Latest news with #AIcomputing
Globe and Mail
28-05-2025
- Business
- Globe and Mail
Bell to open six data centres equipped to power AI models and apps in B.C.
Bell Canada said it will open six data centres in British Columbia equipped to power artificial intelligence models and applications, just a few years after it sold off a group of data facilities. The first of the data centres will come online in June in Kamloops, with the second opening in Merritt by the end of the year. Two more will open in the next two years, with one scheduled for 2026 at Thompson Rivers University. Two others are in the advanced planning stages, according to the company. Altogether, Bell said its data centres will have the capacity for up to 500 megawatts of electricity, a substantial load that has been secured. Upon completion, the facilities will make up the largest AI computing cluster in the country, according to the company. 'There's been a lot of talk about data centres being built, and visions and ambitions,' said Mirko Bibic, chief executive officer of parent company BCE Inc. BCE-T, in an interview. 'We've got the land, we've got the buildings, we've got the chips in there, we've got the power.' The computer processing needs of generative AI, which include applications such as ChatGPT, have created a surge in demand for data centres around the globe. The federal government last year announced a $2-billion spending package over five years to help Canadian companies and researchers access AI computing power, and to back the construction of new data centres. This infrastructure is necessary for Canadian companies, researchers and governments to build and run AI models, which can lead to productivity gains, economic benefits and scientific breakthroughs. Prime Minister Mark Carney highlighted the potential of AI in his recent mandate letter, and this month appointed Evan Solomon as the country's first Minister of AI and Digital Innovation. Having Canadian-owned and operated data centres for AI is also seen as a way to strengthen sovereignty and protect sensitive information. Many Canadian companies, governments and public institutions rely on storage, cloud computing and AI services from U.S. tech companies. As BCE spends on data centres, however, the company is dealing with more than $30-billion in long-term debt. Its stock price has fallen some 35 per cent in the past year and the company cut its dividend by more than half this month. Mr. Bibic said the company will spend a dollar amount in the low nine figures for its data centre rollout at first, but that number does not include all six facilities. He declined to provide further detail. The first few are on the small end in terms of power consumption – two are 7 MW and two more are 26 MW. The other two projects have the total capacity for more than 400 MW, which is substantial for Canada. Large data centre expansions can run into the billions of dollars, based on figures provided by other companies. 'It's a big investment in the scheme of AI in Canada, but we do invest over $3-billion a year,' Mr. Bibic said, referring to the telecom's capital expenditures. The company is not receiving federal funding through the AI infrastructure program, but Mr. Bibic said he is open to applying. 'It'd be wise for us to do,' he said. Telecoms have operated data centres before, in hopes of expanding into cloud computing and other services. But with lots of supply available, companies slashed their prices and struggled to compete. In 2020, BCE announced the sale of 25 of its data-centre facilities to U.S. company Equinix Inc. for just over $1-billion. 'There was no particular differentiation for Bell in that circumstance,' Mr. Bibic said. AI provides a different opportunity, he continued. Not only is there demand, but sovereignty, data protection and cybersecurity are increasingly important issues in Canada. 'You've got AI as a clear winner in terms of massive growth potential now and for the long-term, and we're extremely well-positioned to lean in,' he said, adding the company has a large existing enterprise business and that the revenue potential with AI is significant. BCE is not the only telecom looking to AI for growth. Telus Corp. said in March that it is equipping a facility it owns in Quebec to handle AI, and plans to later do the same in B.C. Alberta's government is trying to attract companies to build AI data centres, emphasizing the province's abundant natural gas reserves for electricity generation. But one of the advantages of B.C. is access to clean hydropower, along with fibre-optic connectivity. 'I won't dismiss the ambition, but what differentiates this is the fact that these are real [facilities], at-scale, backed by Bell and with low-cost sustainable energy,' Mr. Bibic said. Longer term, the company has the potential to expand its AI footprint in Manitoba and Quebec, which also rely on hydropower. The need for data centres depends on the pace of AI adoption, which some analysts say is uncertain. 'A lack of clarity about future demand makes precise investment calculations difficult,' McKinsey & Co. said in a recent report. 'Overinvesting in data centre infrastructure risks stranding assets, while underinvesting means falling behind.' Mr. Bibic said the company is 'methodical' about its expansion. The facility opening in June will be occupied by Groq Inc., a U.S. company that competes with chip maker Nvidia Corp. and offers hardware and software for running AI. Bell also partnered with Thompson Rivers University for the data centre opening next year, allowing students and faculty to train and run AI models. 'We are securing demand and building at the same time,' Mr. Bibic said. 'It's not, 'Build it and they will come.''
