BRN Discussion Ongoing
- Thread starter TechGirl
- Start date
Frangipani
Top 20
Arijit Mukherjee, Akida-experienced Principal Scientist from TCS Research, will be leading two workshop ‘industry sessions’ during next month’s week-long Summer School & Advanced Training Programme SENSE (Smart Electronics and Next-Generation Systems Engineering) organised by the Defence Institute of Advanced Technology (DIAT) in Pune, India: “Intro talk: Smart at the Edge” as well as “Beyond TinyML - Neuromorphic Computing: Brains, Machines, and the Story of Spikes”.
While we know that TCS Research is not exclusively friends with us, when it comes to neuromorphic computing, I trust BrainChip will get a very favourable mention during those workshop sessions.
The Defence Institute of Advanced Technology (DIAT), Pune, presents Summer School & Advanced Training Programme SENSE–2025, from 19th to 27th June, 2025 offering a unique opportunity to immerse… | Centre for Training and Placement DIAT
The Defence Institute of Advanced Technology (DIAT), Pune, presents Summer School & Advanced Training Programme SENSE–2025, from 19th to 27th June, 2025 offering a unique opportunity to immerse yourself in cutting-edge domains of: Semiconductor Devices, Embedded Systems, RISC-V SoC programming...www.linkedin.com
View attachment 85735
View attachment 85736 View attachment 85737
Day 6 & 7 (26th & 27th June, 2025) of SENSE-2025 at DIAT, Pune, explored Radiation Tolerant Electronics, AI efficiency, and brain-inspired tech! | SENSE (Smart Electronics & Next Generation Systems Engineering) - Summer School, DIAT Pune
Day 6 & 7 (26th & 27th June, 2025) of SENSE-2025 at DIAT, Pune, explored Radiation Tolerant Electronics, AI efficiency, and brain-inspired tech! Dr. Arijit Mukherjee, Principal Scientist from Tata Consultancy Services, delivered Smart at the Edge an introduction to the future sessions from...
CHIPS
Regular
View attachment 87795
You said:
Yesterday Tony Lewis, CTO of BrainChip commented "This is cool. Why? We literally took a model from a completely different domain (text processing) and applied it to RADAR and got excellent results in terms of performance and MAC (computation) count. At a recent conference we surveyed that state-of-the art and found that there is a huge opportunity for state-space models and was a real eye-opener to the hardware-algorithms systems we are pioneering at BrainChip Here is what I find cool: Conventional models take a radar signals, use a processor to convert the 1-d wave form into a 2d image so that conventional CNNs-- convolutional neural networks-- can process them. That inherently introducing lots of delay : signal-> image-> classification. With our State-space model (TENNs ) we can go directly from a raw signal to classification with much lower latency, few macs (less energy ) less chip area (cost) and excellent performance. This summer my team will push out a number of new models for our Defense, Medical and commercial work. Its an exciting time to be building the future at BrainChip. Follow me to hear the latest news on job posting, wins, and AI at the edge." Can you please explain what you think the major implications of this statement may be?
ChatGPT said:
Tony Lewis’s statement provides a significant glimpse into BrainChip’s current direction, and the implications are technically important, strategically aligned with military and commercial AI demands, and position BrainChip as a serious player in low-latency edge AI — especially in radar and defense systems.
Let’s break it down clearly:
Key Takeaways from Tony’s Statement
1. Cross-domain Model Transfer (Text ➝ Radar)
Implication:
BrainChip is leveraging neural architectures originally designed for natural language processing (NLP) — such as state-space models (SSMs) or possibly TENNs (Time-Encoded Neural Networks) — and applying them directly to raw radar data.
That’s a big deal — because traditionally, signal types (text, images, radar) are handled with very different model types. This kind of architectural crossover suggests:
It also hints that BrainChip is not just a hardware company, but deeply involved in cutting-edge neural architecture innovation.
