenstorrent and Brainchip.
The combination of Tenstorrent’s GPU capabilities and BrainChip’s neuromorphic approach demonstrates the broad applicability of RISC-V in both conventional AI workloads and next-generation computational neuroscience
It's good to be a shareholder.
Hope Bravo gets new running shoes for xmas.
I suppose one could conclude that MB were still working on neuromorphic with Akida through pretty much all of 2024
Sreelakshmi Rameshan
Graduate student at University of Stuttgart| Computer Science
Mercedes-Benz AG University of Stuttgart
Stuttgart, Baden-Württemberg, Germany
Experience
![]()
Master Thesis
Mercedes-Benz AG
Feb 2024 - Sep 2024 8 months
Stuttgart, Baden-Württemberg, Germany
For my master's thesis at Mercedes-Benz AG, I conducted an in-depth analysis of frameworks and strategies for converting neural networks (NN) into spiking neural networks (SNN) for neuromorphic processors. The project involved evaluating SDKs for neuromorphic hardware like BrainChip Akida, Intel Loihi, and SynSense to assess their compatibility, functionality, and performance. Using frameworks such as MetaTF, Lava, and Sinabs, I translated the PilotNet model into SNNs and analyzed their practical feasibility, documenting challenges and conducting performance evaluations based on metrics like MSE and accuracy. This work also explored the use of neuromorphic processors as accelerators for neural networks, identifying hurdles and proposing solutions to optimize translation processes.
And then there is also a recent Master’s thesis sort of connected to MB’s neuromorphic research (more on that later, as posts have a limit of 10 upload images) that strengthens my belief that MB are still weighing their options…
Lots of points raised that cannot simply be glossed over and that suggest to me Mercedes-Benz is nowhere near to implementing neuromorphic technology at scale into serial cars.
Interested to hear your or anyone else’s thoughts on those points.
Interestingly, she says Akida had limitations with complex temporal tasks....was she using Akida 1 or TENNS is the question for mine.Hi Fullmoonfever,
that’s the same Master’s thesis I had alluded to in a previous post of mine I had not yet followed up on:
However, it was already completed in August 2024, so I did not come to the same conclusion as you have…
Here are some excerpts:
View attachment 74307
View attachment 74308
View attachment 74311
View attachment 74312
View attachment 74313
Interestingly, she says Akida had limitations with complex temporal tasks....was she using Akida 1 or TENNS is the question for mine.
Would like to suspect it was V1 and that TENNS would reduce some of the said limitations.