Connect with us
Anglostratits

Tech

Huawei Pioneers a New Computing Architecture for the AI Era: Making One Million Processors Work as One Computer

Published

on

Huawei

SHANGHAI, CHINA – Media OutReach Newswire – 30 September 2026 – During HUAWEI CONNECT on September 17, 2026, Eric Xu (Huawei’s Rotating Chairman) and Dr. Liao Heng (Chief Scientist of HiSilicon) had a Q&A session with journalists of media outlets to discuss Huawei’s Peerium Computing Architecture for the AI era as well as UnifiedBus (UB).

At last year’s HUAWEI CONNECT, Eric Xu gave a keynote speech and released Ascend chips and their roadmap (including the 950, 960, 970, and SuperPoDs and SuperClusters). He also announced the open release of the UnifiedBus protocol.

 




 
 

At this year’s HUAWEI CONNECT, Huawei announced the Ascend 950 and the Atlas 950 SuperPoD. While the industry is talking about SuperPoDs or supernodes, the ways of making these products differ hugely. The key of SuperPoDs is about allowing several thousand or even tens of thousands of processors to work as one computer.

Today, Huawei announced a SuperPoD based on the Ascend 950 chip. The SuperPoD is powered by a new computing architecture that Huawei has pioneered. This architecture now can make one million processors work as one computer. This architecture is called Peerium, and it can achieve strong scaling to the million-processor level through nested parallelism, unified memory addressing, and peer-to-peer interconnect. It extends the parallelism paradigm of Turing with the introduction of Nested BSP (Nested Bulk Synchronous Parallel).

This architecture also extends the von Neumann single-machine architecture and overruns the master-slave design paradigm that has prevailed for decades. Given that the available technologies in the industry could not support this architecture, Huawei invented a key interconnect technology – UnifiedBus. Ultimately, one million processors can truly work as one larger computer.

UnifiedBus is a high-speed bus that uses an open protocol to scale without a limit to connect CPUs, NPUs, memory, SSDs, interface cards, and switches. With UnifiedBus, peer-to-peer interconnect is achieved across compute, storage, and networking. This fundamentally overruns the traditional master-slave architecture.

The Atlas 950 SuperPoD is a product built upon the Peerium Computing Architecture. A SuperCluster with 256,000 computing cards is currently being deployed and tested.

Q: Dr. Liao, your paper mentioned a SuperPoD that’s made up of 256,000 cards. Is that the upper limit of your technology? Can you tell us more about market demand for your products?

Liao Heng: 256,000 or roughly 200k, this number is not an arbitrary number.

First of all, the purpose of these large clusters is for training. It has a hardware infrastructure that can support the training of foundation models, which today are already in the 5T parameter range.

Over the next two years, there will be roughly 6 or 7 frontier AI labs in China, and all will aim at training foundation language models of sizes ranging from 10 to 40 trillion parameters. These numbers kind of match up to the memory capacity and the size of the SuperPoD. So you need such a scale of infrastructure to train these models.

The second thing is that in China, most data centers have a national plan on this power grid. So 200,000 is a relatively conservative number, given the design of the power delivery system in a single autonomous zone and single data center campus.

As I said, everybody must be quite familiar with the AI model race that’s happening across the globe. And in China, there are at least 3 or 4 tier-1 players that are already well recognized and have achieved tier-1 status.

Q: What are your plans to encourage Chinese model providers to use Ascend chips for training?

Eric Xu: For the Ascend 950, we first launched the 950PR to the market. It’s primarily used for AI inference, and currently the total volume of supply available in the market is not very big. The SuperPoDs based upon the Ascend 950DT are mainly used for AI training. Right now, they are under testing, and their large-scale supply will start at the end of this year or early next year. We are having extensive dialogues and discussions with AI model providers. The testing result of using the Ascend 950DT for training is pretty good. I believe that starting from next year, a lot of AI model training will be based upon SuperPoDs that use the Ascend 950DT. I think at the end of the day, it’s not about how hard we are going to push model providers, but about how much supply capacity we will have. We can only supply based on how much that’s produced. And we hope the value chain will expand its capacity faster to increase supply.

