Pom down under
Top 20
Way too long for me
BRN Discussion Ongoing
- Thread starter TechGirl
- Start date
Frangipani
Top 20
OHB Hellas are not only exploring Akida for their ‘Satellite as a Service’ concept (GIASAAS)
, but also as a consortium partner for an ESA project calledBOLERO (On-Board Continual Learning for SatCom Systems).
Prime contractor of the BOLERO project is KPLabs - both OHB Hellas and Eutelsat OneWeb are subcontractors.
BOLERO
Our mission is to connect Europe. We join engineers, entrepreneurs and investors to forge strong links between institutions, industries, and businesses.
BOLERO On-Board Continual Learning For SatCom Systems
- Next generation of systems
- Status
Ongoing - Status date
2025-05-01 - Activity Code
1B.138
The project identifies, explores, and implements onboard continual machine learning techniques to enhance reliability and data throughput of communication satellites.
The first objective is to identify 3 most promising use cases and applications of continual learning (CL), together with at least two most promising hardware platforms.
The second objective is to implement different CL techniques in the selected scenarios and assess their performance and feasibility for onboard deployment using the selected hardware platforms. The assessment includes the analysis of advantages and trade-offs of CL in comparison to traditional offline machine learning approaches, and the comparative analysis of hardware platforms for CL.
The final goal of the project is to identify a state-of-the-art, potential gaps, and future roadmap for CL in satellite communication systems.
Challenges
The main challenge is related to the limited resources and limited support for continual machine learning mechanisms in existing onboard processing units, e.g., not all operations and layers are supported and model parameters cannot be updated without hardware-specific recompilation. Therefore, common CL approaches are not straightforward to implement on board. Additionally, CL techniques come with the stability-plasticity trade-offs and the need of continuous validation and monitoring.
Benefits
The project offers a complete software and hardware pipeline to implement 3 different continual machine learning approaches (i.e., class-incremental, domain-incremental, and task-incremental) in 3 different application for communication satellites. The comparative analysis helps to identify which approach and hardware platform is best suited for different CL scenarios. The project establishes the foundation for future development of CL in SatCom systems.
Features
- Structured and informed report from selecting 3 most promising applications of CL in communication satellites and 2 hardware platforms
- Code for running a complete onboard CL process for both hardware platforms
- Report containing technology gaps and future roadmap for CL in SatCom systems
System Architecture
The 3 CL applications identified in the project are implemented for two hardware platforms of very different architectures (KP Labs Leopard DPU and BrainChip Akida neuromorphic computer). For each application and platform, there is a complete CL pipeline architecture proposed from data preprocessing to onboard continual learning.
Current status
The 3 most promising applications of continual machine learning in communication satellites have been identified, i.e., domain-incremental beam hopping optimization, task-incremental inter-satellite links routing, and class-incremental telemetry anomaly classification.
For each application, a state-of-the-art CL approach has been implemented for two diverse hardware platforms identified as the most promising ones for CL (KP Labs Leopard DPU and BrainChip Akida neuromorphic computer). The performance of each CL approach has been assessed and main technology gaps have been identified.
documentation
Documentation may be requestedPrime Contractor
KP Labs Sp. z o. o.
Polandhttps://kplabs.space
Subcontractors
OHB HELLAS
GreeceWebsite
Eutelsat OneWeb (OW)
United Kingdomhttps://oneweb.net/
Last update
2025-05-03 12:39
BOLERO: On-Board Continual Learning for SatCom Systems - [Article Title] | Explore KP Labs' Innovations and Insights
Discover insights from KP Labs in our news update, BOLERO: On-Board Continual Learning for SatCom Systems, focusing on key developments and innovations in satellite systems and space exploration.
www.kplabs.space
PROJECT
4 min read
BOLERO: On-Board Continual Learning for SatCom Systems
Published on
January 28, 2025
In an era of exponentially increasing data generation across all domains, satellite communications (SatCom) systems are no exception. The innovative BOLERO project, led by KP Labs, and supported by a consortium including OHB Hellas and Eutelsat OneWeb, is at the forefront of this technological evolution. This project is making significant strides in applying both classic and deep machine learning (ML and DL) techniques within the dynamic realm of satellite data, marking a transformative step in SatCom technology.
Understanding the Need for Continual Learning in SatCom
Traditionally, satellite applications have relied on supervised ML algorithms trained offline, with all training data prepared before the training process begins. This method is effective in stable data scenarios. For example, a deep learning model can accurately identify brain tumor lesions from magnetic resonance images after being trained on a diverse dataset. However, the dynamic space environment presents unique challenges. Factors such as thermal noise, atmospheric conditions, and on-board noise can significantly alter data characteristics, causing these offline-trained models to struggle or fail when encounteringnew, unfamiliar data distributions.The BOLERO Approach
BOLERO addresses these challenges by adopting an online training paradigm. The training process is shifted directly to the target environment, such as an edge device on a satellite. This innovative approach bypasses the need for downlinking large amounts of data for Earth-based retraining, overcoming bandwidth and time limitations. Training models in their deployment environment accelerate the training-to-deployment cycle and significantly improve model reliability under dynamic conditions.Tackling New Challenges
Implementing continual learning brings its own challenges, including catastrophic forgetting, where models may lose previously acquired knowledge. Additionally, the stability-plasticity dilemma must be addressed to ensure models are adaptable and capable of retaining learned information. BOLERO tackles these issues through strategies such as task-incremental learning, allowing models to adapt to new tasks, and domain-incremental learning, enabling them to handle data with evolving distributions.
