AMD specific news
Helios marks AMD’s biggest AI infrastructure push yetedit
AMD has launched Helios, its first complete open rack-scale AI infrastructure system that integrates next-generation Instinct GPUs, EPYC Venice CPUs, Pensando networking, and the ROCm software stack for large-model training and inference. The platform delivers high memory capacity (about 50% more than Nvidia’s competing rack), open standards connectivity, strong performance metrics, and has already secured an early deployment with Microsoft Azure. While Helios aims to challenge Nvidia’s dominance with better value and flexibility, software maturity around ROCm remains a key hurdle compared with CUDA.
Publication: Network World
AMD Quietly Seeds Zen 6 Client Support into Linux, Tipping Olympic Ridge Desktops and Medusa Mobile Ryzen Chipsedit
AMD has quietly added Zen 6 client CPU support to the Linux HSMP driver, covering Olympic Ridge and Medusa families. Olympic Ridge targets AM5 desktop Ryzen chips with up to 24 cores, 48 threads, and large L3 cache. Medusa Point mobile APUs will offer up to 22 cores for mainstream laptops, a major jump over Strix Point, plus a cost-optimized Medusa 2 variant. These Zen 6 client processors are expected to launch in early 2027, following the server EPYC Venice chips.
Publication: ET Retail
AMD Officially Launches Radeon RX 9050 GPU With 8GB VRAMedit
AMD is making a 4GB graphics card in 2026, it can upset gamersedit
Because of price pressure, leaks indicate AMD is introducing the Radeon RX 9050 graphics card with 4GB of VRAM. The industry has not seen such a card for years because modern games struggle even with 8GB of VRAM. The surprising choice is raising eyebrows.
Publication: India Today
Competition in news
Intel providing chip technology to startup led by co-investor of CEO Lip-Bu Tan in rare dealedit
Lenovo is promoting a hybrid AI approach that combines on-premises and cloud resources to help enterprises cut high token costs associated with AI inference and agentic workloads. The strategy emphasizes running AI closer to data and users for better efficiency, control, and lower expenses—claiming significant savings (up to 8x vs. cloud infrastructure). This positions hybrid models as a practical way for businesses to scale AI economically without relying solely on expensive public cloud or API-based services.
Publication: Economic Times
Humanoid Robot Powered by Qualcomm Collapses on Stage at Computex 2026edit
A Qualcomm-powered humanoid robot collapsed on stage during a live demo at Computex 2026 in Taipei while walking and carrying a gadget. Qualcomm attributed the incident to a brief communication glitch in the prototype, noting it performed a safe collapse to protect people and objects. The viral video drew widespread memes and jokes online, including a laughing reaction from Elon Musk.
Publication: Deccan Chronicle
Tessolve collaborates with Intel Foundry to support EMIB-based advanced package designedit
Tessolve has partnered with Intel Foundry to support package designs using EMIB (Embedded Multi-die Interconnect Bridge) technology for advanced multi-chiplet architectures. The collaboration enables Tessolve to help Intel Foundry customers accelerate development of high-performance semiconductors for AI, HPC, and other applications through end-to-end design and packaging expertise. It aims to reduce costs, lower risks, and speed innovation by combining Tessolve’s engineering capabilities with Intel’s advanced packaging solutions.
Publication: Electronics For You Business
Partner in news
Lenovo encourages hybrid AI model to address token costsedit
Lenovo is promoting a hybrid AI approach that combines on-premises and cloud resources to help enterprises cut high token costs associated with AI inference and agentic workloads. The strategy emphasizes running AI closer to data and users for better efficiency, control, and lower expenses—claiming significant savings (up to 8x vs. cloud infrastructure). This positions hybrid models as a practical way for businesses to scale AI economically without relying solely on expensive public cloud or API-based services.
Publication: The Hindu Business Line
Lenovo encourages hybrid AI model to address token costsedit
Lenovo India plans to address concerns around rising token costs with a streamlined AI model that functions across cloud, local infrastructure and customer devices to drive up efficiency by 6-fold, officials said in an interaction with the media.
Publication: The Hindu Business Line
ASRock Industrial, AMD Partner for AI WAVE SHOW 2026edit
ASRock Industrial partnered with AMD and Taiwan AI Cloud at AI WAVE SHOW 2026 to showcase the AI BOX-A395 edge computing platform. Powered by the AMD Ryzen AI Max+ 395 processor (up to 126 TOPS and 128GB memory), it enables high-performance, secure on-premises AI inference and agents for industries like healthcare and manufacturing. The demo highlights generative AI, data security, and flexible deployments to accelerate enterprise AI adoption and digital transformation.
Publication: Energetica
Industry news
India unveils multi-pronged pipeline for semicon talent pipelineedit
India has launched a multi-pronged strategy under Semicon India and Semicon 2.0 to build a skilled semiconductor workforce across design, fabrication, and packaging. Key initiatives include free EDA tools and training via Chips to Startup (68,000 trained), SMART Lab (over 1 lakh trained), new AICTE curricula, and partnerships with Lam Research, IBM, and universities like Purdue. The approach aims to create industry-ready talent through curriculum upgrades, specialized labs, and global collaborations to support the full semiconductor value chain.
Publication: Communication Today
India leads in workplace AI adoption, but records highest failed AI pilotsedit
India leads global workplace AI adoption, with employees 45% more likely than average to use AI in core workflows. However, it also records the highest rate of failed AI pilots at 38%, versus the global average of 28%. Failures stem mainly from lack of business context, poor personalization, and generic tools, prompting calls for more tailored, trustworthy AI solutions.
Publication: The New Indian Express
Nvidia Graphics Cards Are About to Get Even More Expensiveedit
Samsung has raised DRAM prices by 20% amid the ongoing memory shortage, prompting Nvidia to hike graphics card prices by up to 30%. The increase is expected to hit the entire Blackwell lineup, potentially pushing the RTX 5070 from ~$650 to around $850 and the RTX 5090 toward $6,500. This continues a trend of rising GPU costs driven by the prolonged RAM crisis that shows no signs of easing soon.
Publication: IGN
It’s Cool, Calm, Collectingedit
Environmental fears over water and power consumption are overstated, as new tech like liquid cooling, AI-optimized systems, and captive power sources sharply cut resource use. India can pursue AI advancement without sacrificing sustainability by energy consumption innovations and smart regulations. New tech such as liquid cooling, AI-optimized systems and captive power sources can all sharply cut resource use. Securing data sovereignty is vital for national security and progress.
Publication: The Economic Times
Disclaimer - For AMD Internal Use Only