Recogni raises $102m to further development of its AI inference chip

This investment was co-led by Celesta Capital and GreatPoint Ventures, with contributions from existing backers such as Mayfield, DNS Capital, BMW i Ventures, and SW Mobility Fund. The round was further bolstered by new participants Pledge Ventures and Tasaru Mobility Investments, wholly owned by the Public Investment Fund (PIF), alongside debt financing from HSBC Innovation Banking. The infusion of capital is earmarked for the advancement of next-generation systems for AI inference solutions, aimed at enhancing performance and power efficiency, ultimately delivering the lowest total cost of ownership.

The rapid advancement of AI technology, coupled with the significant growth in model sizes, presents a challenge for AI's application across diverse markets. AI inference, the process of applying trained models to real-time data for predictions or solutions, demands more efficient and power-saving solutions than previously available. Recogni's power-efficient AI inference acceleration technology is set to revolutionise AI computing in crucial fields such as generative AI and intelligent autonomy by offering scalable solutions.

Marc Bolitho, CEO of Recogni, points out the urgent gap between rapidly expanding AI models and the slower evolution of computing capability: "The critical need for solutions that directly address the key challenges in AI inference processing – compute capability, scalability, accuracy, and energy savings – is more urgent than ever. Recogni is leading this transformative wave, engineering pivotal advances that will redefine data centres and enterprise and revolutionise industries like automotive and aerospace," he said.

Traditional AI acceleration methods fall short of meeting the future demands of AI. Current practices, which largely depend on energy-intensive GPUs for cloud-based AI training and inference, place a significant burden on the computational, cooling, and power infrastructures of data centres, making the process both financially and ecologically unsustainable. Recogni proposes a groundbreaking approach to inference processing that addresses the AI computing challenge from its foundation, offering scalability for the future.

“When you have a solution that achieves 10x higher compute density, 10x lower power and 13x less cost per query, it’s a no-brainer to invest and help bring that solution to market,” said Ashok Krishnamurthi, Managing Partner at GreatPoint Ventures. “The compute demand for AI applications is going to be significantly more than what the experts are forecasting, and addressing the power consumption piece of the puzzle is critical at this stage resulting in significantly lower operational costs. The repeat investment and high calibre of new investors in the C round is a true testament to Recogni’s team and technology.”

“Generative AI systems in the market today are highly inefficient and consume too much power while adding system complexity, but Recogni’s proven technology is raising the bar on power performance to address the large compute needs of AI workloads,” said Sriram Viswanathan, Celesta Capital Founding Managing Partner and Chairman of Recogni’s Board of Directors. “With the rapid AI proliferation over the past year, there is a clear need for Recogni’s products, which are purpose-built for efficient inference compute. We’re pleased to support the company’s ability to meet market demand with its innovative solutions and look forward to seeing them continue to expand their exciting product offering.”

In December 2022, Recogni introduced its pioneering low-power AI compute product, Recogni Scorpio, marking a significant milestone as the world’s first 1000 TFLOPS (1 Peta-Flop) class inference solution. Scorpio’s performance not only affirms Recogni’s vision but also its capability in overcoming AI compute barriers, ensuring scalable performance across generations.