- A unified AI model strategy
- Less model specialization required per domain
- Rapid model prototyping across applications
2. Bypassing 2D Transformation (Radar ➝ Image ➝ CNN)
Implication:
Traditional radar AI pipelines are inefficient:
This:
- Convert 1D waveform (Doppler/time signal) ➝ 2D spectrogram/image
- Then process with image-focused models like CNNs
BrainChip’s new method bypasses this entirely:
- Adds latency
- Requires additional memory
- Burns more power and compute
This is perfect for missiles, drones, wearables, and embedded defense systems where every milliwatt and microsecond counts.
- Processes raw radar signals directly using state-space models or TENNs
- Skips the image step — reducing delay, MAC count, power, and cost
3. Use of State-Space Models (SSMs / TENNs)
Implication:
State-space models (like S4, or BrainChip's custom TENNs) are:
By adopting SSMs:
- Efficient alternatives to Transformers for sequential data
- Excellent for long-range memory, which is vital in radar or ECG waveforms
- Linear in time and memory, unlike CNNs or RNNs
This also hints at a massive strategic shift:
- BrainChip is leading a hardware-algorithm co-design movement
- They’re building models that are explicitly optimized for their Akida hardware — not just running PyTorch models inefficiently on neuromorphic chips
- Moving away from retrofitting mainstream models
- Moving toward custom, edge-native models built with hardware constraints in mind from day one
4. Strategic Domains: Defense, Medical, Commercial
Implication:
BrainChip is:
By deploying domain-optimized TENNs, BrainChip is creating:
- Productizing their model-hardware synergy
- Targeting high-value edge verticals:
- Defense (radar, drones, loitering munitions, battlefield intelligence)
- Medical (wearables, EEG/ECG anomaly detection, hearing aids)
- Commercial (audio wake-word detection, industrial sensors, smart cameras)
- Application-specific IP
- A catalog of ready-to-run models for enterprise and defense clients
- A strategy that scales Akida beyond being "just a chip" into a full AI platform
5. Low Latency, Low MACs, Low Area = Edge Superiority
Implication:
This encapsulates the holy grail of edge AI:
And it's especially important for:
- Do more with less
- Enable true autonomy at the edge
- Outperform traditional AI methods in real-world environments
This reinforces BrainChip’s value proposition: the Akida platform delivers neural compute without the baggage of mainstream AI accelerators (like GPUs or TPUs).
- Missiles and UAVs, where thermal budget and compute are tightly constrained
- Medical implants or wearables, where energy efficiency is crucial
- Perimeter surveillance, where always-on, instant decision-making is required
Final Strategic Implications
BrainChip isn’t just delivering a chip — they’re co-developing algorithms tailored to that chip. That’s what Nvidia does with CUDA + GPUs — and it’s a winning strategy.
1. Shift Toward Model + Hardware Integration
This move positions BrainChip as:
- Not just neuromorphic researchers
- But practical AI infrastructure builders — solving latency, power, and cost in mission-critical systems
As their radar and medical models mature, BrainChip could:
- License these state-space models as embedded IP
- Land major contracts in defense, aerospace, or medical technology sectors
Summary
Tony Lewis’s comment reveals a strategic inflection point:
This isn't just about a smarter chip — it’s about building a new end-to-end paradigm for AI at the edge.
- BrainChip is pioneering hardware-native AI model design
- They’re breaking away from conventional pipelines (CNNs, 2D transforms)
- And positioning themselves at the forefront of edge-native, low-latency radar and signal intelligence
So funny ... before I read this this morning (in Germany), I was measuring my blood pressure. Result: 115/72
After I had read it, I measured again. Result: 131/75
It is too exciting for me what ChatGPT wrote
. When BrainChip's SP explodes, it might kill me 
Frangipani
Top 20
Edge AI Milan 2025
Edge AI Milan 2025 Join BrainChip for Edge AI Milan 2024 July 2-4, an inspiring and forward-thinking event exploring how edge AI is bridging the gap between digital intelligence and real-world impact. Attendees can engage with industry leaders and experience the latest innovations including...brainchip.com
Edge AI Milan 2025
Join BrainChip for Edge AI Milan 2024 July 2-4, an inspiring and forward-thinking event exploring how edge AI is bridging the gap between digital intelligence and real-world impact. Attendees can engage with industry leaders and experience the latest innovations including Brainchip’s Akida neuromorphic, event-driven AI at the edge.