Q: What is Huawei’s take on recent calls by leading AI model providers in the US to slow down AI development?

Eric Xu: We need to look at the level and cadence of AI development in China and the US. AI models in China are largely open-source models and their progress is relatively transparent. If we look at the mainstream model providers in the US, they have more computing power, so maybe only they themselves know where they are in terms of the level of sophistication of their models. The market feeling about AI risks might be weaker in China than in the US. Model providers in China need to speed up their pace to reach a level where they could also feel the risks from AI development. And then, maybe they will feel the same as the leading US model providers. But I always believe that we need to strike a balance between driving AI development and managing AI risks. At the very least, the bottom line is we need to ensure AI for good, not AI for evil.

Q: Does Huawei have plans to bring the Atlas 950 SuperPoD to other markets outside China? If yes, which markets are you going to target? Can you share with us data about your market share in China? What is your take on China’s push for chip self-sufficiency?

Eric Xu: Since we don’t have enough capacity to even satisfy the demand in China, we don’t have plans to expand to the international market in a fully-fledged way. But indeed, there are several countries that have very strong demand. We are doing some testing, and we are providing some supply to those countries, but the volume is quite limited.

Right now, it’s pretty hard to collect data about the market share of NVIDIA in China. But based on the data we have collected, Ascend has surpassed NVIDIA.

China’s push for chip self-sufficiency is an inevitable path forward. China is a country of 1.4 billion people and it’s an industrial powerhouse, so the way people work and live is closely tied to chips. Chinese people have a keen sense of urgency, and will not accept a future in which others decide whether or not we can have access to certain products. Even though our chips may be less advanced, at least their supply is assured, so that you don’t have to worry about chip supply day in and day out. I think that’s certainly the path forward. For the Chinese government, for the domestic industry, and for Huawei, the path forward is undoubtedly to push for full self-sufficiency in terms of chips and the entire semiconductor value chain. We also believe this will become a reality sooner or later.

To read the full Q&A session: https://www.apmultimedianewsroom.com/multimedia-newsroom/huawei-pioneers-a-new-computing-architecture-for-the-ai-era-making-one-million-processors-work-as-one-computer-2 




 

Business

GITEX World RobotiX 2026 brings humanoid debuts and next-gen embodied Artificial Intelligence (AI) to Dubai

Published

on

GITEX

New launch showcase at GITEX GLOBAL advances the Physical AI phenomenon as global humanoid robot shipments surpass 22,000 in H1 2026

DUBAI, United Arab Emirates, November 10, 2026/APO Group/ –From interactive humanoids and intelligent machines undertaking industrial tasks to multilingual robot assistants and wearable exoskeletons, the latest advances in the Physical AI phenomenon will take centre stage at the world’s largest tech and AI event next month in Dubai.

 




 
 

GITEX World RobotiX, a new launch event at GITEX GLOBAL, will converge the world’s robotics pioneers and leaders in embodied AI from 8-11 December at Dubai Exhibition Centre, Expo City Dubai, to show how machines that perceive, reason and act in the physical world are transforming manufacturing, logistics, healthcare, hospitality and smart infrastructure. GITEX GLOBAL 2026 will kick off with the GITEX Summit on 7 December at Dubai World Trade Centre, ahead of the exhibition at Expo City Dubai.

 

The showcase arrives as robot demand surges; more than 22,000 humanoid robots were shipped globally in the first half of 2026, a 300 percent year-on-year increase (https://apo-opa.co/4kHlnGe), according to Counterpoint Research, with shipments expected to exceed 50,000 units by year-end. Dubai is also advancing its robotics ambitions through its Robotics and Automation Program, targeting the deployment of 200,000 robots over 10 years to 2032 (https://apo-opa.co/4j7b5y2).

 

Unitree, one of the world’s largest humanoid makers with 31 percent (https://apo-opa.co/4sNVwhW) global market share, spearheads participation this year, alongside other headliners including UBTECH, whose humanoids are deployed in car factories and being tested by Airbus; AiMOGA, Chery Auto’s embodied AI robotics brand; DEEP Robotics, a global leader in quadrupeds; and the UAE’s Micropolis Robotics, which provides autonomous patrol robots for Dubai Police.

