The long wait is over – NVIDIA DGX Spark is finally officially here!
Appincloud limited, as an NVIDIA Inception member, has developed the first APP of enterprise-level applications on DGX Spark – the “AppinCloud Assistant” Spark Edition.
Adaptation is no small matter.
The DGX Spark, as NVIDIA’s first AI personal desktop aimed at AI developers, is equipped with the new Grace processor based on the ARM architecture. Furthermore, the new Blackwell chip focuses on optimizing FP4 precision. It has 128GB Unified Memory + 1 PetaFLOP parallel throughput, how should we use these capabilities?
We have successfully run a complete enterprise-grade intelligent agent on the DGX Spark – the AppinCloud Assistant Knowledge Base. Like most knowledge base products on the market, we use the gpt-oss/Qwen3 foundational large language model.
To enable the agent to “see,” we also built in the Qwen VL model!
Simply put all of these into our DGX Spark All-in-One Machine, now DGX Spark comes with built-in “eyes” and a “brain.” It can understand the images you provide, communicate about the content within those images using natural language, answer your questions, and even perform reasoning, simultaneously!
Coupled with the ASR/TTS voice modules, you can communicate with AppinCloud Assistant directly by voice.
We’ve not only made it reading and writing, Also we’ve even given AI a “persona” – with a built-in digital human.(Experemetal features, need to reach to us)
This isn’t just a simple animated character, but a personified interaction interface integrating multiple technologies.
For more information or wanna to a free trial? please send me email: [email protected]