logo
Nvidia and MediaTek rumoured to announce new Windows on Arm laptop chips at Computex 2025

Nvidia and MediaTek rumoured to announce new Windows on Arm laptop chips at Computex 2025

Yahoo09-05-2025

When you buy through links on our articles, Future and its syndication partners may earn a commission.
At the end of last year, I listed a possible Nvidia laptop chip (ie, not just the GPU but the CPU, too) as something I'd really like to see from 2025, and since then we've had plenty of intimations that we might actually see such a thing. Now, we're hearing this might be announced at Computex later this month.
According to ComputerBase (machine translated), "the official unveiling [of an Nvidia PC chip] is now expected at Computex 2025." This wouldn't be a solo venture, though, as it's expected to be made in partnership with MediaTek, based on the Nvidia and MediaTek GB10 chip that sits inside DGX Spark home-user supercomputer (previously called Project Digits).
If it is based on GB10, that would presumably mean a processor using its Arm-designed CPU cores rather than custom Nvidia-designed ones. This wouldn't be a massive surprise, but there was previously at least some reason to believe that it could be using custom-designed chips, given previous rumours and that Nvidia recently announced a custom-designed Arm chip called Vera to succeed its previous Arm-designed Grace chips.
Whether it's custom or Arm-designed, it will be an Arm chip, meaning much will rely on Windows on Arm (WoA). We've seen Arm chips using WoA in the wild over the past year in the form of Qualcomm's Snapdragon X Elite chips that sit inside some laptops. These have great battery life, but there are still some problems with Windows on Arm, especially for gaming. That's mainly thanks to Prism emulation that still struggles with some games.
WoA does seem to be steadily improving, though, and one can only assume Nvidia would have a better time of getting gaming to work than Qualcomm, given the company is already king of the Windows graphics space. Still, there's not a lick that Nvidia can do about Prism emulation itself, so it won't be entirely in the chip giant's hands.
To be clear, that an Nvidia PC chip is in the works isn't all based on rumours: At the start of this year, Nvidia officially confirmed it is working on an end-user system on a chip (SoC) based on DGX Spark.
Now, both Nvidia and MediaTek will be delivering keynotes at Computex on May 19 and 20, respectively, so it makes sense that both companies would use this as a chance to announce an upcoming collaborative consumer SoC.
ComputerBase points out that, according to DigiTimes, MediaTek has booked additional capacity for semiconductor assembly and testing with ASE, implying that this is because of the upcoming laptop chips. Without seeing what this additional capacity is officially booked for, however, we can't say for certain.
We also can't say for certain how such chips will perform. Previous rumours had it that the upcoming SoC should have RTX 4070 mobile and Strix Halo-level performance, and ComputerBase speculates that instead of the 20 CPU cores in the GB10, the laptop chips will likely have "perhaps only eight to twelve, and the RAM [is] likely to be a quarter of that or even less, i.e. 32 or 16 GB—depending on which market segment is ultimately targeted."
Whatever the case, it's certainly an exciting prospect that we could have an Nvidia chip sitting inside some laptops, which would presumably be low-power ones given Arm chips tend to be very efficient. And that they might give us all the benefits of the Blackwell GPU architecture for gaming. And we might not have too long to find out. Roll on Computex.
Best CPU for gaming: Top chips from Intel and AMD.Best gaming motherboard: The right boards.Best graphics card: Your perfect pixel-pusher awaits.Best SSD for gaming: Get into the game first.

Orange background

Try Our AI Features

Explore what Daily8 AI can do for you:

Comments

No comments yet...

Related Articles

Spectro Cloud Integrates Palette with NVIDIA DOCA and NVIDIA AI Enterprise, Empowering Seamless AI Deployment Across Telco, Enterprise, and Edge
Spectro Cloud Integrates Palette with NVIDIA DOCA and NVIDIA AI Enterprise, Empowering Seamless AI Deployment Across Telco, Enterprise, and Edge

Business Wire

time37 minutes ago

  • Business Wire

Spectro Cloud Integrates Palette with NVIDIA DOCA and NVIDIA AI Enterprise, Empowering Seamless AI Deployment Across Telco, Enterprise, and Edge

