NVIDIA announced that its RTX Spark laptops are now available from ASUS, Dell, HP, Lenovo, Microsoft and MSI, pairing a Blackwell RTX GPU with a Grace CPU in a single superchip offering up to 128GB of unified memory and up to 1 petaflop of FP4 AI performance. The company also introduced Nemotron 3.5 Transcribe, an open weight speech model, new controls for sharing GPU resources with background agents, and gaming updates including Phantom Blade Zero's launch.
NVIDIA said its RTX Spark laptops combine the Blackwell RTX GPU, the Grace CPU and up to 128GB of unified memory on one power-efficient superchip, allowing AI models and creative work to draw from the same memory pool. The systems come in designs as thin as 14mm and as light as three pounds, with battery life NVIDIA describes as all-day, and pricing starts at US$2,599.
The company introduced NVIDIA Nemotron 3.5 Transcribe, an open weight speech model that handles transcription, speaker attribution and translation in a single promptable model, and can answer questions about recorded audio. NVIDIA said it transcribes 40 languages, including seven Indic languages, and ranks No. 1 on the Hugging Face Open ASR leaderboard and No. 2 among open weight models on Artificial Analysis. The model runs approximately 4.3 billion active parameters, about 80% fewer than the leading open weight model, and is available today on Hugging Face. Quantized versions are planned so developers can run it locally on RTX Spark and DGX Spark systems, Jetson AGX Thor devices or scale it on Blackwell data center GPUs.
A new resource management feature in the NVIDIA App lets users choose how GPU resources are shared between a foreground application and background-running AI agents. It is accessed from System Advanced in the app and requires hardware-accelerated GPU scheduling to be enabled in Windows Settings.
NVIDIA also announced day-zero support for LFM3, a family of open weight models from Liquid AI spanning text, vision, audio, omni and decision tasks on one shared backbone. Liquid AI's new 1.7-billion-parameter base model outperforms its previous 2.6-billion-parameter base model, according to the announcement. The models run on RTX Spark with native support through llama.cpp and vLLM and can be fine-tuned locally.
On the gaming side, S-GAME's Phantom Blade Zero launches Oct. 28 at 7 p.m. PT on Steam and the Epic Games Store, with ray-traced effects developed with NVIDIA and DLSS 4.5 Dynamic Multi Frame Generation. RTX Remix 1.6, released last week, adds support for rendering Minecraft Java Edition 1.21.1 with path tracing, and brings DLSS 5 with 3D-Guided Neural Rendering and RTX Spark support to Portal with RTX and the Half-Life 2 RTX demo.
Analysis
Why This Matters
- RTX Spark laptops put desktop-class AI compute in portable machines, letting developers run and fine-tune models locally with a path to more than 150 million RTX AI PCs and NVIDIA GPUs in the cloud.
- An open weight speech model covering 40 languages gives developers control over where audio is processed, from a laptop to the cloud, which matters for fields such as healthcare, financial services and customer service.
- New GPU resource controls address the practical question of how foreground apps and long-running AI agents share a single laptop's compute.
Background
RTX Spark is NVIDIA's line of personal AI computers built around a superchip combining its Blackwell RTX GPU with its Grace CPU, with unified memory so creative work and AI models draw on the same pool. The announcements are part of NVIDIA's broader push to extend its CUDA platform, the same software stack used for AI in the cloud, onto local machines.
What to Watch
- Quantized versions of Nemotron 3.5 Transcribe, which will enable local deployment across RTX Spark, DGX Spark and Jetson AGX Thor systems.
- Whether LFM3's day-zero support translates into real local agent workflows such as inbox triage, document reading and support ticket routing.
- The Phantom Blade Zero launch on Steam and the Epic Games Store, and whether RTX Remix 1.6's Minecraft support broadens the modding community.
Checked
12 claims in this story checked against 1 source: 11 supported.
