NVIDIA RTX Spark has arrived, bringing NVIDIA Blackwell RTX graphics, CUDA, and up to 128 GB of unified memory to a new generation of Windows PCs built for local AI, creative work, and personal agents.
Msty is part of the RTX Spark launch.
Our team had early access to RTX Spark hardware. We used it to test Msty Studio, validate compatibility, and prepare support before general availability.
Day-one support starts with Msty Studio.
Meet NVIDIA RTX Spark
Running large AI models locally has traditionally required a capable GPU, enough available memory, and some technical setup. For many people, cloud AI services have therefore been the simpler option, particularly when working with larger models.
NVIDIA RTX Spark brings more of the hardware needed for local AI directly to Windows laptops and compact desktop PCs. The platform combines an NVIDIA Blackwell RTX GPU and NVIDIA Grace CPU, connected through NVIDIA NVLink-C2C, with configurations offering up to 128 GB of unified memory.
That unified memory is especially relevant to local AI. A model and its working context must fit within the memory available to the workload, so higher-memory configurations can give people more options when choosing supported models for on-device use.
RTX Spark also brings NVIDIA CUDA, fifth-generation Tensor Cores, and the broader NVIDIA AI and graphics software stack to these systems. NVIDIA positions the platform for local AI agents, language models, creative applications, development, and gaming.
What RTX Spark Means for Msty
Msty gives people a choice of models and a choice of where to run them. RTX Spark expands the Windows hardware available for local AI.
With early access to RTX Spark hardware, our team tested Msty Studio on the new systems, validated GPU acceleration, and confirmed compatibility with the platform.
In Msty Studio, you can discover, download, and run supported local AI models. GPU offloading is available in each model’s settings.
RTX Spark supplies the hardware. Msty supplies software for selecting models, attaching private knowledge, and running agentic workflows.
Putting RTX Spark to Work with Msty
Our RTX Spark work starts with Msty Studio, where support is available from launch. Studio runs local and cloud AI models in one privacy-first workspace, and its existing tools run on RTX Spark systems.
Run supported models locally
Discover, download, and run supported open-weight models on your computer. Local inference can keep model processing on the device rather than requiring every prompt to be sent to a cloud provider.
Work with your own knowledge
Knowledge Stacks let you bring documents and other reference material into your AI workflows. When a compatible local model is selected, both inference and the supporting data can remain on the device.
Choose the right model for the task
Different models have different strengths. Msty Studio lets you work with local models, compare responses in split chats, and connect to cloud providers when a task calls for them.
Customize how you work
Personas, Prompt Studio, split chats, and other workspace tools give you flexibility in how you interact with models, organize work, and build repeatable workflows.
Stay in control of your data
Msty Studio was built with privacy in mind. You can run supported models locally, work offline, and choose when to use an external AI provider.
RTX Spark also matters beyond Studio. Msty Go is built for agentic work, and NVIDIA designed RTX Spark in part for personal agents running on Windows. We are evaluating where local execution on RTX Spark fits across other Msty products.
Local and Cloud, on Your Terms
Local AI gives people and organizations more choice. Sensitive or proprietary work can remain closer to its source. Developers can experiment without sending every request to a hosted API. Models can remain available when an internet connection is limited or unavailable.
Cloud AI still has an important role. Some workloads benefit from hosted models, specialized services, or infrastructure that is not practical to run locally.
Msty supports both approaches. In Studio, you can run models on your own hardware, connect to cloud providers, or combine the two depending on the work in front of you.
RTX Spark adds a new class of Windows PC to that choice. We look forward to seeing how the Msty community uses it for local models, private knowledge, and the AI workflows that matter to them.
Msty and RTX Spark
We appreciate NVIDIA’s collaboration ahead of the RTX Spark launch. Early access gave our team time to test Msty Studio on the hardware and have support in place from day one.
If you are getting an RTX Spark system, Msty Studio support is available now.