SAN JOSE, Calif.--(BUSINESS WIRE)--Spectro Cloud, a leading provider of Kubernetes management solutions, today announced the integration of NVIDIA DOCA Platform Framework (DPF), part of NVIDIA's latest DOCA 3.0 and NVIDIA AI Enterprise software, into its Palette platform. Building on its proven track record as a trusted partner for major organizations deploying Kubernetes in the cloud, at the data center, and at the edge, Spectro Cloud continues to expand its leadership in enabling production-ready infrastructure for AI and modern applications. This integration empowers organizations to efficiently deploy and manage NVIDIA BlueField-3 DPUs alongside AI workloads across diverse environments, including telco, enterprise, and edge. Spectro Cloud is excited to meet, discuss, and demonstrate this integration at GTC Paris, June 11-12. With the integration of DPF, Palette users gain access to a suite of advanced features designed to optimize data center operations: Comprehensive provisioning and lifecycle management: Palette streamlines the deployment and management of NVIDIA BlueField-accelerated infrastructure, ensuring seamless operations across various environments. Enhanced security service deployment: With the integration of NVIDIA DOCA Argus, customers can elevate cybersecurity capabilities, providing real-time threat detection for AI workloads. DOCA Argus operates autonomously on NVIDIA BlueField, enabling runtime threat detection, agentless operation, and seamless integration into existing enterprise security platforms. Support for Advanced DOCA Networking Features: Palette now supports deployment of DOCA FLOW features, including ACL pipe, LPM pipe, CT pipe, ordered list pipe, external send queue (SQ), and pipe resize, enabling more granular control over data traffic and improved network efficiency. NVIDIA AI Enterprise-ready deployments with Palette Palette now supports NVIDIA AI Enterprise-ready deployments, streamlining how organizations operationalize AI across their infrastructure stack. With deep integration of NVIDIA AI Enterprise software components, Palette provides a turnkey experience to provision, manage, and scale AI workloads, including: NVIDIA GPU Operator Automates the provisioning, health monitoring, and lifecycle management of GPU resources in Kubernetes environments, reducing the operational burden of running GPU-intensive AI/ML workloads. NVIDIA Network Operator Delivers accelerated network performance using DOCA infrastructure. It enables low-latency, high-throughput communication critical for distributed AI inference and training workloads. NVIDIA NIM Microservices Palette simplifies the deployment of NVIDIA NIM microservices, a new class of optimized, containerized inference APIs that allow organizations to instantly serve popular foundation models, including LLMs, vision models, and ASR pipelines. With Palette, users can launch NIM endpoints on GPU-accelerated infrastructure with policy-based governance, lifecycle management, and integration into CI/CD pipelines — enabling rapid experimentation and production scaling of AI applications. NVIDIA NeMo With Palette's industry-leading declarative management, platform teams can easily define reusable cluster configurations that includes everything from NVIDIA NeMo microservices to build, customize, evaluate and guardrail LLMs; to GPU drivers and NVIDIA CUDA libraries; to the NVIDIA Dynamo Inference framework; plus PyTorch/TensorFlow, and Helm chart deployments. This approach enables a scalable, repeatable, and operationally efficient foundation for AI workloads. By integrating these components, Palette empowers teams to rapidly build, test, and deploy AI services, while maintaining enterprise-grade control and visibility. This eliminates the traditional friction of managing disparate software stacks, GPU configurations, and AI model serving infrastructure. "Integrating NVIDIA DPF into our Palette platform marks a significant step forward in delivering scalable and efficient AI infrastructure solutions," said Saad Malik, CTO and co-founder, Spectro Cloud. "Our customers can now harness the full potential of NVIDIA BlueField's latest advancements to drive accelerated networking, infrastructure optimization, AI security, and innovation across telco, enterprise, and edge environments." 'Organizations are rapidly building AI factories and need intelligent, easy-to-use infrastructure solutions to power their transformation,' said Dror Goldenberg, senior vice president of Networking Software at NVIDIA. 'Building on the DOCA Platform Framework, the Palette platform enables enterprises and telcos to deploy and operate BlueField-accelerated AI infrastructure with greater speed and efficiency.' This strategic integration positions Palette as a comprehensive platform for organizations aiming to operationalize AI at scale, including: Telco solutions: High-performance, low-latency infrastructure tailored for telecommunications applications. Enterprise deployments: Scalable and secure AI infrastructure to support diverse enterprise workloads. Edge computing: Lightweight, GPU-accelerated solutions designed for resource-constrained edge environments. Palette is available today for deployment and proof of concept (POC) projects. For more information about Spectro Cloud's Palette platform, visit Learn more about our work with NVIDIA, including technical blogs, here. About Spectro Cloud Spectro Cloud delivers simplicity and control to organizations running Kubernetes at any scale. With its Palette platform, Spectro Cloud empowers businesses to deploy, manage, and scale Kubernetes clusters effortlessly — from edge to data center to cloud — while maintaining the freedom to build their perfect stack. Trusted by leading organizations worldwide, Spectro Cloud transforms Kubernetes complexity into elegant, scalable solutions, enabling customers to master their cloud-native journey with confidence. Spectro Cloud is a Gartner Cool Vendor, CRN Tech Innovator, and a 'leader' and 'outperformer' in GigaOm's 2025 Radars for Kubernetes for Edge Computing, and Managed Kubernetes. Co-founded in 2019 by CEO Tenry Fu, Vice President of Engineering Gautam Joshi and Chief Technology Officer Saad Malik, Spectro Cloud is backed by Alter Venture Partners, Boldstart Ventures, Firebolt Ventures, Growth Equity at Goldman Sachs Alternatives, NEC and Translink Orchestrating Future Fund, Qualcomm Ventures, Sierra Ventures, Stripes, T-Mobile Ventures, TSG and WestWave Capital. For more information, visit or follow @spectrocloudinc and @spectrocloudgov on X.