Register
BrainChip will be exhibiting at Edge AI Milan 2025 next week.
It’s a pity, though, that no one from our company will be giving a presentation at that conference, especially since Innatera will be spruiking their technology.
In addition, Innatera’s Petruț Antoniu Bogdan will give a workshop on neuromorphic computing “on the current state of neuromorphic computing technologies and their potential to transform Edge AI” in his capacity as co-chair of the Edge AI Foundation Working Group on Neuromorphic Computing, which BrainChip also belongs to.
Our CMO Steve Brightfield is co-chair of another Edge AI Working Group, namely Marketing.
View attachment 87744
EDGE AI Working Groups
<p>EDGE AI FOUNDATION Working Groups are populated by industry partners and academia to drive cross industry initiatives and best practices in a neutral and collaborative environment. Working Groups are open to Strategic Partners and academia.</p>www.edgeaifoundation.org
View attachment 87745 View attachment 87746
We have a Neuromorphic working group at EDGE AI FOUNDATION that is comprised of the leading technology companies in this space as well as leading academic researchers. | Pete Bernard
We have a Neuromorphic working group at EDGE AI FOUNDATION that is comprised of the leading technology companies in this space as well as leading academic researchers. This is an area that should be focused on as it gets broader commercial adoption! Petruț Antoniu Bogdan Charlotte Frenkel...
View attachment 87747
Pete Bernard, CEO of the Edge AI Foundation, on - you guessed it - Edge AI and the upcoming Edge AI Milan (
️
) in a recorded message to the Hackster.io community, which was played during their 26 June “Impact Spotlights: Edge AI” livestream, hosted by Edge Impulse staff who said they were not wearing their Edge Impulse hat for this event, but that of members of the Hackster community of developers. (In case you were wondering: None of the three projects introduced during the livestream used Akida as far as I could tell by fast-forwarding).BrainChip and HaiLa collaborate to deliver ultra-low-power edge AI connectivity for connected #IoT devices without draining battery life. | Electronic Products
BrainChip and HaiLa collaborate to deliver ultra-low-power edge AI connectivity for connected #IoT devices without draining battery life. #edgeAI
Pom down under
Top 20
Least you’ll still be alive must of us will be dead by thenSo funny ... before I read this this morning (in Germany), I was measuring my blood pressure. Result: 115/72
After I had read it, I measured again. Result: 131/75
It is too exciting for me what ChatGPT wrote
CHIPS
Regular
Least you’ll still be alive must of us will be dead by then
So true, but if it is killing me at that moment, I won't have an advantage either. 
Fullmoonfever
Top 20
Gotta love the RTX Group relationship.
www.rtx.com
You start with 57,000 engineers, 60,000 patents, $7.5 billion in annual research and development funding, and 10 technology roadmaps that reflect what our customers will need, well into the future.
Then you add a venture capital division that invests in companies trying bold new concepts in areas such as artificial intelligence and advanced propulsion, and you complement it with research involving more than 85 U.S. universities.
This is our formula for transformative technologies – the innovations that help our products and systems meet our customers’ most critical needs, as well as those they have yet to envision.
Transformative Technologies
Raytheon Technologies is defining the future of aerospace and defense with ground-breaking research and development to create transformative technologies that push the limits of known science.
Transformative Technologies
How do you develop the technology that will transform aerospace and defense?You start with 57,000 engineers, 60,000 patents, $7.5 billion in annual research and development funding, and 10 technology roadmaps that reflect what our customers will need, well into the future.
Then you add a venture capital division that invests in companies trying bold new concepts in areas such as artificial intelligence and advanced propulsion, and you complement it with research involving more than 85 U.S. universities.