Globe and Mail
20-05-2025
- Business
- Globe and Mail
Is AMD Stock a Buy, Sell, or Hold on ZT Systems Sale?
Advanced Micro Devices (AMD) opened in the red this morning after announcing plans of selling ZT Systems to San Jose-headquartered Sanmina (SANM) for $3.0 billion in cash and stock. However, it will retain the data center infrastructure firm's engineering team, comprising a total of 1,200 experts, for about $1.6 billion, a press release confirmed on Tuesday. AMD shares have rallied in recent weeks, partly due to the 90-day trade truce between the U.S. and China. At the time of writing, they're up more than 30% versus their April low. What ZT Divestiture Means for AMD Stock The planned divestiture of ZT Systems that is expected to close in the final quarter of 2025 could help the chipmaker streamline its focus on AI computing and cloud-based solutions. Unloading the data center infrastructure manufacturing company brings $2.25 billion in cash that strengthens AMD's balance sheet as well. Additionally, retaining ZT's engineering team will help AMD 'compete better with Nvidia in the data center GPU market via additional system experience and faster hyperscale deployment times,' as per Christopher Danely – a senior Citi analyst. AMD Shares May Still Find It Hard to Fly Despite seeing the chipmaker's divestiture of ZT as a positive, Citi maintained its 'neutral' rating on AMD shares today. Danely's $100 price target indicates potential downside of more than 10% from here. Investors should note that AMD stands to get hurt as President Donald Trump continues to make it harder for U.S. chipmakers to be in business with China moving forward. On the most recent earnings call, the company's management told investors that export controls could result in up to $1.5 billion in lost revenue this year. Plus, Intel's price cuts could threaten the chipmaker's gross margin as well in 2025, which may also make it even more difficult for AMD stock to extend gains. Should You Invest in AMD at Current Levels? Despite aforementioned challenges, other Wall Street analysts remain constructive on AMD stock for 2025. The consensus rating on Advanced Micro Devices currently sits at ' Strong Buy ' with the mean target of nearly $129 indicating potential upside of as much as 13% from current levels.
Yahoo
12-05-2025
- Business
- Yahoo
Aetina Showcases Full-Stack Edge AI Innovations at COMPUTEX 2025, Driving Generative AI and Intelligent Applications at the Edge
TAIPEI, May 12, 2025 /PRNewswire/ -- Aetina, a leading provider of edge AI solutions, is set to showcase its comprehensive edge AI computing portfolio at COMPUTEX 2025 (May 20th –23rd, Taipei Nangang Exhibition Center, Hall 1, Booth J0118). The exhibit will feature scalable solutions across enterprise, edge and embedded deployments —including the SuperEdge short-depth AI server series, the versatile MegaEdge AI inference/training platforms with advanced GPU expandability, and the compact yet powerful DeviceEdge edge AI computers powered by the NVIDIA Jetson Orin modules. Aetina will also spotlight application-driven deployments developed in close collaboration with ecosystem partners, targeting critical use cases across digital transformation, in-vehicle application, and smart manufacturing. The exhibit aims to accelerate real-world adoption of generative AI, and AI agent technologies at the edge. Empowering Industry Transformation with AI Agents and RAG Technology As industries redefine their businesses with AI to foster digital transformation, the need for on-premise AI computing is more critical than ever. Aetina introduces the SuperEdge AEX-2UA1, a short-depth GPU server built on NVIDIA MGX modular architecture. The system supports two dual-slot GPUs, including the latest NVIDIA RTX PRO 6000 Blackwell Server Edition GPU, offering exceptional memory capacity and compute power ideal for deploying LLMs and VLMs. By integrating Retrieval-Augmented Generation (RAG) technology, the AEX-2UA1 enables precise, domain-specific knowledge generation, significantly improving decision-making across sectors like finance, healthcare, law, and enterprise governance. Aetina also showcases how the AEX-2UA1, paired with NVIDIA L40S GPUs and NVIDIA ACE , enables responsive digital avatars for automated workflows, 24/7 customer service, and remote operations—enhancing both and operational agility. Enterprise-ready Scalable, Power-Efficient On-Prem AI in Collaboration with Qualcomm Technologies, Inc. With AI adoption becoming a top priority for business leaders, they seek solutions offering enterprise-grade performance and scale without draining budgets or compromising data. Aetina's flagship product, MegaEdge AIP-FR68, combined