Nvidia RTX Spark

Headlines · 6

1. 6h
   Googlebook laptops may repeat Steam Machine's pricing mistake

   Google has launched the Googlebook laptop platform, described as an AI-focused reimagining of Chromebook with a 16GB RAM requirement and prices starting at $999, apart from a limited-time $899 Acer offer, with Lenovo and Asus models reaching MacBook Air territory. The platform's Android phone integration requires a device running Android 17, an OS released only four months ago and so far limited to a handful of expensive phones. In this opinion piece, Tom's Hardware argues the timing is poor given the author's claim that RAM costs rose roughly five-fold in a year and PC shipments are down 20%, drawing a parallel to Valve's Steam Machine, which was delayed about six months and priced above $1,000; it notes Nvidia's AI-coder-focused RTX Spark laptops start at $2,599 and run up to $6,999.

   Tom’s Hardware · 🔥 8
2. 1d
   Adobe and NVIDIA expand partnership to speed agentic creation

   Adobe and NVIDIA announced an expansion of their partnership, bringing NVIDIA accelerated computing and open models into Adobe's flagship creative tools. Photoshop, Premiere and Substance 3D have been optimized for the NVIDIA RTX Spark superchip launching Oct. 16, which Adobe says delivers up to 2x faster AI processing, editing, color and effects. In the cloud, Frame.io uses NVIDIA NVDEC hardware decoding and TensorRT; early results on selected transcode and search workloads show roughly 5x faster embedding generation, a 4-5x improvement in P95 processing latency and an approximately 80% smaller search proxy. Adobe Brand Concierge has also been optimized with NVIDIA Nemotron open models and will reach enterprise customers later this year, and NVIDIA has selected Brand Concierge to power personalized conversational experiences on NVIDIA.com.

   NVIDIA Newsroom · 🔥 5
3. 1d
   Nvidia RTX Spark laptops launch, built to run AI agents locally

   Nvidia RTX Spark laptops went on sale on Oct. 16 from ASUS, Dell, HP, Lenovo, Microsoft and MSI, built around the GB10 superchip that pairs an Arm-based CPU with a Blackwell GPU, up to 128GB of unified memory and up to 1 petaflop of FP4 AI compute for running large language models and AI agents on the device itself. Nvidia also released Nemotron 3.5 Bolt 74B-A9B, an open-weight mixture-of-experts model (74B parameters, 9B active, packed into 24GB, supported by llama.cpp, Ollama, LM Studio, Unsloth and vLLM) that scores 21 on the Artificial Analysis Intelligence Index versus 58 for Claude Opus 5.5, plus the open-source Switchyard model router for splitting work between local and cloud models and the PAIR tool for serving models to other devices on a home network. Windows officially supports Hermes, Perplexity and OpenClaw (GitHub Copilot is also mentioned) with two-click installation from the store, while Copilot Plus features such as Recall are absent at launch and are expected for Windows Insiders later this year.

   The Verge AI · 🔥 33
4. 1d
   Llama for Windows released: one-click local model downloads

   Llama for Windows, the Windows counterpart to the macOS Llama app, is a free, open-source Windows 11 application built on Microsoft's WinUI 3 that lives in the system notification area and opens its control panel from a taskbar icon. It connects directly to the Hugging Face model hub, shows a curated model list with parameter counts, disk usage and vision support, and downloads GGUF-quantized models driven by the llama.cpp ecosystem, all without admin rights, manual compilation or complex launch parameters. The app offers an Alt+Space floating chat window, a built-in web UI for loading and unloading local models, and can start an OpenAI-compatible local server at localhost:9931 that scripts and AI agents can call, with multiple apps able to share one model file. According to the project, prompts and inference stay on the device and no telemetry, analytics events or installation identifiers are collected, though network access is still needed to fetch models and updates; it supports Windows 11 x64 and ARM64 plus NVIDIA's RTX Spark platform, and does not list Windows 10 as supported.

   cnBeta · 🔥 8
5. 4d
   Nadella pitches Windows PC as 'infinite software factory' with local MAI-Code-1.1 Flash

   After Microsoft's October 7 Windows and Surface event, CEO Satya Nadella said on X that Copilot Code plus on-device models let users build the software they need on the desktop with no cloud token spend, declaring "your PC becomes an infinite software factory." The engine is Microsoft's in-house MAI-Code-1.1 Flash, a 137-billion-parameter coding model with a 256K context window that the post says is now optimized to run on a PC; the on-device version is quantized to about 53GB and peaked at 75.5GB of memory at full context on a Surface Laptop Ultra using NVIDIA RTX Spark. Yet Microsoft's own engineering documentation notes that GitHub Copilot's automatic mode decides task by task whether work runs locally or in the cloud, with one diagram stating that "local inference does not make the session offline" — routing ships at the end of October 2026, hybrid Copilot features reach Copilot+ PCs over coming months, and Code rolls out first via the Frontier early-access program.

   TechRadar · 🔥 0
6. 6d
   Mechrevo to launch RTX Spark mini PC Master Road X10 with AI SSD

   Mechrevo plans to launch a mini PC called Master Road X10 based on NVIDIA's RTX Spark superchip, according to tipster @巨炮快评; the product also appeared in a slide from a September 2026 NetEase Future Conference talk by Mechrevo founder Pan Chunjie. The report says the machine will carry Mechrevo's self-developed AI SSD, which offloads very large local models from DRAM to NAND and cuts redundant computation and loading by saving and reusing processed context and optimizing model loading. Mechrevo has not officially announced the product, and no specs, pricing or launch date were disclosed.

   IT之家 AI · 🔥 0

Experience and discussion from the community

Share my Nvidia RTX Spark experienceAsk about Nvidia RTX Spark