This is our formula for transformative technologies – the innovations that help our products and systems meet our customers’ most critical needs, as well as those they have yet to envision.
Advanced sensing
We have a deep heritage in developing radio frequency, electro-optical/infrared sensing, communications and electronic warfare systems for space, air, ground, sea and undersea environments.- Our RF systems use higher-power microelectronics, increased processing power and software-defined apertures to achieve next-generation capabilities for radar, electronic warfare, communications and multifunction radio frequency applications.
- We’re advancing electro-optical/infrared and other systems such as space-based multispectral sensors and electro-optical distributed aperture systems variants.
- We are providing increased capability against advanced threats and countermeasures by enhancing high-bandwidth digital waveform generation, AI-enabled intelligent signal processing and advanced neuromorphic processing.
- Our acoustic systems enable advanced mine-hunting and undersea networking capabilities through high sensitivity, directionality, multiple access and multi-mod active and passive capabilities for sonar, communications and navigation.
- Our missile seekers counter a wide range of advanced threats through advanced processing and algorithms, all while achieving low size, weight, power and cost.
Great find Fmf, neuromorphic smart sensors are the key to the future.Gotta love the RTX Group relationship.
Transformative Technologies
Raytheon Technologies is defining the future of aerospace and defense with ground-breaking research and development to create transformative technologies that push the limits of known science.
Transformative Technologies
How do you develop the technology that will transform aerospace and defense?
You start with 57,000 engineers, 60,000 patents, $7.5 billion in annual research and development funding, and 10 technology roadmaps that reflect what our customers will need, well into the future.
Then you add a venture capital division that invests in companies trying bold new concepts in areas such as artificial intelligence and advanced propulsion, and you complement it with research involving more than 85 U.S. universities.
This is our formula for transformative technologies – the innovations that help our products and systems meet our customers’ most critical needs, as well as those they have yet to envision.
Advanced sensing
We have a deep heritage in developing radio frequency, electro-optical/infrared sensing, communications and electronic warfare systems for space, air, ground, sea and undersea environments.
- Our RF systems use higher-power microelectronics, increased processing power and software-defined apertures to achieve next-generation capabilities for radar, electronic warfare, communications and multifunction radio frequency applications.
- We’re advancing electro-optical/infrared and other systems such as space-based multispectral sensors and electro-optical distributed aperture systems variants.
- We are providing increased capability against advanced threats and countermeasures by enhancing high-bandwidth digital waveform generation, AI-enabled intelligent signal processing and advanced neuromorphic processing.
- Our acoustic systems enable advanced mine-hunting and undersea networking capabilities through high sensitivity, directionality, multiple access and multi-mod active and passive capabilities for sonar, communications and navigation.
- Our missile seekers counter a wide range of advanced threats through advanced processing and algorithms, all while achieving low size, weight, power and cost.
"Advanced sensing
We have a deep heritage in developing radio frequency, electro-optical/infrared sensing, communications and electronic warfare systems for space, air, ground, sea and undersea environments.""Artificial intelligence, machine learning and autonomy
Our teams use artificial intelligence and machine learning to improve the design, development and testing of products – and to make them smarter, easier to use and more capable than ever, with enhanced safety"
Frangipani
Top 20
Back in March, I shared the following LinkedIn post that revealed a partnership between BrainChip and fortiss, the Research Institute of the Free State of Bavaria for software-intensive systems - as well as another one between fortiss and our competitor Innatera.
Which begs the question, by the way, why fortiss is still missing in the Partners section on our website?