 

Robots take on real-world tasks

 

UAE-based Mangobot, a robotics solutions provider specialising in deploying and integrating intelligent machines, will present a robotic workforce featuring autonomous general-purpose robots, patrol robots, multilingual robot receptionists and live robotic performances. Visitors can also experience wearable exoskeleton technology firsthand. “It’s less ‘come look at robots’ and more ‘come see what working alongside robots actually feels like,’” said Mazen El-Kurdi, CTO of Mangobot.

 

Making its Middle East debut, MOS 2, developed by Chinese embodied intelligence innovator Lumos Robotics, is a wheeled ‘AI Worker’ with dual arms capable of handling payloads of up to 50 kg and operating for more than eight hours.

 

According to Lumos Robotics, its first-generation MOS robots completed over 1,000 hours of continuous, failure-free operation at Mitsubishi Electric’s Changshu factory, undertaking material handling, quality inspection and screwdriving tasks. The company will also showcase NIX S3, which will deliver a live dance performance demonstrating its 21 degrees of freedom, high-torque joints, agility and dynamic motion capabilities.

 

Cecile Wei, PR Director, Lumos Robotics, said: “Our demonstrations are designed to show the philosophy behind our products: robots need to do more than move. They need to perform useful tasks reliably in real environments.”

 

UAE-based Lootah Robotics, a robotics solutions and systems integration company, will showcase humanoid, quadruped, service, cleaning and telepresence robots from manufacturers including Unitree, AGIBOT, Pudu Robotics, UBTECH and EngineAI.

 

“We are not simply supplying individual robots; we are building an integrated robotics ecosystem,” said Hazem Ghazal, General Manager of Lootah Robotics. “Our stand will showcase how organisations can select different robots for different requirements while maintaining centralised visibility and control.”

 

The next frontier of embodied AI and human-robot interaction

 

Advances in embodied AI – where AI models are brought into a physical body or hardware system – are creating opportunities for robots to interact more naturally with people across consumer services, education, entertainment and domestic environments.

We are not simply supplying individual robots; we are building an integrated robotics ecosystem

 

China’s Light Origins, which develops humanoid robots and intelligent robotic systems, will make the first public appearance of its ‘bunny-shaped’ robot, with live demos of household scenarios offering visitors hands-on interaction. “We believe our bunny-shaped robot is ready to enter consumers’ homes,” said Tingxiang Fan, Co-founder and CTO, Light Origins.

 

Meanwhile, Raydiculous, a bionic robotics innovator from the Chengdu Humanoid Robot Innovation Centre, will make its Middle East debut with two innovations. Its TP01 humanoid, designed for commercial interaction and public services, can communicate in English and Arabic and perform synchronised movements. Alongside it, DINOBOT, described by Raydiculous as the world’s first intelligent bipedal walking bionic dinosaur, will demonstrate realistic movement and interactive capabilities.

 

Sandy Liu, Overseas Sales Manager, Chengdu Humanoid Robot Innovation Centre, said: “With its integrated capabilities of walking, interaction, performance, and education, our solutions turn Physical AI into a visible and usable business tool.”

 

Supporting future robotics talent, UAE-based technology and robotics education company Besomi will introduce six humanoid and mobile robots known as the ‘Besomi Rangers’. Unitree, Seeed Studio, Arduino and Qualcomm will be joining Besomi on the stand. The showcase will feature the latest humanoid robot from Unitree, a global leader in humanoid robotics. A live Robotics & AI Lab will allow visitors to explore how robots are engineered, programmed and trained, with hands-on engineering training using the latest Arduino VENTUNO platform, developed in collaboration with Qualcomm.

 

“We want young engineers and innovators to ask, ‘How does this work?’

and then go build the next generation of robotics technologies right here in the UAE. Showcasing the latest Unitree humanoid robot gives visitors a look at the forefront of robotics. Our hands-on training with Arduino and Qualcomm gives engineers practical experience to turn their own ideas into working robotics and AI applications. We are also pleased to welcome Seeed Studio to our stand as part of this focus on building engineering capabilities,” said Mohammad Hani, CEO, Besomi Robotics & AI.

