Inflection AI has partnered with Intel to launch a new enterprise artificial intelligence system called Inflection for Enterprise. This system is powered by Intel’s Gaudi 3 AI accelerator and Tiber AI Cloud platform, and is set to be fully operational in the first quarter of 2025.

Currently functioning as a cloud-based service, the system aims to provide a robust solution for businesses working with large language models (LLMs). Inflection for Enterprise offers customized AI models tailored to the unique data and needs of each organization, delivering specialized solutions for companies.

At the core of the system is the Inflection 3.0 model, which comes in two different versions: one optimized for chatbots and the other capable of following more complex user instructions. Additionally, the second version supports JSON data format, facilitating the integration of model outputs into applications.

This initiative from Inflection AI enables businesses to maintain control over their AI solutions. The AI models, customized based on the companies’ data, provide more efficient and secure process management, while ownership of these models remains entirely with the company.

Sean White, CEO of Inflection AI, stated, “Inflection for Enterprise is the only system that allows businesses to have complete control over their intelligence. You manage your data, your finely-tuned model, and your architecture. Whether you run it in the cloud, in a hybrid environment, or on-premises is up to you.”

Intel’s Gaudi 3 accelerator provides three times the number of AI cores compared to its predecessor, delivering high performance. Moreover, the new Ethernet module enhances the speed of data sharing, offering businesses a cost advantage that allows AI applications to operate more efficiently. Thanks to the infrastructure provided by Intel, Inflection AI’s software can operate at twice the cost efficiency compared to some competing processors.


Source