The Consortium’s Collaborative Dynamics in BOLERO
The BOLERO project is propelled by the synergistic efforts of its consortium members. As the project leader, KP Labs is primarily responsible for developing the Synthetic Data Generators (SDGs) and the continual learning models, ensuring their efficacy across multiple SatCom applications and hardware architectures. OHB Hellas contributes by exploring novel machine learning methodologies suitable for streaming data, assessing continual learning applications in and beyond the space sector, and implementing two use cases in different hardware modalities. Eutelsat OneWeb focuses on identifying strategic space-based applications for continual learning, evaluating their business impact, and analyzing the benefits of continual learning models, particularly in terms of performance and cost-efficiency. Together, these entities combine their unique strengths to advance the BOLERO project, addressing the evolving demands of SatCom systems.
Real-World Applications and Future Impact
The applications of BOLERO are diverse, ranging from monitoring the operational capabilities of space devices to gas-level sensing and object detection in satellite imagery. These applications highlight the potential of continual learning to enhance the efficiency and accuracy of SatCom systems, potentially revolutionizing the management and processing of satellite data for more responsive, agile, and efficient operations.The BOLERO project, led by KP Labs and supported by a consortium including OHB Hellas and Eutelsat OneWeb, represents a groundbreaking step in harnessing the full potential of continual learning for SatCom systems. By confronting the unique challenges associated with satellite data and leveraging the latest in ML technology, BOLERO is poised to significantly improve the adaptability and efficiency of SatCom systems, setting a new standard in the field of satellite communications.
The other promising hardware platform being tested is KPLabs’ own Leopard DPU: https://www.kplabs.space/solutions/hardware/leopard
Frangipani
Top 20
OHB Hellas are not only exploring Akida for their ‘Satellite as a Service’ concept (GIASAAS)
BOLERO (On-Board Continual Learning for SatCom Systems).
Prime contractor of the BOLERO project is KPLabs - both OHB Hellas and Eutelsat OneWeb are subcontractors.
View attachment 83668
BOLERO
Our mission is to connect Europe. We join engineers, entrepreneurs and investors to forge strong links between institutions, industries, and businesses.connectivity.esa.int
BOLERO On-Board Continual Learning For SatCom Systems
Objectives
- Next generation of systems
- Status
Ongoing- Status date
2025-05-01- Activity Code
1B.138
The project identifies, explores, and implements onboard continual machine learning techniques to enhance reliability and data throughput of communication satellites.
The first objective is to identify 3 most promising use cases and applications of continual learning (CL), together with at least two most promising hardware platforms.
The second objective is to implement different CL techniques in the selected scenarios and assess their performance and feasibility for onboard deployment using the selected hardware platforms. The assessment includes the analysis of advantages and trade-offs of CL in comparison to traditional offline machine learning approaches, and the comparative analysis of hardware platforms for CL.
The final goal of the project is to identify a state-of-the-art, potential gaps, and future roadmap for CL in satellite communication systems.
Challenges
The main challenge is related to the limited resources and limited support for continual machine learning mechanisms in existing onboard processing units, e.g., not all operations and layers are supported and model parameters cannot be updated without hardware-specific recompilation. Therefore, common CL approaches are not straightforward to implement on board. Additionally, CL techniques come with the stability-plasticity trade-offs and the need of continuous validation and monitoring.
Benefits
The project offers a complete software and hardware pipeline to implement 3 different continual machine learning approaches (i.e., class-incremental, domain-incremental, and task-incremental) in 3 different application for communication satellites. The comparative analysis helps to identify which approach and hardware platform is best suited for different CL scenarios. The project establishes the foundation for future development of CL in SatCom systems.
Features
- Structured and informed report from selecting 3 most promising applications of CL in communication satellites and 2 hardware platforms
- Code for running a complete onboard CL process for both hardware platforms
- Report containing technology gaps and future roadmap for CL in SatCom systems
System Architecture
The 3 CL applications identified in the project are implemented for two hardware platforms of very different architectures (KP Labs Leopard DPU and BrainChip Akida neuromorphic computer). For each application and platform, there is a complete CL pipeline architecture proposed from data preprocessing to onboard continual learning.
Current status
The 3 most promising applications of continual machine learning in communication satellites have been identified, i.e., domain-incremental beam hopping optimization, task-incremental inter-satellite links routing, and class-incremental telemetry anomaly classification.