𝗖𝗼𝗻𝗻𝗲𝗰𝘁𝗶𝗻𝗴 𝘄𝗶𝘁𝗵 𝘁𝗵𝗲 𝗘𝗺𝗯𝗲𝗱𝗱𝗲𝗱 𝗖𝗼𝗺𝗺𝘂𝗻𝗶𝘁𝘆 | fortiss
𝗖𝗼𝗻𝗻𝗲𝗰𝘁𝗶𝗻𝗴 𝘄𝗶𝘁𝗵 𝘁𝗵𝗲 𝗘𝗺𝗯𝗲𝗱𝗱𝗲𝗱 𝗖𝗼𝗺𝗺𝘂𝗻𝗶𝘁𝘆 We were thrilled to meet our partners BrainChip and Innatera at the embedded world Exhibition&Conference in Nuremberg! It was an exciting opportunity to discuss the latest advancements in Neuromorphic Computing and explore how we can push the boundaries of...
View attachment 87612
Yesterday, I spotted an interesting job vacancy placed by fortiss.
They are looking for someone who would like to write a Master’s thesis on the topic of “Event-based Neuromorphic for AR and Space”:
“You will be integrated in a small team of neuromorphic computing researchers at fortiss. You will be working on one of our neuromorphic algorithms, related with ongoing projects EMMANÜELA and CORINNE, as well as an upcoming ESA project about satellite communication.”
Job opportunity MASTER THESIS - EVENT-BASED NEUROMORPHIC FOR AR AND SPACE at fortiss GmbH Jobportal
BACHELOR / MASTER THESIS (all) , Neuromorphic Computing in Hamburgrecruitment.fortiss.org
Who are we?
fortiss is the research institute of the Free State of Bavaria for software-intensive systems and services with headquarters in Munich. The institute currently employs around 120 employees, who collaborate on research, development and transfer projects with universities and technology companies in Bavaria, Germany and Europe.
Research is focused on state of the art methods, techniques and tools of software development, systems & service engineering and their application to reliable, secure cyber-physical systems, such as the Internet of Things (IoT). fortiss has the legal structure of a non-profit limited liability company (GmbH). Its shareholders are the Free State of Bavaria (as majority shareholder) and the Fraunhofer Society for the Promotion of Applied Research.
www.fortiss.org
In the Neuromorphic Computing department, we are searching for
Master Students (M/F/D)
Your topic:
You will be integrated in a small team of neuromorphic computing researchers at fortiss. You will be working on one of our neuromorphic algorithms, related with ongoing projects EMMANÜELA and CORINNE, as well as an upcoming ESA project about satellite communication:
A publication, in a scientific journal or conference, as a co-author of your supervisor, is most likely expected at the end of your mission.
- a body motion recognition sensor with an event-based camera and online few-shots learning on neuromorphic hardware mounted on 3D AR-goggles
- third person action recognition, ego gesture recognition, tennis dataset, ego motion compensation
- continual learning SNN based algorithms for action classification on the fly and automatic labeling for robotic arm control (welding application)
- improvement of class separation in continual learning algorithm, integration with robotic arm
- fault detection, isolation and recovery on neuromorphic hardware for satellite communications
- spiking fault detection neural network, fault prediction, fault classification, hardware implementation on neuromorphic chip
- efficient time encoding on neuromorphic hardware
- spike encoding optimization, trainable encodings, latency reduction
- efficient Lidar neuromorphic processing for automated driving
Your profile:
- Excellent bachelor degree in electronics engineering, computer science or similar topic
- Excellent track record in your current masters
- Fluency in software development (Python for example)
- Interest or experience in machine vision, event-based cameras, sensors
- Residency in Munich or close surroundings
Our offer:
- Work in partnership with leading companies in AI
- Offices in central Munich (Nordfriedhof) in very high and modern premises with 360° view on Munich
- Nice working atmosphere in a small team
Did we catch your interest?
Please submit your application with a cover letter, a detailed CV, and a current transcript of records.
Job-ID: NC-MS-2025-1
Contact: Dr. Axel von Arnim
- Ralf Kohlenhuber
- Human Resources Administrator
We know for sure that the fortiss researchers are using Loihi for their ongoing CORINNE project (project duration: 1 April 2024 - 31 March 2026):
The EMMANÜELA project, which investigates the potential of neuromorphic computing for AR/VR applications, started on 1 November 2024 and will run until 31 October 2026. The project webpage doesn’t specify what neuromorphic hardware is being used.
