AI News Feed
Market watch
Products & Applications

Lenovo YOGA Pro 15 RTX Spark Opens Blind Preorders With Nvidia N1X Superchip

Lenovo opened blind preorders for the YOGA Pro 15 RTX Spark on Oct. 8. The laptop is among the first local-agent AI PCs powered by Nvidia’s RTX Spark N1X superchip and targets AI creators, developers and researchers.

The RTX Spark N1X superchip integrates Nvidia AI and RTX graphics into a thin and powerful Windows PC. QbitAI reported the chip includes up to a 6,144-core Nvidia Blackwell RTX GPU and up to a 20-core Nvidia Grace CPU, delivering up to 1 Petaflop FP4 of local AI compute. Nvidia’s CUDA platform, the global standard for AI acceleration software, runs natively on the device, supporting frontier AI tools and model development.

The YOGA Pro 15 RTX Spark offers up to 128GB unified memory and up to 2TB high-speed storage. The superchip combines GPU, CPU, and unified memory, which Lenovo says breaks the memory wall that has constrained traditional on-device AI. It allows very large parameter models to stay resident locally and supports complex multitasking. With 80W full-power tuning and energy-efficiency claims, the machine is designed to sustain output during large-model inference, high-resolution video rendering, and compilation of large code projects.

The laptop runs Lenovo’s Tianxi AI, described as a personal super intelligent agent. Its new Super Mode is designed to handle complex tasks on the YOGA Pro 15 RTX Spark through private local workflows without network or token limits. It can work with multiple productivity tools and generate videos, web pages, and documents, according to the report. Users can switch between local and standard modes, using local mode for private tasks.

Tianxi AI Super Mode has been optimized for the device. Relying on RTX Spark, Lenovo says the agent can provide uninterrupted AI operation and independently plan, execute, and deliver editable, interactive results. The laptop supports local operation of models with more than 100 billion parameters and ships with a 35B local model ready out of the box, enabling millisecond-level local responses without complex configuration.

Lenovo lists three core advantages: offline usability with zero token cost, privacy and asset security, and autonomous agent collaboration. Because on-device AI runs locally, it can still dispatch large models when disconnected, avoiding cloud API token consumption and subscription fees. Data remains local, reducing the risk of cloud leaks. With Tianxi AI, the system moves from passive question-and-answer commands to AI-to-AI agent collaboration and self-executed tasks.

For AI creation, development, and research, Lenovo says the device reshapes workflows. Local script generation, text-to-image, audio and video materials, 3D modeling and UI design can proceed without network delays. Offline on-device AI supports private code repository review, intelligent code completion, refactoring, and debugging. For researchers, Tianxi AI can help with paper reading and summarization, local literature analysis, and local cleaning and modeling of large research datasets.

The hardware includes a 15.3-inch 2.5K OLED 165Hz touchscreen with high color accuracy. It supports multimodal interaction through a full-area writing pad and dedicated stylus ecosystem, described as allowing users to write, draw, tap, and speak to AI. Lenovo says the device natively supports more than a thousand professional creation applications and is deeply compatible with Nvidia Studio and RTX acceleration for 3D rendering, video editing, and AI creation workflows.

The laptop comes in a restrained Thunder Gray color, with a thin and light body for mobile work. It has a 92.5Wh high-density battery intended to reduce battery anxiety for outdoor and mobile workstations. Lenovo says the YOGA Pro 15 RTX Spark marks a shift from cloud-dependent tools to autonomous local intelligent agents. The blind preorder channel opened at 9:00 on October 8. The article was provided by Lenovo and republished by QbitAI with authorization; views belong to the original author.