Home > News content

Asus plans Nvidia chip-based localized AI service to 'keep data secure'

via:IT之家     time:2023/5/30 20:15:04     readed:113

Asustek has unveiled an AI service called AFS Appliance, based on an Nvidia chip that will allow business users to take control of their data while exploring the potential of generative AI, according to multiple media reports.

What makes Asus's new service special is that all the hardware will be installed in corporate customers' own devices, ensuring data security and control. The AI computing platform, built on Nvidia chip technology, will be operated and updated by Asus.

One of the well-known concerns with AI services like OpenAI is that they are served through online data and have the potential to expose sensitive information. For example, Samsung banned employees from using OpenAI's ChatGPT platform after they were found to have uploaded sensitive code to the platform.

Image source Pexels

Asustek will preload the AFS hardware with its own large language model, Formosa. The model is said to be equivalent to ChatGPT 3.5 and is mainly trained in traditional Chinese.

The company plans to price the service at around $6,000 a month, said Wu Hanzhang, general manager of Asustek Cloud and Taizhiyun, in an interview. The highest-profile solution will use the DGX AI supercomputer platform and will cost around $10,000 per month.

He expects the service to start with 30 to 50 corporate clients in Taiwan and expand internationally by the end of the year.

Taijiyun previously released its large-scale language model Formosa in the middle of this month, and announced the launch of a full range of enterprise-level generated AI solutions, providing AI Foundry Service (AFS) series of services that rapidly mobilize AI supercomputer computing power, easily optimize models and flexibly deploy models.

"Nvidia is our partner and will accelerate the adoption of this technology by companies," says Mr Ng. "Before ChatGPT, companies didn't know why they needed so much computing power. Due to the sensitivity of accessing their proprietary data, most enterprises will deploy generative AI on their premises. And because it involves the most sensitive data, they need a smart head under their control and management."

translate engine: Youdao

China IT News APP

Download China IT News APP

Please rate this news

The average score will be displayed after you score.

Post comment

Do not see clearly? Click for a new code.

User comments