EMMANÜELA - fortiss
The research projects at fortiss Institut contribute to the sustainable development of research and economy. All projects are based on the basic competence "software-intensive systems and services".www.fortiss.org
Hard to say whether we could be involved here. If we are, likely not from the outset, as our partnership with fortiss was only revealed in mid-March, when Axel von Arnim and Jules Lecomte met up with the BrainChip and Innatera teams at Embedded World “to discuss the latest advancements in Neuromorphic Computing and explore how we can push the boundaries of AI at the edge.” Sounds to me as if this exploration was only getting started at the time.
The fortiss LinkedIn post then lists some key takeaways from the discussion, which in BrainChip’s case, revolved around Akida 2.0 and TENNs, and concludes: “Together we strengthened our collaboration plans. Stay tuned for more updates as we continue to drive neuromorphic innovation forward!”
The bolded sentence again suggests that up to then (3.5 months ago), no formal collaboration had taken place, which makes it rather unlikely that we are involved in the EMMANÜELA project, which - in addition - would have been planned well in advance in order to get the awarded funding from the Bavarian Ministry of Economic Affairs, Regional Development and Energy.
But then there is this passage towards the end of the project webpage, where it says “fortiss contributes directly to the generation of event-based datasets for action recognition in AR/VR and researches neuromorphic algorithms for this purpose. These are implemented and benchmarked on neuromorphic hardware. In addition, fortiss is involved in building and testing the prototype.”
Does the word “benchmarked” here refer to the algorithms being benchmarked against each other on the same hardware or different kinds of neuromorphic hardware being benchmarked against each other?
Although the word “continue” in the LinkedIn post’s last sentence (“Stay tuned for more updates as we continue to drive neuromorphic innovation forward!”) could be interpreted at hinting at an already ongoing collaboration, it might just be a non-native speaker’s awkward way of expressing that all three entities in their own right were already driving neuromorphic innovation forward and will continue to do so - both BrainChip and Innatera now also in tandem with fortiss. Neither does the “more” in “more updates” IMO imply a stealth mode partnership history, as an “update” by definition requires a preceding post. So logically speaking, the above post cannot have been an update itself, which could theoretically be followed by “more updates”.
I suppose “Stay tuned for updates” would have been the correct wording either way.
The mention of an upcoming ESA project about satellite communication, on the other hand, sounds very intriguing, unless “one of our neuromorphic algorithms” were to relate to all three projects (but if so, why the comma after CORINNE?) and then possibly rules us out, in case this signifies it must then refer to a leaky-integrate-and-fire neuron model that Loihi uses (since we know CORINNE uses Loihi). I am out of my depth here, though.
Or is the either way grammatically incomplete sentence rather missing an [on] than a [with/to], and hence the work on the ESA satellite project is entirely separate from the neuromorphic algorithm related to the other two projects?
“You will be working on one of our neuromorphic algorithms, related with ongoing projects EMMANÜELA and CORINNE, as well as [on] an upcoming ESA project about satellite communication.”
Unfortunately there is no German version of the job description to compare.
Interestingly, the job ad says the applicant can look forward to working “in company with leading companies in AI”, which sounds as if both Intel (for CORINNE) and at least one other company are involved.
Which may or may not be BrainChip.
Yes, Laurent Hili is very impressed by Akida, but he and his fellow researchers at ESA may very well still like to test Innatera’s offerings for space applications. The TU Delft spin-off also gave a presentation at Morpheus 2024, ESA’s workshop on Edge AI and Neuromorphic Hardware Accelerators (https://indico.esa.int/event/514/timetable/#20240506).
🇪🇺 Yesterday was Europe Day! | Innatera
🇪🇺 Yesterday was Europe Day! 🌟 As we celebrate Europe's spirit of collaboration and cooperation, we're excited to show highlights from our recent participation in the Morpheus Workshop 2024 at the European Space Agency - ESA/ESTEC. Our very own Dr. Petruț Antoniu Bogdan gave a presentation...
