TL;DR:
- Asus plans to introduce the AFS Appliance, an on-site AI server, to tackle security and data control issues associated with cloud-based AI systems.
- The AFS Appliance will utilize Nvidia chips and run the Formosa language model, equivalent to OpenAI’s GPT-3.5.
- Pricing for the service is expected to start at $6,000 per month, with higher-powered options available for $10,000 per month.
- The servers will be owned and operated by Asus, offering greater control over data management.
- Initially, the service will be provided to 30 to 50 enterprise customers in Taiwan before expanding globally.
- Asus aims to address concerns about data sensitivity and privacy by installing the AFS Appliance hardware directly in clients’ own facilities.
Main AI News:
Asustek Computer, widely known as Asus, is set to revolutionize the business landscape with its innovative offering of local ChatGPT-style AI servers for office use. In a bid to address concerns surrounding security and data control associated with cloud-based AI systems, Asus is introducing the AFS Appliance, a rental business AI server that will operate on-site. This groundbreaking service, equipped with powerful Nvidia chips and running the cutting-edge AI language model “Formosa,” is poised to rival OpenAI’s GPT-3.5.
In an exclusive interview with Bloomberg, Asus Cloud and TWS President Peter Wu revealed that the company aims to provide the AFS Appliance service at a monthly rate of approximately $6,000. For those seeking top-tier performance, the highest-powered server, built on an Nvidia DGX AI platform, will be available for around $10,000 per month. These servers, leveraging Nvidia’s A100 GPUs, will be owned and operated by Asus, ensuring complete control over data management. Initially, Asus plans to cater to 30 to 50 enterprise customers in Taiwan before expanding its reach globally later in 2023.
Asus’s collaboration with Nvidia to drive enterprise adoption of this groundbreaking technology highlights their shared commitment to propelling the industry forward. Peter Wu emphasized, “Nvidia is a partner with us to accelerate the enterprise adoption of this technology. Before ChatGPT, the enterprises were not aware of why they needed so much computing power.” This strategic partnership aims to provide businesses with the necessary computing resources to fully harness the potential of AI.
At the core of the AFS Appliance lies the “Formosa Foundation Model,” a sophisticated language model developed by TWS, a subsidiary of Asustek. Designed to generate text infused with traditional Chinese semantics, this model empowers users with AI-driven text generation and coding capabilities, akin to the renowned ChatGPT.
While the demand for generative AI services continues to surge, concerns regarding data sensitivity have become increasingly prominent. With many major AI services relying on online data centers, the risk of exposing sensitive information has grown exponentially. In response, Samsung took measures to prohibit its employees from using OpenAI’s ChatGPT after discovering instances of sensitive code being uploaded to the platform. Apple, too, recently banned its employees from utilizing external AI tools that could potentially leak data to competitors.
Recognizing the paramount importance of privacy for businesses, Microsoft announced its intention to launch a privacy-focused version of ChatGPT tailored to large organizations, such as banks and hospitals, in May. This isolated version would address concerns related to data leaks and regulatory compliance. However, Microsoft’s servers would still be located in their data centers. In contrast, Asus takes privacy protection to new heights by installing the AFS Appliance hardware directly within clients’ own facilities, assuaging concerns surrounding data exposure.
Bloomberg highlights that amid the burgeoning demand for AI training chips, companies like Asus seek to secure their share of the market by offering “holistic AI systems” that provide comprehensive AI solutions in a convenient service package. Asus’s established partnership with Nvidia ensures a steady supply of Nvidia’s chips as the AFS Appliance service becomes available, cementing Asus’s position as a formidable player in the AI landscape.
Conclusion:
Asus’s introduction of on-site AI servers through the AFS Appliance marks a significant development in the market. By offering a comprehensive solution to security concerns and data control, Asus provides businesses with greater confidence in leveraging AI technologies. This move aligns with the growing demand for localized AI solutions that prioritize privacy and data protection. Asus’s strategic partnership with Nvidia and its commitment to ensuring a consistent supply of AI training chips position them as a strong contender in the evolving AI landscape.
