BRN Discussion Ongoing

[itsol4605]

A competitor's product with analog neurons.
Perhaps someone could make a comparison to Akida.

#analog | Neuromorphicism

Neuronova is ready to launch their first Evaluation Kit Silk 1 in the following months! Their H1 Neuromorphic Processor has programmable #analog neurons and synapses with a configurable analog front-end. Give them a follow to stay updated!

www.linkedin.com

 

[Pom down under]

A competitor's product with analog neurons.
Perhaps someone could make a comparison to Akida.

#analog | Neuromorphicism

Neuronova is ready to launch their first Evaluation Kit Silk 1 in the following months! Their H1 Neuromorphic Processor has programmable #analog neurons and synapses with a configurable analog front-end. Give them a follow to stay updated!

www.linkedin.com

 

[genyl]

They hire a whole bunch of sales managers around the world but wth are they even doing? Not selling anything is my guess. Jesus christ zzzzzzzzz

 

[Esq.111]

Morning Chippers ,

Little article on Anduril ,

‘I have got this s- figured out’: Anduril unveiling EagleEye mixed-reality device at AUSA - Breaking Defense https://share.google/4v1pjqxX0rKgmD2Qr

Regards,
Esq.

 

[Pom down under]

 

[Pom down under]

They hire a whole bunch of sales managers around the world but wth are they even doing? Not selling anything is my guess. Jesus christ zzzzzzzzz

Trying to close ya shorts

 

[7für7]

AOpening 23.5 just to go 22.3 while other stocks yesterday recovered

 

[Guzzi62]

Morning Chippers ,

Little article on Anduril ,

‘I have got this s- figured out’: Anduril unveiling EagleEye mixed-reality device at AUSA - Breaking Defense https://share.google/4v1pjqxX0rKgmD2Qr

Regards,
Esq.

Can someone please update me what this has to do with BRN?

Is it the display being used in the helmet, perhaps?

ChatGPT can't find any BRN connection with either his company nor him.

 

[Esq.111]

Afternoon Guzzi62 ,

Hard to say at this point in time , though unlike most , Lucky seems to have his shite together ( ability running Anduril ) , rapidly deploying new fangled technology.

Also our Chief Technical boffin was wandering around Washington the other day , chewed the fat with a few big wigs , then posted on LinkedIn that ....... words to the equivalent...... BrainChip , Anduril & Plantier are the goods.

Long Bow yes , though uncanny non the less.



Regards,
Esq.

 

[HopalongPetrovski]

Afternoon Guzzi62 ,

Hard to say at this point in time , though unlike most , Lucky seems to have his shite together ( ability running Anduril ) , rapidly deploying new fangled technology.

Also our Chief Technical boffin was wandering around Washington the other day , chewed the fat with a few big wigs , then posted on LinkedIn that ....... words to the equivalent...... BrainChip , Anduril & Plantier are the goods.

Long Bow yes , though uncanny non the less.



Regards,
Esq.

One would have to consider that such an application would benefit greatly from Akida's abilities and characteristics in regard sensor fusion, on chip learning, both the privacy and immediacy of cloud independence and power efficiency gains. And Anduril likes to work with COTS componentry.
If the IP can't be baked into existing hardware I'm sure BrainChip would be happy to facilitate a production run for an appropriate fee.

 

[Pom down under]

Can someone please update me what this has to do with BRN?

Is it the display being used in the helmet, perhaps?

ChatGPT can't find any BRN connection with either his company nor him.

 

[Fullmoonfever]

From memory, this ESA project was discussed late last year and a few times this year and believe there could be an aim for a rocket launch in 2026 as Frangipani had pointed out previously.

I thought originally might be this year but not to be.

Anyway, I recall Frangipani also pointing out it was odd some time ago that EdgX weren't prominent on our website and that maybe Akida had been dropped off their DPU offering as the add on neuromorphic which does seem correct when looking at the current specs.

Though a LinkedIn post by Satcomsearch just 4 mths ago was still touting it as an add on.

Happy to welcome EDGX as a satsearch trusted supplier. Introducing the EDGX DPU, a compact, powerful and modular AI processing system purpose-built for smallsats — and hardened for space. 🔹… | Narayan Prasad

Happy to welcome EDGX as a satsearch trusted supplier. Introducing the EDGX DPU, a compact, powerful and modular AI processing system purpose-built for smallsats — and hardened for space. 🔹 Powered by NVIDIA Jetson Orin NX, this DPU delivers 157 TOPS of AI compute in a <1 kg, 0.5U form...

www.linkedin.com

If our relationship is not there then curious as to why, given EdgX and BRN had the MOU signing exercise.

