BRN Discussion Ongoing

[Guzzi62]

The logic and the mathematics in relation to the metrics provided and interpreted (by some here) makes no business sense. Do Hoppy's interpretation of the metrics after researching variable costs to manufacture a silicon chip and research if there are other costs associated with manufacturing silicon chips that a foundry, not the IP/Chip designer has to factor in when pricing their production costs. It's well worth the effort.

Bla Bla you are clogging up the tread with meaningless posts no one cares about.

Off to ignore you go, bye bye.

 

[perceptron]

Where on page 14 is there wording that suggests your statement to be true?


View attachment 95375

It states it clearly in the graphic.

Price/chip (volume dependent)
The Price per chip is volume dependent.

The words are Price/chip (volume dependent). There is no explicit reference to sales to customer. Are you implying that a foundry is only charging the variable costs to produce a chip? That makes no sense.

 

[HopalongPetrovski]

It’s like anything that you buy
Volume gives you room to move
In any manufacturing the more you purchase the better deal you can get
It’s a win win
Also the better it is for everyone from the production level down to the buyer at the shop.
I don’t see the point of the discussion.

Hi Food.
I was just trying to ascertain and understand what P's issue's were, and if they had any validity.
Apparently not.
Or none that I can decipher anyway.
It seems that the majority opinion here is that Perceptron should just present his/her findings rather than asking more cryptic questions.
That is my opinion anyway. Without further clarification I'm not prepared to give this matter any further attention.

I doubt there will ever be general public knowledge of either the "real" price BrainChip pays per unit for its product nor the "real" price it gets when it sells it.
Of course there will be reported figures for those interested to dissect and as time passes and numerous deals are made and paid for, info. will gradually leak out.
But initially, (and I would counsel them for as long as they can maintain it), both of these figures are commercially sensitive in a competitive market and will be subject to the most rigorous NDA's their lawyers can construct.
I suspect they will be obfuscated by both BrainChip, their source, and whoever our product is sold to, ad infinitum, down the line.
Nobody willingly discloses info that can be used by others to firmly establish a profit margin.
It puts one in a weak bargaining/negotiating position especially when talking contracts, both external and within the company.
And, at the root of all the calculations is net profit.
And, at a certain point, commercial company's (like BrainChip) need to get off the investor's teat and start producing profit, or they go under.

 

[overpup]

This Page is just a research linking page with the hope of something,
What a disgrace its all crap this NDA'S, Whats a NDA when its not in a product , for 5 years if not more there been throwing these lines out
Companies can throw that line out all day its time shine and get information out there if companies are with us or not and akida is in any product,
Im sick of hearing garbage and the IBM Line is a classic, mercedes , toyota, and there is many more, why haant anything come thru

In this context I think NDA stands for "No Dollars Associated"

 

[7für7]

The logic and the mathematics in relation to the metrics provided and interpreted (by some here) makes no business sense. Do Hoppy's interpretation of the metrics after researching variable costs to manufacture a silicon chip and research if there are other costs associated with manufacturing silicon chips that a foundry, not the IP/Chip designer has to factor in when pricing their production costs. It's well worth the effort.

Shh shh shh shhhhhh
Everything will be okay

 

[perceptron]

Bla Bla you are clogging up the tread with meaningless posts no one cares about.

Off to ignore you go, bye bye.

I guess real discussion vs copyrighted gifs have priority.

 

[Diogenese]

#starlinkd2c #directtocell #reinforcementlearning #spacexakida #neuromorphicai #brazildigitalinclusion #satelliteedgecomputing #leoinnovation #spacex #akida #starlink #d2c | Daniel Azevedo Novais

✨ ⚙️ Starlink Direct‑to‑Cell: RL‑Driven Autonomous Beam Management for Remote Brazil The objective is to maximize population coverage and spectral efficiency in critical regions including the Legal Amazon, Pantanal wetlands, and the Northeastern backlands. As of February 2026, Starlink...

www.linkedin.com

View attachment 95378 View attachment 95379 View attachment 95380

Hi ILL,

"It may come cannonballing down from the sky" Q4 2026?!!!! -

if we're not all being hallucinated.

Until then, it is "classified"?

We have seen snippets of references to electronically steering antennae and SDR, so could be, who knows?

 

[HopalongPetrovski]

The words are Price/chip (volume dependent). There is no explicit reference to sales to customer. Are you implying that a foundry is only charging the variable costs to produce a chip? That makes no sense.

The only correlation between the price you pay for something and the price you on sell it for, is that the sales price should exceed the purchase price, plus all associated costs of being in business, plus a profit margin, which is the incentive.

