Inflection AI, Intel Launch Enterprise AI System

Intel Corp. and machine intelligence company Inflection AI Inc. have announced a collaboration to accelerate the adoption and impact of artificial intelligence for enterprises and developers.

Intel Gaudi and Intel Tiber AI Cloud will support Inflection for Enterprise, an enterprise-grade AI system that Inflection AI will launch to deliver conversational, employee-centric AI capabilities. It will also provide the controls, customization, and scalability required for large-scale deployments.

“With support for open-source models, tools, and competitive performance per watt, Intel Gaudi 3 solutions make deploying gen AI accessible, affordable, and efficient for enterprises of any size," said Justin Hotard, Intel executive vice president and general manager of the Data Centre and AI Group.

Building an AI system typically demands substantial infrastructure-extensive model development and training, and collaboration among engineers, data scientists, and application developers. Inflection for Enterprise will allow enterprises to leverage a virtual AI co-worker trained on their company data, policies, and culture. 

The solution will enable fine-tuning of a model to be native to an organisation, potentially improving the usefulness of use cases through alignment with the company’s tone, purpose, product, service, and operating information.

“Enterprise organisations need more than generic off-the-shelf AI, but they don’t have the expertise to fine-tune a model themselves. We’re proud to offer an AI system that solves these problems,” said Inflection AI Chief Operating Officer Ted Shelton.

The solution provides application templates designed to let businesses skip hardware testing and model building and avoid capital expenses to scale quickly. While Inflection AI’s Pi consumer application was previously run on Nvidia GPUs, Inflection 3.0 will be powered by Gaudi 3 with instances on-premises or in the cloud powered by AI Cloud.

The solution also allows enterprises to own their models without sharing outside their organisation. Additionally, customers can host and run the model on their preferred architecture, whether hosted on-premises, in the cloud or hybrid.

The system is available presently through the AI Cloud and will be shipping to enterprises in Q1 2025.