NVIDIA RTX Spark computers on October 16 While putting it up for sale, it added Istanbul to its global promotional events. The company’s October 15 announcement explains that Vatan Bilgisayar is planning local AI, gaming and content creation demonstrations at its Topkapi store. NVIDIA does not require advance registration for this meeting. However, it does not disclose the prices or store stocks of all models in Türkiye in the same announcement.
The Istanbul event will take place in the Topkapı store at Mevlana Street, number 138. NVIDIA states that visitors can also follow the program about local content producers. On the other hand, the company also has events with different participation conditions in San Francisco, Berlin, Taipei, Beijing and Seoul on the same day.
There is also a separate preparation process on the Windows side of the platform. In his statement to The Verge, Microsoft program manager Peter Dawoud explains that the new startup experience will reduce the process of finding and installing AI agents to two clicks. This experience will offer the user agent options during the first installation of the RTX Spark computer.
In addition, Microsoft supports installing and updating agents via the store. Thus, users do not have to search for dependencies and update packages for each agent separately. However, the agents offered in the store do not automatically gain unlimited access to every file or account on the computer. NVIDIA’s product page emphasizes that access permissions will remain up to the user via Windows controls.
Grace processor and Blackwell GPU use shared memory
NVIDIA’s technical product page shows a 20-core Grace CPU and a 6,144-core Blackwell RTX GPU in the top option of the RTX Spark. The platform supports up to 128GB of unified memory. Thus, the processor and graphics unit use the same capacity instead of separate memory pools. This approach aims to open up space for large local AI models.
The company also places CUDA support at the center of the platform. Developers can run their models on the computer with tools such as CUDA and TensorRT. In addition, content producers benefit from hardware encoding and decoding support for H.264, HEVC and AV1. NVIDIA also states on the product page the possibility of processing three-dimensional scenes up to 90 GB with combined memory.
Initial devices include ASUS ProArt P16, Dell XPS 16, HP OmniBook Ultra 16, Lenovo Pro 9n, Microsoft Surface Laptop Ultra and MSI Prestige N16 Flip AI+. RTX Spark is also coming to small desktops. The maximum memory and core values on the page do not indicate that all devices will be sold with the same hardware. Each manufacturer determines the memory, screen and other options of its own model separately.