It may just be that the released DPU is a commercial version currently available and not with neuromorphic as yet till testing phases complete and acceptable like with the below project?

Anyway, this project has been running since last year but I've just noticed an update recently and guess it will come out soon which neuromorphic they went with.

EdgX are the lead contractor and below is a snip of the original objectives and then a snip of the latest updates.

Neuromorphic AI Onboard – ESA CSC: Connectivity & Secure Communications

connectivity.esa.int

Neuromorphic AI OnboardNeuromorphic AI Onboard​

• Competitiveness & Growth

• Status
  Ongoing
• Status date
  2024-10-08

Objectives
The proposed activity entails the definition (definition phase) of an onboard neuromorphic AI data processing unit designed for satellite communication (satcom) Low Earth Orbit (LEO) constellations, targeting smallsats. The activity contains three main high level chapters:

1. Identification of potential neuromorphic satcom applications and scenarios.
2. Preliminary benchmarking of neuromorphic algorithms.
3. Define the product architecture, user requirements and technical specifications.

Current status
The project is approaching the Phase Completion Review. Achievements include the analysis and selection of use cases, a revised set of product requirements, and the initial DPU architecture. Trade-off and selection of neuromorphic processors and algorithms has been conducted, with the overall hardware and software architecture now proposed. The business plan has been consolidated. Preliminary benchmarking of algorithms with demonstrations of selected use cases have been designed.

Documentation​

Documentation may be requested

Prime Contractor​

EDGX​

Belgium
https://www.edgx.space/

Last update
2025-10-13 02:06

 

[MDhere]

Bring on the hour of power to $1.50 !

 

[7für7]

Bring on the hour of power to $1.50 !

I personally hope that a shorter/manipulator will not bring the 20 cent back shortly before closing

 

[IloveLamp]

Bring on the hour of power to $1.50 !

 

[Guzzi62]

View attachment 91954

I had expected nothing less from our resident clown, well done sir!

 

[7für7]

Oh yeah… here we go …

 

[manny100]

Politicians come and go but its only Tech advances that really cause change or allow change. That is why we are in the 'box' seat.
It's just patience from here.
........... and that is why i always say that our tech is our value safety net.

 

[Bravo]

A competitor's product with analog neurons.
Perhaps someone could make a comparison to Akida.

#analog | Neuromorphicism

Neuronova is ready to launch their first Evaluation Kit Silk 1 in the following months! Their H1 Neuromorphic Processor has programmable #analog neurons and synapses with a configurable analog front-end. Give them a follow to stay updated!

www.linkedin.com

ChatGPT's Response


Here’s a concise, apples-to-apples comparison of Neuronova’s H1 vs BrainChip’s Akida (Gen-2) based on what’s publicly available.

What each one is (at a glance)​

• Neuronova H1 — a fully-analog neuromorphic processor for spiking neural networks (SNNs) aimed at ultra-low-power edge sensing; the company positions it as “doing more with less,” with a roadmap toward sensor-in-package integration. Public materials emphasise analog in-memory/time-domain compute and extremely low power (trade press cites <1 µW for always-on sensing use cases). NeuromorphicCore.ai+3Neuronova+3Tech.eu+3
• BrainChip Akida (Gen-2) — a fully-digital neuromorphic processor IP (licensable block or FPGA dev platform) designed for event-based inference and on-chip/edge learning; supports CNN→SNN conversion and BrainChip’s TENN (Temporal Event-based Neural Networks) models, with model cards and documented HW limits. BrainChip+3BrainChip+3doc.brainchipinc.com+3

Side-by-side​

Dimension	Neuronova H1	BrainChip Akida Gen-2
Core compute style	Fully analog SNN compute at the sensor/edge	Fully digital event-domain neural processor IP
Target use	Ultra-low-power always-on sensing; roadmap to sensor-level integration	Edge AI inference and on-device learning (incremental/edge learning) for wearables, IoT, automotive, space, etc.
Power narrative	“Do more with less”; reporting around sub-µW always-on claims in media profiles (for tiny tasks)	“Millwatts-class” continuous AI; event-sparsity to cut energy; supports edge learning without cloud round-trips Tech.eu+2PoliHub+2
Sensors	Messaging about “towards full sensor integration” (front-ends that adapt to drift/non-ideality)	Works with conventional frame sensors and event cameras; model zoo & conversion tools; not sensor-hardwired Neuronova+1
Software/tooling	Public details limited (no open model zoo/docs on H1 at time of writing)	Published product brief, capability limits, examples docs, and Akida 2 model cards (incl. 8-bit / 4-bit quantisation) BrainChip+3BrainChip+3doc.brainchipinc.com+3
Learning	(Not detailed publicly)	Supports on-chip/edge learning; TENNs for temporal efficiency; incremental learning flows BrainChip+1
Business model	Early-stage startup; pre-seed funded, building chips	IP/licensing + dev platforms; 2nd-gen IP available (docs, model zoo) PoliHub+1