with the Qualcomm® Cloud AI 100 Ultra accelerator cards and the Qualcomm® AI Inference Suite, provide a turnkey solution for generative AI workloads in air-gapped environments that meet these critical needs. This supports applications like voice agents, image generation, Retrieval-Augmented Generation (RAG), and code generation. Also on display is the next-generation MegaEdge AIP-FR68S, featuring a built-in 1600W power supply unit for reduced setup complexity, improved system stability, and simplified deployment even in space-constrained areas. It includes a Gen5 PCIe switch, enabling models twice as large (up to 200 billion parameters) or twice the performance, enhancing accuracy in natural language processing, image recognition, and predictive analytics applications that require context-rich and complex AI models. The MegaEdge AIP-FR68S supports scalability for up to three Qualcomm Cloud AI 100 Ultra cards, empowering enterprises to further expand AI inference performance for evolving needs. Unlocking New Potential for Autonomous and In-Vehicle AI As an NVIDIA Elite Partner, Aetina will showcase its high-performance robotics platform featuring the NVIDIA Jetson AGX Thor. Offering up to 2 PetaFLOPS AI Compute, this platform is built to run advanced generative AI models - like the GR00T foundation model- with ultra low latency, all within an efficient 130W power envelope. This unlocks new possibilities for humanoid and generative physical AI applications across robotics, smart space, industrial, medical, telecommunication, and more. Additionally, Aetina introduces the rugged AIE-VX28 fanless in-vehicle AI system, powered by NVIDIA Jetson AGX Orin. Supporting 8x GMSL2 automotive cameras and 10G Ethernet, it is IP66-rated and E-Mark certified, making it ideal for deployment in logistics fleets, and intelligent transportation systems. In addition, Aetina will present its intelligent driving solution integrating NVIDIA Jetson Orin NX with the Digital Media Professionals Inc. (DMP) AI suite for advanced Driver Monitoring System (DMS) and ADAS functions. It enables real-time detection of driver fatigue, distraction, and gaze deviation, while also identifying external driving hazards, enabling dual-layer AI perception for enhanced road safety and intelligent transportation operations. Accelerating Smart Manufacturing with Advanced AI Vision Aetina is driving the growth of edge AI capabilities across industries through partnerships with leaders in inference acceleration. Together with EdgeCortix, Aetina showcases an edge-deployed visual language model (VLM). Powered by the EdgeCortix SAKURA-II, traditional visual detection systems are transformed into interactive decision-making tools. This generative AI technology enhances video search, scene comprehension, and asset tracking while enabling real-time queries that deliver deeper insights, expedite decisions, and improve safety across security, retail, industrial, and healthcare applications. To further address real-time quality inspection and labor shortages, Aetina partners with Axelera AI and Accordance Systems to deliver a semiconductor inspection and classification solution. The system combines Aetina's Rockchip-based AIE-KR1B-B1-AXE edge AI system with Axelera AI M.2 inference card and Accordance AiPILOT vision inspection technology, delivering high-speed, multi-task visual defect detection with exceptional accuracy. In another co-developed solution with Macnica and PrefactorTech, Aetina presents an advanced AI defect inspection solution integrating Epson robotic arms, NVIDIA Jetson Orin-powered computing platform, and Innodisk industrial cameras. This system delivers high-precision image recognition and real-time processing for PCB inspection, part defect detection, and product positioning with automated handling. The solution enhances production efficiency and quality control while optimizing deployment costs and flexibility for smart manufacturing implementation. A voice-interactive robotic arm demonstration also highlights how conversational AI capability can bring voice-enabled automation to life in smart factory environments. Committed to a "scenario-driven, problem-solving" approach, Aetina continues to empower customers with a robust Edge AI portfolio and industry-proven expertise, enabling smarter operations, lower costs, and faster AI deployment. Welcome visiting Aetina at COMPUTEX Booth #J0118 and join us in shaping an intelligent future powered by Edge AI. View original content to download multimedia: SOURCE Aetina Corporation Error in retrieving data Sign in to access your portfolio Error in retrieving data Error in retrieving data Error in retrieving data Error in retrieving data
