Recogni Secures Funding to Drive the Next Generation of AI Hardware Innovation

Founded seven years ago, Recogni set out with a bold vision to revolutionize AI compute systems. What began as a mission to create the most compute-dense and energy-efficient accelerator for autonomous driving has expanded into a groundbreaking approach that could reshape the entire AI landscape. Today, Recogni is at the forefront of the rapidly advancing generative AI (GenAI) field, leveraging its innovative Pareto logarithmic number system to build AI inference systems that are not only faster and more accurate but also more energy-efficient and cost-effective than ever before.

Juniper has participated in Recogni’s $102 million Series C funding round, which has been co-led by Celesta Capital and GreatPoint Ventures. The investment signals a pivotal moment in the company’s journey, enabling Recogni to accelerate the development of its cutting-edge technologies and collaborate with Juniper Networks on AI inference solutions for data centers. With hyperscalers, cloud providers, and enterprises clamoring for more efficient AI hardware, Recogni’s partnership with Juniper comes at a critical time, as the demand for faster and more sustainable AI systems continues to grow.

“Juniper’s commitment to pushing the boundaries of networking aligns perfectly with Recogni’s mission to make AI more economical and accessible,” said Marc Bolitho, CEO of Recogni. “We are collaborating on a solution that will allow the world’s largest models to run at unprecedented speed, accuracy, and efficiency—drastically lowering total cost of ownership for our customers.” This vision is embodied in Recogni’s revolutionary Pareto system—a first-of-its-kind logarithmic number system that transforms traditional multiplication operations into addition, drastically improving power efficiency without sacrificing accuracy.

Recogni’s Pareto system addresses one of the most pressing challenges in the AI industry: the skyrocketing demand for compute power and the strain it places on energy resources. As data centers race to support the ever-expanding AI models, energy consumption and space limitations have become critical concerns. Recogni’s technology allows for significantly smaller, faster, and more energy-efficient chips, helping data centers maximize their compute power without draining excessive resources.

“The market for multimodal GenAI inference is accelerating rapidly and Recogni is at the forefront of exploring power efficiency without compromising performance or accuracy,” said Rami Rahim, CEO of Juniper Networks. “In this phase of inference innovation, cost-effectiveness and power efficiency—aligned with scalable fabric and network solutions—are extremely important. We were impressed with the team’s expertise and ability to approach complex issues with new ideas, so providing venture funding and collaboration made a great deal of sense.”

Through this investment and partnership, Recogni plans to scale its innovative AI hardware stack, supporting the growing need for GenAI inference in cloud environments and data centers. The collaboration will focus on improving the efficiency of GenAI models while maintaining high performance and accuracy—two key factors for cloud service providers and enterprises looking to deploy AI solutions at scale.

Recogni’s vision extends beyond just scaling AI for hyperscalers and enterprises. With the Pareto system, Recogni is driving down the cost of deploying and running AI models, ensuring that these cutting-edge systems remain accessible and viable for a broader range of applications. By providing a solution that reduces power consumption and operational expenses, Recogni is positioning itself as a leader in the next wave of AI infrastructure.

Earlier this year, Recogni demonstrated the power of its Pareto system, unveiling its second-generation chip for generative AI in collaboration with Taiwan Semiconductor Manufacturing Co. (TSMC). This chip is designed to take advantage of Pareto’s efficiency, providing developers with a powerful tool to train and run AI models faster, cheaper, and with far less energy. The system is set to go into production in 2026.

Recogni’s commitment to innovation doesn’t stop at hardware. The company’s team, with a strong presence in both Europe and North America, is composed of some of the most experienced talents in AI, deep learning, silicon engineering, and systems design. This expertise has been crucial in developing Pareto and optimizing Recogni’s GenAI hardware stack.

“We’ve spent years developing Pareto to address the core challenges of AI—speed, power efficiency, and accuracy,” said Bolitho. “As the AI race heats up, we are proud to be part of a movement that will help companies and individuals scale their AI models in an affordable and sustainable way. We’re excited for what lies ahead.”

Recogni’s Pareto system isn’t just a technological marvel—it’s a key to unlocking the true potential of AI while tackling the global challenges of energy consumption, data center space, and sustainability. With this new round of funding and Juniper Networks as a strategic partner, Recogni is positioned to lead the charge in building AI systems that are not only more powerful but also more responsible.