 

Building the intelligence behind autonomous machines

 

JWIPC, a Chinese developer of AI computing hardware and industrial computing solutions, will launch robotic controllers built on NVIDIA’s Thor module, which the company says delivers up to 2,070 TFLOPS of FP4 computing performance. Designed for humanoids, robotic arms and autonomous mobile robots, the controllers enable AI processing and real-time decision-making.

 

“The NVIDIA Thor module provides a high-compute edge brain for embodied intelligence, running multi-modal models locally to integrate sensor fusion, task planning, and millisecond motion control,” said Nick Zhang, Regional Sales Manager, JWIPC.

 

Meanwhile, Physical AI software developer Songying Technology will demonstrate ORCA OS and ORCA Lab, integrating simulation, synthetic data generation, AI training, validation and deployment into a connected robotics development environment. Taiwan-based embedded computing specialist Vecow will unveil advanced, rugged Edge AI and robotics platforms powered by NVIDIA, Intel, AMD, Qualcomm, and NXP, while officially introducing its dedicated UAE team to accelerate mission-critical regional deployments

 

The 46th edition of GITEX GLOBAL, organised by inD, opens with The GITEX Summit at Dubai World Trade Centre (DWTC) on 7 December, before the four-day GITEX GLOBAL Expo begins at Dubai Exhibition Centre alongside Expand North Star, its startup and investor event.

 

The dual-blockbuster showcase converges 6,000-plus enterprise and startup exhibitors with tens of thousands of government leaders, C-suite executives, investors and buyers from 180 countries, reaffirming Dubai’s status as an epicentre of tech collaboration.

 

More information is available at www.Gitex.com and (https://apo-opa.co/4iyQ67d)

*Sources: https://apo-opa.co/4kHlnGe

https://apo-opa.co/4j7b5y2

Distributed by APO Group on behalf of GITEX Global.

 

 




 

Continue Reading

Business

2 million hacking tool detections in six months: Kaspersky reveals the most used lateral movement utilities

Published

on

For the period of 6 months, April to September 2026, the solution detected 2,171,513 hacking tool usage attempts

JOHANNESBURG, South Africa, November 9, 2026/APO Group/ –Recent statistics compiled from detection alerts within Kaspersky Anti Targeted Attack (KATA)* reveal the hacking tools most often used for lateral movement across enterprise infrastructures, highlighting the pressing need for reliable network security solutions to protect corporate environments.

 




  

Recent investigations into cyber-espionage groups show that attackers increasingly avoid detectable malware, choosing instead to blend into legitimate corporate network traffic (https://apo-opa.co/4BWK3AY). By weaponising dual-use tools and administrative utilities already present in the system, they move laterally across the domain undetected to harvest credentials and locate high-value assets. In this era of silent intrusion, Kaspersky shares insights into the most frequently used hacking tools to help organisations defend their networks.

For a decade, KATA, Kaspersky’s mature NDR solution, has been delivering comprehensive anti-APT protection. It features advanced threat detection, in-depth network traffic analysis, and centralised incident response to ensure full visibility across the entire network. Equipped with nearly 12,500 network detection rules in its database, KATA detects an average of 312,705 potential attacks daily.

For the period of 6 months, April to September 2026, the solution detected 2,171,513 hacking tool usage attempts. The following types of tools were detected most frequently:

This year, we are proudly celebrating the 10th anniversary of KATA

HackTool.Cobalt – 136,084 detections. While designed for security assessments, Cobalt Strike is also frequently used by malicious actors across various attack stages, including post-exploitation, lateral movement, and similar activities.

HackTool.NetExec – 129,560 detections. NetExec is used for automated network security assessments and discovering poorly secured network services. Its capabilities can also be exploited by attackers for reconnaissance and gaining system access.