If this is the point you are making I think that is already pretty well understood by most here.

 

[perceptron]

The only correlation between the price you pay for something and the price you on sell it for, is that the sales price should exceed the purchase price, plus all associated costs of being in business, plus a profit margin, which is the incentive.

If this is the point you are making I think that is already pretty well understood by most here.

So by your logic and other's, a foundry is only charging it's customer's to produce their IP/Chip designs the variable costs? Moreover, what is the logic behind the metric "volumes < $10" on the same page?

 

[buena suerte :-)]

에이직랜드, 글로벌 뉴로모픽 AI 선두주자 ‘브레인칩’과 신규 계약 - 더스탁(The Stock) | Maeve Jung

I’m so proud to share our new project with BrainChip, a global leader in neuromorphic AI powered by their innovative Akida™ technology. Facilitating the initial discussions and engaging directly with the Brainchip leadership team has been an inspiring experience for me. Asicland is now...

www.linkedin.com

View attachment 95351

https://www.the-stock.kr/news/articleView.html?

에이직랜드, 글로벌 뉴로모픽 AI 선두주자 ‘브레인칩’과 신규 계약​

김태영 기자다른 기사 보기

• 입력 2026.02.23 13:05

기사스크랩하기다른 공유 찾기본문 글씨 키우기본문 글씨 줄이기

브레인칩 핵심 설계 파트너 낙점…해외 고객사 유치 및 수익 구조 다변화 가속​

에이직랜드X브레인칩. 사진=회사 제공
[더스탁=김태영 기자] 주문형반도체 디자인 솔루션 전문기업 에이직랜드가 세계 최초로 뉴로모픽 프로세스를 상용화한 글로벌 AI반도체 기업 '브레인칩(BrainChip)'과 신규 반도체 설계 및 공급계약을 체결했다고 23일 밝혔다.
이번 계약은 브레인칩의 2세대 뉴로모픽 AI 프로세서인 ‘AKD2500’ 개발을 위한 프로젝트다. 에이직랜드는 이번 프로젝트에서 고객의 뉴로모픽 Core를 적용한 ASIC 디자인과 웨이퍼 패키징, 테스트에 이르는 Level 1 범위의 종합 설계 서비스를 제공할 예정이다.
브레인칩은 인간 뇌의 신경망 구조를 모방한 ‘뉴로모픽(Neuromorphic)’ 기술을 세계 최초로 상용화한 기업이다. 핵심 제품인 ‘Akida’ 프로세서는 데이터 변화가 있을 때만 연산이 이뤄지는 스파이킹 신경망(SNN) 구조를 채택해 에너지 효율을 극대화했으며, 기기 자체에서 실시간 학습이 가능한 ‘온 칩 러닝(On-chip Learning)’ 기능을 제공한다. 이를 통해 자율주행, 로보틱스, 웨어러블, 지능형 IoT 등 다양한 피지컬 AI(Physical AI) 분야에서 글로벌 시장을 확대하고 있다.
특히, 이번에 협력하는 ‘AKD2500’은 브레인칩의 2세대 아키텍처를 기반으로 하며, 최근 화두가 되고 있는 로보틱스 및 피지컬 AI 분야에서 요구되는 초저전력ㆍ고효율 연산에 최적화된 칩이다.

에이직랜드는 이번 계약을 통해 기존 NPU 설계를 넘어 차세대 기술인 뉴로모픽 AI 반도체까지 설계 포트폴리오를 확장하게 됐다. 이는 국내 시장을 넘어 글로벌 팹리스 고객사를 유치할 수 있는 중요한 기술적 도약이 될 것으로 기대된다.

에이직랜드 이종민 대표는 “세계적인 뉴로모픽 AI 기업인 브레인칩과의 협력은 에이직랜드의 ASIC 설계 역량이 글로벌 수준임을 입증하는 사례”라며, “이번 프로젝트의 성공적인 수행으로 글로벌 팹리스 고객사 유치를 위한 기술적 교두보를 공고히 하고, 글로벌 매출 비중을 적극적으로 확대해 나갈 것” 이라고 밝혔다.

한편, 에이직랜드는 국내 유일의 TSMC VCA(Value Chain Alliance)로서 축적된 선단 공정 설계 노하우를 바탕으로 다양한 산업 분야의 고객사들과 협력을 넓혀가고 있다.


Translation into English by Google Translate:


Asicland signs a new contract with 'Brain Chip', a global neuromopic AI leader

Reporter Kim Tae-young

Input 2026.02.23 13:05

Brainchip core design partner winning... Attracting overseas customers and accelerating revenue structure diversification

Asicland X Brain Chip.