For each application, a state-of-the-art CL approach has been implemented for two diverse hardware platforms identified as the most promising ones for CL (KP Labs Leopard DPU and BrainChip Akida neuromorphic computer). The performance of each CL approach has been assessed and main technology gaps have been identified.
documentation
Documentation may be requested
Prime Contractor
KP Labs Sp. z o. o.
Poland
https://kplabs.space
Subcontractors
OHB HELLAS
Greece
Website
Eutelsat OneWeb (OW)
United Kingdom
https://oneweb.net/
Last update
2025-05-03 12:39
BOLERO: On-Board Continual Learning for SatCom Systems - [Article Title] | Explore KP Labs' Innovations and Insights
Discover insights from KP Labs in our news update, BOLERO: On-Board Continual Learning for SatCom Systems, focusing on key developments and innovations in satellite systems and space exploration.
PROJECT
4 min read
BOLERO: On-Board Continual Learning for SatCom Systems
Published on
January 28, 2025
In an era of exponentially increasing data generation across all domains, satellite communications (SatCom) systems are no exception. The innovative BOLERO project, led by KP Labs, and supported by a consortium including OHB Hellas and Eutelsat OneWeb, is at the forefront of this technological evolution. This project is making significant strides in applying both classic and deep machine learning (ML and DL) techniques within the dynamic realm of satellite data, marking a transformative step in SatCom technology.
Understanding the Need for Continual Learning in SatCom
Traditionally, satellite applications have relied on supervised ML algorithms trained offline, with all training data prepared before the training process begins. This method is effective in stable data scenarios. For example, a deep learning model can accurately identify brain tumor lesions from magnetic resonance images after being trained on a diverse dataset. However, the dynamic space environment presents unique challenges. Factors such as thermal noise, atmospheric conditions, and on-board noise can significantly alter data characteristics, causing these offline-trained models to struggle or fail when encounteringnew, unfamiliar data distributions.
The BOLERO Approach
BOLERO addresses these challenges by adopting an online training paradigm. The training process is shifted directly to the target environment, such as an edge device on a satellite. This innovative approach bypasses the need for downlinking large amounts of data for Earth-based retraining, overcoming bandwidth and time limitations. Training models in their deployment environment accelerate the training-to-deployment cycle and significantly improve model reliability under dynamic conditions.
Tackling New Challenges
Implementing continual learning brings its own challenges, including catastrophic forgetting, where models may lose previously acquired knowledge. Additionally, the stability-plasticity dilemma must be addressed to ensure models are adaptable and capable of retaining learned information. BOLERO tackles these issues through strategies such as task-incremental learning, allowing models to adapt to new tasks, and domain-incremental learning, enabling them to handle data with evolving distributions.
The Consortium’s Collaborative Dynamics in BOLERO
The BOLERO project is propelled by the synergistic efforts of its consortium members. As the project leader, KP Labs is primarily responsible for developing the Synthetic Data Generators (SDGs) and the continual learning models, ensuring their efficacy across multiple SatCom applications and hardware architectures. OHB Hellas contributes by exploring novel machine learning methodologies suitable for streaming data, assessing continual learning applications in and beyond the space sector, and implementing two use cases in different hardware modalities. Eutelsat OneWeb focuses on identifying strategic space-based applications for continual learning, evaluating their business impact, and analyzing the benefits of continual learning models, particularly in terms of performance and cost-efficiency. Together, these entities combine their unique strengths to advance the BOLERO project, addressing the evolving demands of SatCom systems.
Real-World Applications and Future Impact
The applications of BOLERO are diverse, ranging from monitoring the operational capabilities of space devices to gas-level sensing and object detection in satellite imagery. These applications highlight the potential of continual learning to enhance the efficiency and accuracy of SatCom systems, potentially revolutionizing the management and processing of satellite data for more responsive, agile, and efficient operations.
The BOLERO project, led by KP Labs and supported by a consortium including OHB Hellas and Eutelsat OneWeb, represents a groundbreaking step in harnessing the full potential of continual learning for SatCom systems. By confronting the unique challenges associated with satellite data and leveraging the latest in ML technology, BOLERO is poised to significantly improve the adaptability and efficiency of SatCom systems, setting a new standard in the field of satellite communications.
The other promising hardware platform being tested is KPLabs’ own Leopard DPU: https://www.kplabs.space/solutions/hardware/leopard
View attachment 83667
BOLERO – OHB-Hellas
www.ohb-hellas.gr
Because it would need to be like a 25-1 consolidation and they would issue a heap more shares at the same time to raise capital
Thankyou For you reply,
My shares are in fact in Australian Super also so I guess I’ll be selling them sooner rather than later.. at a loss unfortunately.
The AGM will hopefully shed some light on this 5 year plan and more detail on current engagements. No doubt it will be high jacked by questions on moving to the US to a degree but I sincerely hope that we are afforded the insight, without divulging detail, on where we “actually” stand in our commercial journey.
