Neureality Raises $8M for AI Inferencing Platform

NeuReality Emerges from Stealth with $8 Million Seed Funding

NeuReality, an Israeli startup focused on AI hardware, has officially launched and secured $8 million in seed funding. The company is pioneering a new method for enhancing AI inferencing by shifting away from traditional CPU-dependent systems.

Key investors in this round include Cardumen Capital, the crowdfunding platform OurCrowd, and Varana Capital. Furthermore, NeuReality announced the addition of Naveen Rao to its board of directors.

Rao brings significant expertise, having previously served as the General Manager of Intel’s AI Products Group and as CEO of Nervana System, a company acquired by Intel.

Founding Team Expertise

The company’s leadership team comprises CEO Moshe Tanach, VP of Operations Tzvika Shmueli, and VP for Very Large-Scale Integration Yossi Kasus.

Their collective background spans both AI and networking technologies. Tanach previously worked at Marvell and Intel, while Shmueli and Kasus both have experience at Mellanox and Habana Labs.

A New Approach to AI Infrastructure

NeuReality’s approach is rooted in the team’s deep understanding of networking and storage hardware. They aim to create an AI platform that overcomes the limitations of CPU bottlenecks.

In a recent interview, CEO Tanach indicated that the company is developing an architecture designed to allow hyperscale clouds and data centers to efficiently offload machine learning models.

“We’ve integrated techniques from the storage and networking domains,” Tanach explained. “We’ve applied these principles to AI, creating a bottom-up system built around the core engine requirements.”

Performance and Efficiency Gains

NeuReality anticipates its system will deliver substantial performance improvements.

The company projects a 15x increase in performance per dollar for basic deep learning offloading, with even greater gains achievable when the entire processing pipeline is migrated to their platform.

Future Plans and Target Markets

Currently, NeuReality is utilizing Xilinx FPGAs for prototyping. They plan to release their custom hardware solution in early next year.

Their target customers include major cloud providers, data center solution providers like WWT, and OEMs/ODMs seeking to implement specialized AI solutions for applications such as fraud detection.

The team’s work with Xilinx has laid the foundation for their custom chip development, and they are already planning a subsequent funding round to support this endeavor.

The Vision for Accessible AI

“AI is becoming increasingly prevalent in our daily lives and will continue to expand rapidly,” stated Tanach. “To ensure AI is accessible to all organizations, we must develop affordable infrastructure.”

He further emphasized that NeuReality’s technology will facilitate innovation in areas like disease treatment, public safety, and education, ultimately contributing to a smarter, cleaner, and safer world.

NeuReality believes that reducing the cost of AI infrastructure and AI-as-a-Service will remove significant barriers to adoption.

Image Credits: NeuReality