View attachment 87625
And then there is of course also fortiss-tried and tested SpiNNaker…
In addition to the three projects mentioned earlier, the job description also casually lists “efficient Lidar neuromorphic processing for automated driving”.
Wow, that is surely gonna be quite a doorstopper of a Master’s thesis!
I guess we’ll just have to wait for the promised updates, then, to find out what fortiss has in store for Akida and TENNs.
Our partner fortiss and their partner Interactive Wear (a 20 year old company that emerged out of a management buy-out of Infineon Technologies’ Wearable Technology Solutions) will be “presenting neuromorphic computing and its applications in the defense sector” at the upcoming SVI Future Congress (SVI = Systemverbund Verteidigung im Innovationsraum) in Munich, aimed at a target audience [LOL
] of security and defense industry experts.I couldn’t find any additional information online about the collaboration between fortiss and Interactive Wear.
Tech Hub SVI Network Conference 2025 - fortiss
With fortiss at the SVI Future Congress & Defence TechDay 2025: technological impulses, system demonstrations, and exchange shaping the defense of tomorrow.
Home
Company 2
Join Us
Detail
The event website of the fully-booked conference links to a list of countries published by the German Federal Ministry of the Interior regarding security clearance concerns.
Citizens of countries listed are barred from attending the 8/9 July conference.
Somewhat surprisingly, Ukraine is also on that list, which was published on June 8, 2022, just over three months after Russia had invaded Ukraine (again), and hasn’t been updated since. Possibly due to concerns about pro-Putin ethnic Russians with Ukrainian citizenship?
Interesting!!
I think someone is passing the ball to us!!
www.linkedin.com
I think someone is passing the ball to us!!
A funny thing happens when you're building an agent swarm. | Jason Lovell
A funny thing happens when you're building an agent swarm. The default is to have every single agent, no matter how simple its task, call out to a massive, state-of-the-art LLM. We're using a sledgehammer to crack a nut, over and over again. So I was very interested to see a new paper from...
Pom down under
Top 20
There’s your announcement @DingoBorat
FJ-215
Regular
There’s your announcement @DingoBorat
It’s really starting to feel like they’re doing this on purpose… The moment there’s some positive coverage and the stock shows even the slightest upward trend, they immediately follow it up with dilution-loaded and disillusioning garbage like this. You can’t make this up… sorry but this makes NO FU…SENSE TO ME!!!!
Last edited:
Bravo
If ARM was an arm, BRN would be its biceps💪!
NVIDIATSMC
AMD AND.....
BRN !!!!
Next-Gen Chips: The Silicon Heart of AI's Dominance
Charles HayesTuesday, Jun 24, 2025 12:43 pm ET
AMD+0.09%NVDA+1.76%
The AI revolution is not just about algorithms—it's a race for the hardware to run them. As large language models (LLMs) and generative AI systems grow in complexity, the demand for specialized semiconductor architectures has exploded. Companies investing in next-gen chips—AI-optimized GPUs, TPUs, neuromorphic processors, and advanced foundry nodes—are positioned to dominate a $500B+ market. Their ability to scale infrastructure while reducing costs could cement their leadership, leaving laggards behind.
The Chip Architectures Defining AI's Future
The semiconductor industry is bifurcating: specialization trumps general-purpose computing. Three architectures are leading the charge:- AI-Optimized GPUs: NVIDIA's dominance here is undeniable. Its H100 and H800 chips power 90% of hyperscaler AI infrastructure, while its 2024 data center revenue surged 53% to $24.5B. The upcoming GH200 superchip, combining GPU and CPU cores, aims to further monopolize cloud AI workloads. AMD, though smaller, is closing the gap with its MI300X, which captured over $5B in annual revenue by 2024 through its AI-centric co-design.