Partnerships are great but no business survives on minimal income and massive overheads forever.
Labsy
Regular
Anduril! Yeah! Was thinking the same.
For those who are not familiar, they should look this up. They just ousted Microsoft from defense for VR technology in defence amongst providing advanced drones etc etc ....
Space Cadet
Regular
My Super fund must be wrong then because I asked and they said I would have to sell them. I can't transfer them.
SC
Okeydokey01
Regular
I agree with you dippY22 that engagements along the lines of those you outline seem to be the most probable catalyst. For every line of reasoning I have gone down these continue to remain the most likely. Assuming the water tight NDA and obvious security requirements around these types of engagements I wouldn’t expect anything to be able to be released to market.
However perhaps this would be a time where actions could speak louder than words. Something along the lines of the board suspending performance bonuses and awarding of RSU's until revenue of $xxx was achieved.
If this move is indeed to satisfy the contractual requirements of one of these types of partners and we are that close. Then a pause in bonuses by the board, knowing what is coming would not be unreasonable, in my opinion. It would certainly speak volumes and go a long way to rebuilding trust for me.
Pom down under
Top 20
If we join as a 3rd tier company on the nasdaq then the requirement is $1 USD per share so 5-1 if the share price gets to 0.3 au. Information about 3rd tier supplied by other so can’t confirm if this information is correct. But still not good but loads better than 25-1. So maybe I good announcement with $$$ value then no consolidation I recon
HopalongPetrovski
I'm Spartacus!
What are the tiers of the Nasdaq market?
The Nasdaq Stock Market has three distinctive tiers: The Nasdaq Global Select Market®, The Nasdaq Global Market® and The Nasdaq Capital Market®. Applicants must satisfy certain financial, liquidity and corporate governance requirements to be approved for listing on any of these market tiers.
What Is the Nasdaq Capital Market?
The Nasdaq Capital Market is one of Nasdaq's U.S. market tiers containing early-stage companies that have relatively lower market capitalizations. Listing requirements for companies on the Nasdaq Capital Market are less stringent than for the two other Nasdaq market tiers, which focus on larger companies with higher market capitalization.KEY TAKEAWAYS
- The Nasdaq Capital Market (Nasdaq-CM) is one of three listing tiers on the Nasdaq exchange, specifically for companies that need to raise capital.
- Companies listed here may be small companies with a need to grow capital or shell corporations designed to raise capital in public markets for the purpose of acquiring other business entities.
- Companies that don't qualify for the Nasdaq National Market trade on Nasdaq-CM.
- Nasdaq Capital Market companies are required to meet a net income standard of at least $750,000, a minimum public float of 1,000,000 shares, at least 300 shareholders, and a share bid price of at least $4 (with certain exceptions).
Understanding the Nasdaq Capital Market
The Nasdaq Capital Market, known until 2005 as the Nasdaq SmallCap Market, primarily lists so-called small-cap stocks (typically those with market capitalizations of around $300 million to $2 billion). The name change reflected a shift in focus towards listing companies that need to raise capital. It's meant to be a less encumbered entrance for a smaller company or a Special Purpose Acquisition Company (SPAC) to capitalize and grow through a Nasdaq listing.12Although the initial listing requirements are relaxed, the corporate governancerequired to maintain a Nasdaq listing is the same across all tiers. This means Nasdaq Capital Market companies must have a code of conduct, an audit committee, independent directors, and so on.3
Listing Requirements for the Nasdaq Capital Market
The Nasdaq Capital Market makes it easier for early-stage companies to get listed, especially when compared to other senior exchanges with more onerous requirements. To list initially on the Nasdaq Capital Market, companies must meet all of the criteria under at least one of three listing standards—the equity standard, the market value of listed securities standard, or the net income standard.All the standards share some requirements such as one million publicly held shares, 300 shareholders, and three market makers (MMs). However, these also differ in important ways. The equity standard requires stockholders' equity of $5 million, where the other two require only $4 million, and it also requires an operating history of two years, while the other two do not require an operating history.
The market value of listed securities requires a market value of listed securities of $50 million and a market value of publicly held shares of $15 million. The net income standard is the only one requiring a net income, $750,000 in the latest fiscal year or in two of the last three years, but has the lowest requirement for market value of publicly held shares at $5 million.4
Nasdaq. “5500. The Nasdaq Capital Market.”
Although companies can pick the standard that best fits their situation, the listing standards and the required governance are more stringent than some early-phase capital markets. Because of the costs involved with meeting these standards, companies listing on the Nasdaq Capital Market often handily exceed the minimum requirements before they decide to list.5
Have you asked BRN whether they still have a commercial relationship with Marquee clients?
I have and a commercial relationship still exists..
They list the Marquee clients on the 'Why invest' part of the website and it has been pointed to them that if there is a change it must be reflected immediately for legal reasons.