Photo=provided by the company

[The Stock=Reporter Kim Tae-young]
Asicland, a company specialising in on-demand semiconductor design solutions, announced on the 23rd that it had signed a new semiconductor design and supply contract with BrainChip', a global AI semiconductor company that commercialised the world's first neuromorphic process.

This contract is a project for the development of 'AKD2500', Brainchip's second-generation neuromorphic AI processor. In this project, Asicland will provide comprehensive design services ranging from Level 1 to ASIC design, wafer packaging, and testing that apply the customer's neuromopic Core.

Brainchip is the world's first company to commercialise 'Neuromorphic' technology that imitates the neural network structure of the human brain. The core product, the 'Akida' processor, maximises energy efficiency by adopting a spiked neural network (SNN) structure that performs calculations only when there is a data change, and provides 'On-chip Learning' function that allows real-time learning on the device itself. Through this, we are expanding the global market in various physical AI (Physical AI) fields such as autonomous driving, robotics, wearable, and intelligent IoT.

In particular, the 'AKD2500' that we are cooperating with this time is based on the second-generation architecture of brain chips, and it is a chip optimised for ultra-low power and high-efficiency computing required in the field of robotics and physical AI, which has become a hot topic recently.

Through this contract, A-Group will expand its design portfolio beyond the existing NPU design to the next-generation technology of neuromopic AI semiconductors. This is expected to be an important technological leap that can attract global fabless customers beyond the domestic market.

"The cooperation with Brainchip, a global neuromographic AI company, is a case that proves Asicland's ASIC design capabilities are at a global level," and said, "With the successful implementation of this project, we will consolidate the technical bridgehead to attract global fabricless customers and actively expand the proportion of global sales."

On the other hand, Asicland is expanding cooperation with customers in various industries based on the accumulated side process design know-how as the only TSMC VCA (Value Chain Alliance) in Korea.

Thanks @Frangipani,

Exciting progression.

Cheers

 

[smoothsailing18]

So by your logic and other's, a foundry is only charging it's customer's to produce their IP/Chip designs the variable costs? Moreover, what is the logic behind the metric "volumes < $10" on the same page?

PISS OFF YOUR A WASTE OF FRESH AIR

 

[smoothsailing18]

Product Release Timeline and Availability
Renesas began its formal commercial journey with BrainChip technology following a definitive intellectual property (IP) license agreement signed in December 2020.[5] [6] This agreement granted Renesas the rights to integrate the Akida 1.0 neural network IP into its own SoC designs.[7]

The rollout of products featuring BrainChip technology has occurred in phases:

Initial Integration (2022–2023): Renesas first implemented the Akida IP on its R-Car V3H SoC platform.[4] This device was designed to accelerate AI applications in automotive ADAS (Advanced Driver Assistance Systems) and industrial automation. By December 2022, Renesas confirmed it was "taping out" a new chip using SNN technology on a 22nm CMOS process node.[1]

Pilot Production and OSAT (2025–2026): Through a joint venture with CG Power and Stars Microelectronics, Renesas is establishing an Outsourced Semiconductor Assembly and Test (OSAT) facility in Sanand, India.[8] A pilot production line is expected to be completed by July 2025, with the first chips (potentially including those with integrated AI IP) expected to roll out by mid-2026.[9] [10]

Mass Production (2027): Full-scale mass production for the next generation of these integrated chips at the Sanand facility is slated to begin in October 2027.

 

[perceptron]

Product Release Timeline and Availability
Renesas began its formal commercial journey with BrainChip technology following a definitive intellectual property (IP) license agreement signed in December 2020.[5] [6] This agreement granted Renesas the rights to integrate the Akida 1.0 neural network IP into its own SoC designs.[7]

The rollout of products featuring BrainChip technology has occurred in phases:

Initial Integration (2022–2023): Renesas first implemented the Akida IP on its R-Car V3H SoC platform.[4] This device was designed to accelerate AI applications in automotive ADAS (Advanced Driver Assistance Systems) and industrial automation. By December 2022, Renesas confirmed it was "taping out" a new chip using SNN technology on a 22nm CMOS process node.[1]

Pilot Production and OSAT (2025–2026): Through a joint venture with CG Power and Stars Microelectronics, Renesas is establishing an Outsourced Semiconductor Assembly and Test (OSAT) facility in Sanand, India.[8] A pilot production line is expected to be completed by July 2025, with the first chips (potentially including those with integrated AI IP) expected to roll out by mid-2026.[9] [10]

Mass Production (2027): Full-scale mass production for the next generation of these integrated chips at the Sanand facility is slated to begin in October 2027.

What is your source?

