On-premises GPU server 'Edgestar'
Liberate yourself from cost and complexity! Build interactive generative AI in a secure environment.
"Edgestar" is an on-premise GPU server that can perform AI inference and learning all in one unit within a secure environment. By balancing the safety of on-premise servers, it enables the utilization of valuable data that may be dormant within the company, allowing for the construction of interactive generative AI in a secure environment and the creation of new value. Additionally, it allows for fine-tuning of LLMs at a scale of 70B without worrying about the HBM or GDDR memory capacity of the GPU cards. aiDAPTIV+ integrates flash memory and DRAM to achieve a larger memory space. 【Features】 ■ Equipped with AI SSD developed by PHISON, capable of executing training up to the 70B class ■ Provides a simultaneous connection environment for 20 clients at less than half the price of a new multi-GPU server ■ Supports a wide range of business applications and can be optimized to meet specific business needs *For more details, please download the PDF or feel free to contact us.
- Company：Forcesteed Robotics
- Price：5 million yen-10 million yen