Have you done some research about how long new and complex technology takes to be I produced into safety regulated autos and health?
Some posters use the word transperancy but never say precisely what it is.
Do you have a precise list of issues where you see there is a lack of transparency on? They cannot be dealt with unless precisely identified.
From time to time I ask questions for clarification and yes, I fish for Information on occasions like others do. Sometimes I get an answer and sometimes due to NDA's, competitive reasons etc I do not get a answer.
They did in fact provide their reasons for a redomicile in the Ann.
For perspective BRN is still a spec stock.
I think it will be difficult for BRN to get a 75% vote to redomicile if a majority of retail holders vote.
Around 74% of SOI is held by retailers.
BRN would need to provide significant incentive to holders to get a yes vote through IMO. That would likely involve a deal or 2 which give the SP a decent push along.
Sean did say in the Quarterly he expected a substantial increase in booking this year.
Watch for a significant placement to a US industry participant prior to a meeting (if there is one)..
I’m not gonna lie. I need Sean to present something at the AGM to convince me that deals and revenue are coming this year. Antonio deflecting criticism by telling shareholders that one big deal with a tech giant will make us profitable over night is not going to go down well this year.
Pom down under
Top 20
I'd assume the Brainchip team have or will be arriving in Sydney for their Monday morning meeting prior to the AGM on Tuesday, it's
important that they unite and are all on the same page before heading into the Lions Den!!
Some may agree with my choice of theme music...
Bravo
If ARM was an arm, BRN would be its biceps💪!
Published April 29, 2025
How Beken Leverages Arm to Pioneer Low-Power AI and Wireless Connectivity
Learn about low-power AI wireless connectivity and the Beken BK7259 chip's role in enhancing smart devices and IoT.
newsroom.arm.com
How Beken Leverages Arm to Pioneer Low-Power AI and Wireless Connectivity
Beken Corp.'s BK7259 chip, featuring Arm's Ethos-U65 NPU and Cortex-M cores, transforms smart devices with low-power AI processing and seamless wireless connectivity.By Arm Editorial Team
Artificial Intelligence (AI)Internet of Things (IoT)
Share
Reading time 6 min
Imagine a smart door lock that recognizes your face in under 200 milliseconds—securely, locally, and without draining battery life. That’s just one-fifth of a second—virtually instant to the human eye. Or AI toys that respond in real time using natural voice interactions. At the heart of these breakthroughs is Beken Corp.’s latest chip, the BK7259, a powerhouse of ultra-low power AI processing built for the edge.
A global leader in wireless connectivity and Wi-Fi microcontrollers, Beken is pushing the boundaries of what’s possible in smart devices, leveraging Arm’s Ethos-U65 NPU and Cortex-M cores.
As AI and IoT converge, Beken isn’t just adapting—it’s shaping the future of intelligent connectivity.
Beken’s AI Edge Apps Feature Low Power Consumption
Headquartered in Shanghai, China, Beken is a leading innovator in Wi-Fi microcontrollers (MCUs) and wireless connectivity for IoT. With expertise in RF-CMOS design and digital signal processing, Beken delivers high-performance, low-power SoC solutions that drive next-generation connected devices.Serving global leaders like Amazon, AT&T, LG, Samsung, and Sony, Beken generates over 1 billion RMB in annual revenue and holds a dominant market share in IoT Wi-Fi and China’s national ETC standard chips, shaping the future of wireless technology.
Recognizing the growing demand for AI-driven edge applications, Beken has expanded into smart devices and intelligent systems. Its latest innovation, the BK7259, integrates multiple Ethos-U65 NPU and Arm Cortex-M cores, delivering fast, efficient AI processing for edge applications.
A key differentiator of the BK7259 is its industry-leading low power consumption. Its Wi-Fi keep-alive mode draws less than 50 microamps, setting a world record for efficiency. This advancement enables battery-powered, always-connected AI devices, making edge AI more viable than ever. Beken also capitalizes on Arm’s robust AI ecosystem, utilizing CMSIS-NN and the Vela compiler to streamline AI model deployment and accelerate solution development.
As AI and IoT continue to converge, Beken remains at the forefront of innovation, enabling scalable, power-efficient AI solutions that drive the future of intelligent connectivity.
Empowering Development with Arm’s Ecosystem
Beken’s adoption of the Ethos-U65 NPU has been pivotal, offering access to open-source algorithms and AI development tools. The broad Arm developer ecosystem provides extensive support, enabling Beken to optimize AI workloads efficiently. The growing open-source AI community further enhances development by fostering collaboration and innovation at the edge.Optimizing AI Inference for Smart Devices and Intelligent Systems
One critical AI function in smart devices is real-time audio processing. Tasks like echo cancellation, noise suppression, and environmental noise filtering require both accuracy and efficiency. With the Ethos-U65 and Arm’s inference library, Beken optimizes these functions for low-power, real-time execution.Beken has also prioritized RTOS (Real-Time Operating Systems) over traditional Linux-based AI processing because of its potential for enabling efficient, low-latency applications.