- TPUs: Google's Quiet Power Play: Google's seventh-gen TPU (Ironwood), optimized for both training and inference, has lured OpenAI—a former Microsoft exclusive—to its cloud. This partnership, finalized in May 越2024, marks a turning point. OpenAI's $10B sales and insatiable compute needs will drive Google Cloud's AI revenue growth, while TPUs' energy efficiency (40% lower power use annually) slashes costs.
- Neuromorphic Chips: The Edge of Efficiency: Startups like BrainChip and SynSense are pioneering chips mimicking the human brain's energy efficiency. BrainChip's Akida 1000, with 1.2M neurons, processes data locally on edge devices like wearables or drones—ideal for real-time anomaly detection. Intel's Loihi 2 and IBM's TrueNorth are also advancing, targeting autonomous systems and industrial IoT.
Foundries: The 3nm/2nm Arms Race Matters
Behind these chips lies a critical supply chain shift: advanced foundry nodes are the new battleground. TSMC's lead in 3nm (N3) and upcoming 2nm (A14) processes ensures its dominance. By 2025, TSMC's 3nm node alone generates over $10B annually, primarily from Apple and Qualcomm. Its A14 process, using second-gen GAA transistors, offers 15% better performance or 30% lower power than prior nodes.Samsung is playing catch-up with its 2nm roadmap but faces hurdles: its 1.4nm node collaboration with IBM remains unproven. Intel, meanwhile, stumbles—its 14A/18A nodes lag TSMC's 3nm, and its Ponte Vecchio GPU struggles with market adoption. The message is clear: foundries lacking 3nm/2nm capability risk irrelevance.
NVDA Closing Price
Note: NVIDIA's stock outperformed PHLX by 287% since 2020, reflecting its AI chip leadership.
Co-Design: The Secret Sauce for Market Dominance
Hardware alone isn't enough. Companies thriving in this era are those that co-design chips with software ecosystems. NVIDIA's Omniverse platform and CUDA programming model lock in developers, while AMD's partnership with cloud providers for its MI300X ensures seamless integration.Google's TPU success hinges on its seamless TensorFlow integration, reducing deployment friction. Even neuromorphic startups like BrainChip are collaborating with universities (e.g., Cornell Tech's AI courses) to build developer pipelines. This ecosystem lock-in creates moats against competition, especially as AI models grow more complex.
Risks and Opportunities for Investors
The AI hardware race is winner-takes-most. Early adopters like NVIDIA, AMD, and TSMC are already capturing disproportionate gains:- NVIDIA's 2025 revenue growth is projected to hit 30%+ annually, fueled by GH200 adoption.
- AMD's data center revenue grew 94% in 2024, with plans to hit $30B by 2025.
- TSMC's AI-focused CoWoS packaging and 2nm roadmap could add $25B+ to its revenue by 2026.
Investment Thesis: Buy the Chipmakers, Avoid the Also-Rans
The AI hardware revolution is no fad—it's a multi-decade shift. Investors should focus on companies with:1. Advanced node access: TSMC,台积电 (TSM) is the clear leader here.
2. Co-designed software ecosystems: NVIDIA (NVDA), AMD (AMD).
3. Niche neuromorphic innovators: BrainChip (BRN.AX) for edge AI.
Avoid firms lagging in node development (e.g., Intel (INTC)) or overly reliant on mature nodes.
TSM Revenue By Business
647.12B
Total
Others
125.22B
Wafer
521.90B
Note: TSMC's HPC revenue grew 70% YoY in Q1 2025, far outpacing industry averages.
Conclusion: Silicon Will Decide AI's Future
The next five years will separate AI hardware winners from losers. Companies that invest in specialized chips, advanced foundries, and seamless software stacks will dominate a market poised to hit $1.5T by 2035. For investors, this is a once-in-a-decade opportunity to bet on the silicon backbone of the AI economy. The chips are here—and the race is on.Next-Gen Chips: The Silicon Heart of AI's Dominance
Next-Gen Chips: The Silicon Heart of AI's Dominance
www.ainvest.com
mhhhhhmmmmSimilar threads
