Preferred Networks: A Deep Learning Startup Powers the Internet of Things

Preferred Networks, Inc. (PFN), a start-up specialized in deep learning technologies, a branch of artificial intelligence (AI) research, differentiated itself early on by aligning with Japan’s manufacturing might and bringing deep learning to the internet of things (IoT). The case follows the start-up as it evolves into a highly valued company with over 200 employees and global partners across various industries. It offers an overview of the AI business landscape and an explanation of deep learning. PFN’s trajectory shows how technology-heavy research firms spark innovation, attract business partners and collaborators, manage as they grow, and decide what business model best suits their needs. The case is intended for use in classes on artificial intelligence, technology and operations management, marketing of complex products and technologies, entrepreneurship and strategic partnerships for research-heavy startups.

The purpose is to learn how to build a business around new and emerging technologies such as deep learning and other advanced machine learning methods. In addition, it can be used to study • Alliances and partnerships between tech-oriented startups and large organizations interested in the technology; • The marketing of new and complex technologies without a marketing team; • Team management and recruitment in an organization with abundant technical talent but limited business experience; • How deep learning and other machine learning methods can be applied to industries such as autonomous driving, factory automation, and healthcare. Key takeaways: • Understand the meaning of “artificial neural network”, “machine learning”, “artificial intelligence”, “shallow” vs. “deep learning” and their role in the AI boom; • An overview of the AI industry and its growing importance across various sectors; • (as seen through the lens of PFN) Specific applications of deep neural networks in autonomous driving, factory automation, and healthcare; • Assess different business models for technology and research-heavy start-ups (in this case integrator/partnerships, profit shares and new product development).