Balancing Power Efficiency with High-Performance AI
Balancing performance and power efficiency is one of AI’s biggest challenges—especially at the edge. Beken has tackled this head-on by focusing on low-power, high-performance solutions like the BK7259.Built on a dual-domain AI architecture, the BK7259 combines the efficient M52 core for lightweight tasks and the high-performance M55+U65 for more demanding AI workloads. Unlike typical market configurations, this setup includes the advanced U65 NPU, delivering robust processing power for complex neural networks.
This architecture allows for flexible workload allocation—for example, routing low-power audio tasks to the M52 and running video analytics on the M55+U65—while enabling hierarchical wake-up strategies to reduce both idle and active power consumption. The result: a solution that meets the strict performance and energy demands of modern edge AI.
By integrating low-power Wi-Fi transceivers with AI compute, the BK7259 ensures real-time AI execution with minimal energy impact, making it an ideal platform for next-generation edge intelligence.
Leveraging Wireless Expertise for Smarter Devices
Beken’s expertise in Wi-Fi connectivity has played a critical role in its expansion into robotics. The company has developed Wi-Fi 4 to Wi-Fi 6E solutions that provide reliable connectivity while maintaining low power consumption.For IoT applications, stable wireless connectivity is essential for real-time data transmission and processing. Beken optimizes system-level operations to ensure reliable AI-driven decision-making while minimizing energy consumption.
Robust Security Architecture
The Beken system achieves PSA Certified Level 2 by implementing a multi-layered security strategy built on Arm TrustZone technology. This creates a hardware-isolated secure processing environment, further strengthened by dedicated components such as a Crypto Accelerator for efficient encryption and PUF OTP for generating unique device keys and secure storage.To ensure a standardized and robust security foundation, we use Trusted Firmware-M (TF-M), Arm’s open-source secure OS implementation. Together, these integrated hardware and software features enable our platform to meet the rigorous requirements of PSA Certified Level 2.
Balancing Edge Processing and Cloud Connectivity for AI Applications
Beken’s shift from IoT chip design to AI-driven smart devices presented several technical challenges, including hardware-software integration, low power consumption, and seamless cloud connectivity. AI applications require a balance between edge processing and cloud resources, ensuring reliability and minimal latency.For example, smart locks with AI-powered facial recognition need real-time local processing rather than relying on cloud servers. The BK7259’s integration of the Ethos-U65 NPU enables fast, energy-efficient authentication, providing high-speed recognition without compromising security.
Real-World AI Applications Powered by Beken
Beken’s chips power a range of AI-driven applications, showcasing the fusion of AI, wireless connectivity, and power efficiency in next-generation smart devices.- AI Toys – Interactive smart toys, from interactive robots to talking action figures, combine cloud and edge AI to deliver rich, multimodal experiences blending voice recognition with visual understanding. These toys can “think” on their own—processing voice interaction on the device side with ultra-low latency, making toys truly “understand and speak”.
- Smart Door Locks – Advanced facial recognition enables local authentication in under 200 milliseconds. Powered by the Ethos-U65 NPU for fast face detection and a dedicated ISP for clear imaging in low-light conditions, the system also features an H.264 codec for efficient video compression, plus a 2.5D GPU and DPU for high-quality graphical displays. Together, these components ensure smooth video streaming and responsive user interaction.
The Future of AI and IoT
Beken envisions a future where smart devices transition from basic remote-control functions to intelligent, autonomous interactions. AI will play a greater role in automating decision-making, reducing user intervention, and enhancing smart home experiences.To achieve this, Beken is developing a comprehensive range of Wi-Fi microcontrollers, from simple control units to AI-intensive processing cores. These solutions will drive seamless AI integration in home automation, predictive maintenance, and next-generation smart appliances.
Driving AI Innovation in Partnership with Arm
Beken’s partnership with Arm has been instrumental in its AI evolution. Arm’s power-efficient CPU and NPU architectures, coupled with its extensive AI software ecosystem, have provided a strong foundation for edge AI innovation.One key advantage is Arm’s investment in RTOS for edge AI applications, which aligns with Beken’s long-term vision. As AI at the edge continues to advance, Arm’s ongoing innovation in power-efficient AI computing will further enhance Beken’s capabilities and accelerate industry-wide adoption.
Advancing Edge AI with Power-Efficient Computing
Beken’s journey from wireless connectivity to AI-driven applications highlights the importance of power-efficient, high-performance computing. By leveraging Arm’s ecosystem, optimized semiconductor IP, and advanced AI processing, Beken is pushing the boundaries of what’s possible in smart devices and IoT.As edge AI adoption continues to grow, Beken remains committed to developing AI-powered solutions that balance energy efficiency, performance, and intelligent automation—ushering in the next era of smart connectivity and AI innovation.
Bravo
If ARM was an arm, BRN would be its biceps💪!