HackTool.SQLMap – 107,697 detections. The SQLMap tool is used to automate the detection and exploitation of SQL injection vulnerabilities, which can allow unauthorised access to databases via web applications.

“Since advanced hacking tools may bypass endpoint defences, deploying Network Detection and Response (NDR) solution is critical to exposing silent anomalies and neutralising advanced attacks,” says Aren Torosyan, Product Manager at Kaspersky. “This year, we are proudly celebrating the 10th anniversary of KATA. Over this decade of innovation, backed by deep infrastructure expertise and rigorous research, KATA has established itself as a highly effective solution, helping clients worldwide protect their environments from even the most advanced, hidden threats across networks and endpoints. With the transition to the Open Single Management Platform planned for next year, advanced NDR capabilities will become even more comprehensive and accessible to security teams”.

To get full information about Kaspersky Anti Targeted Attack and request a demo, please visit the website (https://apo-opa.co/4BStEgE).

*Statistics are based on detection alerts from Kaspersky products, collected from users who consented to provide statistical data to Kaspersky Security Network. Some of the detections may be attributed to authorised security assessments.

Distributed by APO Group on behalf of Kaspersky.

 




 

Continue Reading

Business

PhotonPay Strengthens North American Regulatory Footprint with Canada’s RPAA Registration

Published

on

PhotonPay

The registration reinforces PhotonPay’s regulatory foundation for payment activities in Canada, with requirements covering operational risk management, incident response and the safeguarding of end-user funds.
HONG KONG SAR – Media OutReach Newswire – 10 November 2025 – PhotonPay today announced that its Canadian operating entity, Photon Dance CA Inc., is now registered with the Bank of Canada as a payment service provider (PSP) under Canada’s Retail Payment Activities Act (RPAA). The entity was previously registered with the Financial Transactions and Reports Analysis Centre of Canada (FINTRAC) as a Money Services Business (MSB). The RPAA registration further strengthens PhotonPay’s regulatory foundation in Canada and supports its continued development of local financial services across North America.

 




  

The RPAA establishes a federal framework for the registration and supervision of PSPs conducting specified retail payment activities in Canada. Administered by the Bank of Canada, it sets requirements for operational risk management, incident response and regulatory reporting. PSPs that hold end-user funds are also required to implement measures to safeguard those funds, strengthening oversight and accountability across Canada’s retail payments ecosystem.

The Canadian registration adds to PhotonPay’s broader regulatory presence across North America. In the United States, PhotonPay is registered as a Money Services Business with the Financial Crimes Enforcement Network (FinCEN) and holds money transmitter licenses across multiple U.S. states. Across the Americas, Europe and Asia, PhotonPay holds 35 payment licenses, regulatory authorizations and registrations, reflecting the regulatory breadth behind its operations across key financial markets.

For PhotonPay, building a global financial operating system requires more than connecting payment capabilities across markets. It also means developing the local regulatory and financial infrastructure needed to support those capabilities in each jurisdiction. As businesses expand internationally, their financial operations increasingly span different account structures, payment systems, currencies and regulatory environments. Bringing these functions together requires infrastructure that can address local market requirements while giving businesses a more unified way to manage their global finances.

This is the vision behind PhotonPay’s next-generation Financial OS. Through an integrated, API-first architecture, PhotonPay brings together multi-asset wallets, local collections, global payouts, foreign exchange, card issuing and embedded finance. By supporting both fiat currencies and stablecoins within its broader financial infrastructure, the platform connects traditional payment rails with emerging digital-asset capabilities, helping businesses manage increasingly complex financial operations across markets.

“Global financial infrastructure must be designed around the realities of the markets it serves. Regulatory frameworks, payment systems and the responsibilities involved in handling funds differ across jurisdictions, and these differences need to be accounted for at the infrastructure level,” said JP Pi, Global Chief Compliance Officer at PhotonPay. “Our approach is to build a Financial OS that brings together global capabilities while respecting the regulatory and operational requirements of each local market. The RPAA registration is another step in putting that approach into practice, as we continue developing infrastructure to support businesses operating across an increasingly interconnected financial landscape.”
 




 

Continue Reading

Trending