 

[Flenton]

BRN price per unit = (fixed costs + (variable costs * number of units)) / number of units

Cost to customer per unit = BRN price per unit * (1 + % profit)

Make sense?

 

[HopalongPetrovski]

So by your logic and other's, a foundry is only charging it's customer's to produce their IP/Chip designs the variable costs? Moreover, what is the logic behind the metric "volumes < $10" on the same page?

Hi Perceptron.
I don't follow how you have come to this conclusion, but, one last time....

Q....a foundry is only charging it's customer's to produce their IP/Chip designs the variable costs?
A. the sales price should exceed the purchase price, plus all associated costs of being in business, plus a profit margin, which is the incentive.

You have had ample opportunity to state your case or issue simply and clearly yet you persist with this line of interrogation and a continually evolving agenda, which frankly is just pissing everyone off.

I can only conclude that you either cannot or will not come to your point.
And so, I don't see any value in continuing this dialogue.

GLTAgenuine Holder's

 

[7für7]

I will sell everything just to see if it will skyrocket to 5 dollars… then I will buy in again! And if it goes down again I will sell at the bottom! Old T&J wisdom!!

Hahaha you losers

 

[TopCat]

Now this could be something that takes off. While we think about robots and defence and cars, this side of Akida could prove fruitful as well. Note: I really know nothing about lines of code , especially billions of them, and I may be wrong, but if AILabs are excited….

Minsky CodeAudit — AI That Audits, Fixes & Modernizes Your Code | Ai Labs Inc. | AiLabs Inc | Ai Labs Inc

Anthropic launches AI tool to convert COBOL to modern languages. This is a big signal. AI-driven legacy code translation is no longer experimental — it’s becoming mainstream. And yes — this directly challenges traditional legacy modernization consulting models (including firms that built...

www.linkedin.com

Global enterprises are sitting on billions of lines of legacy code — especially in:
• Banking
• Insurance
• Utilities
• Government
• Manufacturing

Traditional consulting-led migration is:
* Expensive
* Slow
* Human dependent
* Risk heavy

AI-native platforms like Minsky CodeAudit change the economics completely.
This isn’t just code conversion.

It’s AI-powered system transformation.

The modernization wave is here.
The firms that leverage AI-first migration platforms will lead the next decade of enterprise transformation.

 

[Diogenese]

Hi ILL,

"It may come cannonballing down from the sky" Q4 2026?!!!! -

if we're not all being hallucinated.

Until then, it is "classified"?

We have seen snippets of references to electronically steering antennae and SDR, so could be, who knows?

If Stralink is proposing to use Akida, it would probably be the 1500 which is 22 nm FDSoI (Although FG has opted for 28 nm for improved radhard).

That said, Starlink is lower than FG, so less rad.

 

[Pom down under]

The logic and the mathematics in relation to the metrics provided and interpreted (by some here) makes no business sense. Do Hoppy's interpretation of the metrics after researching variable costs to manufacture a silicon chip and research if there are other costs associated with manufacturing silicon chips that a foundry, not the IP/Chip designer has to factor in when pricing their production costs. It's well worth the effort.

Hope you’ve factored in the postage as well as it’s getting quite expensive these days

 

[manny100]

Its important for holders and potential holders to understand the stages of adoption and hold/add/ Buy/sell/avoid according to their personal risk tolerance.
See article below concerning adoption stages of new tech.
Unfortunately there are no shortcuts.
It breaks down into % of all total eventual adopters how many % at each stage. These of course are estimates and may vary.
At the moment we are in the Early adopters stage heading firmly towards the Early majority stage where we will start making real in roads - not to far away IMO. There are plenty of articles on the web.
So we are only at the stage where only about 16% of all eventual adopters are on board - but not far off the next stage.
Breakdown:
Innovators 2.5% - we have padded this stage.
Early Adopters 13.5% - We are in this stage heading firmly toward the next stage.
Early Majority 34% - The early majority are interested in technology but want proof of its effectiveness. We need clients demonstrating AKIDA's usefulness. Eg Megachips with Robots, Parsons with Defense, MetaGuard.
Late Majority 34% - Much like the early majority, the late majority want a data-driven reason to adopt technology. Convincing people in this adopter category requires research and solid proof that the technology is worth their time. Risk averse.
Laggards 16% - Laggards are wary of new technology. Before they consider getting on board, they need answers to their “what’s in it for me?” (WIIFM) questions. This stage is about changing 'mindsets'.

Technology Adoption Curve: 5 Stages of Adoption | Whatfix

Improve your next technology implementation by learning to identify all 5 adopter categories in the technology adoption curve.

whatfix.com