Show each claim
- SupportedNVIDIA is expanding RTX Spark laptops with a new open speech model and gaming updates.NVIDIA Blog: “RTX Spark laptops are now available from ASUS, Dell, HP, Lenovo, Microsoft and MSI. ... New and expected RTX Spark updates, coming soon, include: NVIDIA Nemotron 3.5 Transcribe ... Phantom Blade Zero”
- SupportedRTX Spark laptops are available from ASUS, Dell, HP, Lenovo, Microsoft and MSI.NVIDIA Blog: “RTX Spark laptops are now available from ASUS, Dell, HP, Lenovo, Microsoft and MSI.”
- SupportedThe laptops combine a Blackwell RTX GPU and Grace CPU in a superchip with up to 128GB of unified memory and up to 1 petaflop of FP4 AI performance.NVIDIA Blog: “These powerful systems combine an NVIDIA Blackwell RTX GPU and an NVIDIA Grace CPU in a single superchip with up to 128GB of unified memory and up to 1 petaflop of FP4 AI performance.”
- SupportedNVIDIA introduced Nemotron 3.5 Transcribe, an open-weight model for transcription, speaker attribution and translation that can also answer questions about audio.NVIDIA Blog: “Introduced today, NVIDIA Nemotron 3.5 Transcribe is a new, open weight speech model ... one promptable model that handles all three, plus questions about the audio.”
- SupportedNemotron 3.5 Transcribe handles 40 languages, including seven Indic languages, and ranks No. 1 on Hugging Face Open ASR and No. 2 among open-weight models on Artificial Analysis.NVIDIA Blog: “It transcribes 40 languages, including seven Indic languages. ... Nemotron 3.5 Transcribe ranks No. 1 on the Hugging Face Open ASR and No. 2 among open weight models on Artificial Analysis”
- SupportedThe model has approximately 4.3 billion active parameters, about 80% fewer than the leading open-weight model.NVIDIA Blog: “approximately 4.3 billion active parameters, 80% fewer than the No. 1 open weight model”
- Partly supportedRTX Spark laptops start at US$2,599.NVIDIA Blog: “RTX Spark laptops are available now from ASUS, Dell, HP, Lenovo, Microsoft and MSI, starting at $2,599”
- SupportedThe laptops are as thin as 14mm and as light as three pounds, with all-day battery life.NVIDIA Blog: “RTX Spark laptops are as thin as 14mm and as light as three pounds, with all-day battery life.”
- SupportedThe NVIDIA App lets users choose how GPU resources are shared between a foreground application and background-running agents, and the feature requires hardware-accelerated GPU scheduling in Windows Settings.NVIDIA Blog: “With the new resource management feature in NVIDIA App, users can now choose how GPU resources are shared. ... This feature requires hardware-accelerated GPU scheduling to be enabled in Windows Settings.”
- SupportedLFM3 is a family of open-weight models for text, vision, audio, omni and decision tasks built on one shared backbone; its new 1.7-billion-parameter base model outperforms its previous 2.6-billion-parameter base model.NVIDIA Blog: “The family spans text, vision, audio, omni and decision models built on one shared backbone, with base and post-trained models available as open weights. ... new 1.7-billion-parameter base model outperforms its previous 2.6-billion-parameter base model.”
- SupportedPhantom Blade Zero launches Oct. 28 at 7 p.m. PT on Steam and the Epic Games Store, with ray-traced effects and DLSS 4.5 Dynamic Multi Frame Generation.NVIDIA Blog: “S-GAME’s Phantom Blade Zero launches tonight, Oct. 28, at 7 p.m. PT on Steam and the Epic Games Store. ... NVIDIA worked with S-GAME Studio and S-GAME Publishing to add ray-traced effects to the title. The NVIDIA DLSS Dynamic Multi Frame Generation feature”
- SupportedRTX Remix 1.6 supports Minecraft Java Edition 1.21.1 with path tracing and brings DLSS 5 with 3D-Guided Neural Rendering and RTX Spark support to Portal with RTX and the Half-Life 2 RTX demo.NVIDIA Blog: “Both creators and players can render Minecraft Java Edition 1.21.1 through RTX Remix with DLSS-accelerated path tracing ... Starting today, those upgrades come to Portal with RTX and the Half-Life 2 RTX demo. Both now support RTX Spark, with DLSS 3D-Guided Neural Rendering”