Nvidia Teams with HPE for New Supercomputer, Debuts Climate AI Model
Nvidia Teams with HPE for New Supercomputer, Debuts Climate AI Model

Yahoo

timean hour ago

  • Yahoo

Nvidia Teams with HPE for New Supercomputer, Debuts Climate AI Model

Nvidia (NASDAQ:NVDA) is building a new supercomputer in Germany with Hewlett Packard Enterprise (NYSE:HPE) and the Leibniz Supercomputing Centre. The system, called Blue Lion, will be built using HPE's next-generation Cray technology and Nvidia's powerful GPUs. Warning! GuruFocus has detected 4 Warning Signs with NVDA. What's under the hood? Blue Lion will run on Nvidia's Vera Rubin architecture a combo of its new Rubin GPU and its first-ever custom CPU, called Vera. According to Nvidia, the system will be roughly 30 times more powerful than Germany's current flagship machine, SuperMUC-NG. The supercomputer is set to go live for researchers across Europe in early 2027, as first reported by Reuters. Meanwhile, Nvidia has also rolled out an AI foundation model built to simulate Earth's climate at kilometer-level precision. Named Climate in a Bottle or cBottle, the tool is part of Nvidia's Earth-2 platform, which blends AI, GPU acceleration, and physics-based modeling. The idea is to generate highly detailed atmospheric data faster and more energy-efficiently than traditional simulation methods and without giving up on accuracy. cBottle can take into account time of day, season, and sea surface temperatures to create realistic atmospheric models. Institutions like the Max Planck Institute for Meteorology and the Allen Institute for AI are already testing how cBottle can help compress climate data and make Earth's complex systems easier to explore. The launch reflects Nvidia's broader push into science-focused applications for AI and high-performance computing, beyond the data center and enterprise cloud. This article first appeared on GuruFocus. 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

Nvidia, HPE Unveil 'Blue Lion' Supercomputer Using Next-Gen AI Chips
Nvidia, HPE Unveil 'Blue Lion' Supercomputer Using Next-Gen AI Chips

Yahoo

timean hour ago

  • Yahoo

Nvidia, HPE Unveil 'Blue Lion' Supercomputer Using Next-Gen AI Chips

June 10 - Nvidia (NASDAQ:NVDA) and Hewlett-Packard Enterprise (NYSE:HPE) will build a new supercomputer for Germany's Leibniz Supercomputing Centre. The system, named Blue Lion, will use Nvidia's upcoming "Vera Rubin" chips. Warning! GuruFocus has detected 4 Warning Signs with NVDA. Scientists expect to gain access to Blue Lion in early 2027. The companies revealed the project at a supercomputing conference in Hamburg. This follows Nvidia's earlier news that the U.S. Lawrence Berkeley National Lab also plans a Vera Rubin-based system for next year. Separately, Nvidia noted the "Jupiter" supercomputer, powered by its technology at Germany's Forschungszentrum Julich, is now ranked as Europe's fastest system. The projects highlight European efforts to remain competitive with the U.S. in supercomputing. These powerful machines are vital tools for research in areas like biotechnology and climate science. Based on the one year price targets offered by 52 analysts, the average target price for NVIDIA Corp is $174.71 with a high estimate of $372.87 and a low estimate of $100.00. The average target implies a upside of +22.49% from the current price of $142.63. Based on GuruFocus estimates, the estimated GF Value for NVIDIA Corp in one year is $273.40, suggesting a upside of +91.68% from the current price of $142.63. This article first appeared on GuruFocus. 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

DOWNLOAD THE APP

Get Started Now: Download the App

Ready to dive into the world of global news and events? Download our app today from your preferred app store and start exploring.
app-storeplay-store