​

BrainChip – partners & collaborators​

• Arm – official Arm Partner Catalog + Arm AI Partner Program (incl. Keil device packs/tools). Arm+2investor.brainchip.com+2
• Prophesee, Socionext, VVDN – joint demos/solutions at CES and other events. investor.brainchip.com
• RTX / Raytheon – teaming on an AFRL neuromorphic radar contract. BrainChip
• Frontgrade Gaisler – licensed Akida IP for space-grade SoC. Edge AI and Vision Alliance
• ARQUIMEA – water safety/search-and-rescue detection solution. BrainChip
• DigiKey – channel/distribution for Akida boards & dev kits. investor.brainchip.com
• Eastronics (Israel), SalesLink (Europe) – regional sales partners. BrainChip

• Intel (IFS) — Member of the Intel Foundry Services (IFS) Accelerator / IP Alliance. BrainChip+2EENews Europe+2
• SiFive — Joint edge-AI platform (Akida + RISC-V). BrainChip+1
• Edge Impulse — Platform integration / ecosystem partner. BrainChip+2edgeimpulse.com+2
• emotion3D — In-cabin automotive safety UX with Akida. BrainChip+2Electronicsclap+2
• Frontgrade Gaisler — Licenses Akida IP for space-grade SoCs. BrainChip+1
• Bascom Hunter — Defense/Intel integrations noted in ASX materials. ASX Announcements
• Onsor Technologies — Wearable seizure-prediction glasses on Akida. BrainChip+2FT Markets+2
• Chelpis / Mirle Group — Security & robotics collaboration (Akida AKD1000/M.2). BrainChip+2TechBullion+2
• Quantum Ventura — Akida used in CNRT cyber-threat tool (project partners include Lockheed Martin MFC & Penn State). BrainChip+1
• Valeo — Joint development agreement for automotive AD/ADAS use. Small Caps

Neuronova – partners & collaborators​

• On Neuronova’s site and recent profiles/funding notes, no Arm (or similar large-ecosystem) partnership is listed publicly. Pages focus on the H1 analog neuromorphic chip and funding news; no partner roster is shown.

Practical takeaways​

• If you need proofable tools, docs, and near-term deployability: Akida Gen-2 is more mature and documented today (model zoo, conversion pipelines, HW limits), with a defined IP path for partners. BrainChip+2BrainChip+2
• If your top priority is extreme standby/always-on power at the sensor itself: Neuronova’s H1 vision (analog SNN + sensor-in-package) is compelling for microwatt-class sentry tasks, but public specs are thin; you’ll likely need NDAs/eval hardware to validate claims and fit. Neuronova+2Tech.eu+2
• Architectural trade-off: analog neuromorphics can be exceptionally efficient but may face calibration/variability and tooling challenges; digital neuromorphics (Akida) are somewhat higher power but typically easier to integrate, reproduce, and scale across foundries and toolchains. (This analog-vs-digital tooling friction is a theme in SNN literature.) MDPI

Where the info is thin (so you know what to ask vendors)​

• Neuronova H1: exact MAC/synapse counts, memory topology, input modalities, supported learning rules, compiler/SDK maturity, I/O (SPI/PCIe), latency/accuracy vs standard benchmarks — not detailed on the public site. Neuronova
• Akida Gen-2: plenty of docs exist, but real-world PPA depends on the mapped model and operating point; check the model cards and HW-limit page for your network shapes and quantisation plans. BrainChip+1

---

Bottom line​

• Akida Gen-2 is the safer pick if you want a documented, digital, licensable neuromorphic platform with existing tools and examples today. BrainChip+1
• Neuronova H1 looks aimed at ultra-low-power sensing with a bold analog approach and sensor-integration roadmap, but you’ll need direct vendor engagement to validate specs and tooling readiness beyond the website/press. Neuronova+1

 

[keyeat]