HopalongPetrovski
I'm Spartacus!
I think general sentiment is turning against the management.
There have always been people critical and others urging patience, but that very patience is wearing thin now, for many, myself included.
In the last few AGM's they have managed to retain support with various exhortations that revenue was coming and just hang on a bit longer and we'll all be rewarded.
But this time, if it's just more generalised motherhood statements and razzle dazzle techno pop, with more "look what we've got to sell" stuff, I don't think that'll cut it.
They need to start proving they can perform commercially or at the least show us they have a viable plan to get there.
I understand its a tough business and takes time but, it is a business, after all, and businesses that cannot eventually support themselves, tend to fail.
So does Belen's chip utilise Akida ... or is Beken opposition to Akida?View attachment 83684
Published April 29, 2025
How Beken Leverages Arm to Pioneer Low-Power AI and Wireless Connectivity
Learn about low-power AI wireless connectivity and the Beken BK7259 chip's role in enhancing smart devices and IoT.
How Beken Leverages Arm to Pioneer Low-Power AI and Wireless Connectivity
Beken Corp.'s BK7259 chip, featuring Arm's Ethos-U65 NPU and Cortex-M cores, transforms smart devices with low-power AI processing and seamless wireless connectivity.
By Arm Editorial Team
Artificial Intelligence (AI)Internet of Things (IoT)
Share
Reading time 6 min
Imagine a smart door lock that recognizes your face in under 200 milliseconds—securely, locally, and without draining battery life. That’s just one-fifth of a second—virtually instant to the human eye. Or AI toys that respond in real time using natural voice interactions. At the heart of these breakthroughs is Beken Corp.’s latest chip, the BK7259, a powerhouse of ultra-low power AI processing built for the edge.
A global leader in wireless connectivity and Wi-Fi microcontrollers, Beken is pushing the boundaries of what’s possible in smart devices, leveraging Arm’s Ethos-U65 NPU and Cortex-M cores.
As AI and IoT converge, Beken isn’t just adapting—it’s shaping the future of intelligent connectivity.
Beken’s AI Edge Apps Feature Low Power Consumption
Headquartered in Shanghai, China, Beken is a leading innovator in Wi-Fi microcontrollers (MCUs) and wireless connectivity for IoT. With expertise in RF-CMOS design and digital signal processing, Beken delivers high-performance, low-power SoC solutions that drive next-generation connected devices.
Serving global leaders like Amazon, AT&T, LG, Samsung, and Sony, Beken generates over 1 billion RMB in annual revenue and holds a dominant market share in IoT Wi-Fi and China’s national ETC standard chips, shaping the future of wireless technology.
Recognizing the growing demand for AI-driven edge applications, Beken has expanded into smart devices and intelligent systems. Its latest innovation, the BK7259, integrates multiple Ethos-U65 NPU and Arm Cortex-M cores, delivering fast, efficient AI processing for edge applications.
A key differentiator of the BK7259 is its industry-leading low power consumption. Its Wi-Fi keep-alive mode draws less than 50 microamps, setting a world record for efficiency. This advancement enables battery-powered, always-connected AI devices, making edge AI more viable than ever. Beken also capitalizes on Arm’s robust AI ecosystem, utilizing CMSIS-NN and the Vela compiler to streamline AI model deployment and accelerate solution development.
As AI and IoT continue to converge, Beken remains at the forefront of innovation, enabling scalable, power-efficient AI solutions that drive the future of intelligent connectivity.
Empowering Development with Arm’s Ecosystem
Beken’s adoption of the Ethos-U65 NPU has been pivotal, offering access to open-source algorithms and AI development tools. The broad Arm developer ecosystem provides extensive support, enabling Beken to optimize AI workloads efficiently. The growing open-source AI community further enhances development by fostering collaboration and innovation at the edge.
Optimizing AI Inference for Smart Devices and Intelligent Systems
One critical AI function in smart devices is real-time audio processing. Tasks like echo cancellation, noise suppression, and environmental noise filtering require both accuracy and efficiency. With the Ethos-U65 and Arm’s inference library, Beken optimizes these functions for low-power, real-time execution.
Beken has also prioritized RTOS (Real-Time Operating Systems) over traditional Linux-based AI processing because of its potential for enabling efficient, low-latency applications.
Balancing Power Efficiency with High-Performance AI
Balancing performance and power efficiency is one of AI’s biggest challenges—especially at the edge. Beken has tackled this head-on by focusing on low-power, high-performance solutions like the BK7259.
Built on a dual-domain AI architecture, the BK7259 combines the efficient M52 core for lightweight tasks and the high-performance M55+U65 for more demanding AI workloads. Unlike typical market configurations, this setup includes the advanced U65 NPU, delivering robust processing power for complex neural networks.
This architecture allows for flexible workload allocation—for example, routing low-power audio tasks to the M52 and running video analytics on the M55+U65—while enabling hierarchical wake-up strategies to reduce both idle and active power consumption. The result: a solution that meets the strict performance and energy demands of modern edge AI.
By integrating low-power Wi-Fi transceivers with AI compute, the BK7259 ensures real-time AI execution with minimal energy impact, making it an ideal platform for next-generation edge intelligence.
Leveraging Wireless Expertise for Smarter Devices
Beken’s expertise in Wi-Fi connectivity has played a critical role in its expansion into robotics. The company has developed Wi-Fi 4 to Wi-Fi 6E solutions that provide reliable connectivity while maintaining low power consumption.
For IoT applications, stable wireless connectivity is essential for real-time data transmission and processing. Beken optimizes system-level operations to ensure reliable AI-driven decision-making while minimizing energy consumption.
Robust Security Architecture
The Beken system achieves PSA Certified Level 2 by implementing a multi-layered security strategy built on Arm TrustZone technology. This creates a hardware-isolated secure processing environment, further strengthened by dedicated components such as a Crypto Accelerator for efficient encryption and PUF OTP for generating unique device keys and secure storage.
To ensure a standardized and robust security foundation, we use Trusted Firmware-M (TF-M), Arm’s open-source secure OS implementation. Together, these integrated hardware and software features enable our platform to meet the rigorous requirements of PSA Certified Level 2.
Balancing Edge Processing and Cloud Connectivity for AI Applications
Beken’s shift from IoT chip design to AI-driven smart devices presented several technical challenges, including hardware-software integration, low power consumption, and seamless cloud connectivity. AI applications require a balance between edge processing and cloud resources, ensuring reliability and minimal latency.
For example, smart locks with AI-powered facial recognition need real-time local processing rather than relying on cloud servers. The BK7259’s integration of the Ethos-U65 NPU enables fast, energy-efficient authentication, providing high-speed recognition without compromising security.
Real-World AI Applications Powered by Beken
Beken’s chips power a range of AI-driven applications, showcasing the fusion of AI, wireless connectivity, and power efficiency in next-generation smart devices.
These use cases demonstrate how Beken’s technology brings intelligent, efficient performance to everyday AI applications.
- AI Toys – Interactive smart toys, from interactive robots to talking action figures, combine cloud and edge AI to deliver rich, multimodal experiences blending voice recognition with visual understanding. These toys can “think” on their own—processing voice interaction on the device side with ultra-low latency, making toys truly “understand and speak”.
- Smart Door Locks – Advanced facial recognition enables local authentication in under 200 milliseconds. Powered by the Ethos-U65 NPU for fast face detection and a dedicated ISP for clear imaging in low-light conditions, the system also features an H.264 codec for efficient video compression, plus a 2.5D GPU and DPU for high-quality graphical displays. Together, these components ensure smooth video streaming and responsive user interaction.
The Future of AI and IoT
Beken envisions a future where smart devices transition from basic remote-control functions to intelligent, autonomous interactions. AI will play a greater role in automating decision-making, reducing user intervention, and enhancing smart home experiences.
To achieve this, Beken is developing a comprehensive range of Wi-Fi microcontrollers, from simple control units to AI-intensive processing cores. These solutions will drive seamless AI integration in home automation, predictive maintenance, and next-generation smart appliances.
Driving AI Innovation in Partnership with Arm
Beken’s partnership with Arm has been instrumental in its AI evolution. Arm’s power-efficient CPU and NPU architectures, coupled with its extensive AI software ecosystem, have provided a strong foundation for edge AI innovation.
One key advantage is Arm’s investment in RTOS for edge AI applications, which aligns with Beken’s long-term vision. As AI at the edge continues to advance, Arm’s ongoing innovation in power-efficient AI computing will further enhance Beken’s capabilities and accelerate industry-wide adoption.
Advancing Edge AI with Power-Efficient Computing
Beken’s journey from wireless connectivity to AI-driven applications highlights the importance of power-efficient, high-performance computing. By leveraging Arm’s ecosystem, optimized semiconductor IP, and advanced AI processing, Beken is pushing the boundaries of what’s possible in smart devices and IoT.
As edge AI adoption continues to grow, Beken remains committed to developing AI-powered solutions that balance energy efficiency, performance, and intelligent automation—ushering in the next era of smart connectivity and AI innovation.
Smoothsailing
Regular
Bravo
If ARM was an arm, BRN would be its biceps💪!
Hi mcm,
No. Unfortunately Beken doesn't utilize neuromorphic computing. It uses a more traditional AI approach. It's a SoC combining Arm's Cortex-M cores and Ethos-U65 NPU.
Arm's partnerships with Beken and BrainChip both focus on advancing ultra-low power AI at the edge.
In terms of absolute lowest power consumption, it looks like Beken's solution might be the winner, but it isn't capable of on-chip learning and only seems capable of handling simple NN tasks.
I guess we've lost our lead somewhat, and will likely start to see more competitors at the edge as we move forward.
Similar threads
