Armed with new funding, Axelera AI debuts Titania, its latest low-power chip for edge AI inference – News

Artificial intelligence chipmaker Axelera AI B.V. today announced Titania, the next generation of its low-power yet high-performance silicon for running generative AI and computer vision inference workloads at the network edge.

Coinciding with the launch of Titania, the company also revealed yet more funding. It has raised €61.6 million ($66.5 million) from the EuroHPC Joint Undertaking Digital fund, as part of its Autonomy of RISC-V for Europe or DARE project. The latest cash injection comes less than a year after the startup closed on its oversubscribed $68 million Series B funding round, and brings its total amount raised to more than $200 million.

Founded in 2021, Axelera AI is the developer of specialized computer chips optimized to run AI software at the network edge, outside data centers. Its compute hardware accelerates AI inference, the processing that chatbots and computer vision do, more efficiently and closer to the point where the data is produced.

The startup believes that by putting its technology closer to the point where the AI processing needs to happen, it can speed up AI interactions. This will allow AI to have greater impact and control when its deployed on connected devices such as internet of things systems and industrial robots.

The new Titania chip is the latest iteration of the startup’s edge computing platform, powered by its proprietary Digital In-Memory Computing technology that leverages a scalable, multicore architecture. It’s based on the open-source RISC-V processor architecture that rivals Arm-based chips.

Axelera AI says multiple Titania chiplets will be packaged within a System-in-Package. The chips are enhanced with proprietary RISC-V vector extensions that can help them to meet the increased demands for AI across various market segments, including high-performance computing, enterprise data centers, robotics, automotive and other industries. It does this while maintaining the efficiency of an edge-oriented architecture. The company hasn’t yet provided any performance specifications, but it insisted that its D-IMC technology enables “near-linear scalability” without the substantial power and cooling overheads associated with traditional AI accelerators, such as graphics processing units.

“Uniquely, our architecture facilitates scaling from the edge to the cloud, streamlining expansion and optimizing performance in ways that traditional cloud-to-edge approaches cannot,” said Axelera AI co-founder and Chief Technology Officer Evangelos Eleftheriou.

Axelera AI believes it’s solving a pressing challenge in the AI industry, where concerns over the performance, high costs and sustainability of existing cloud-based AI accelerators are intensifying amid the rise of DeepSeek Ltd.’s R1 and other low-cost large language models. At the same time, these models still require significant inference computing resources. The company said Titania is designed to address these requirements, enabling rapid throughput and cost-efficiency for the most data-intensive applications.

Axelera AI Chief Executive Fabrizio Del Maffeo said Titania represents a key milestone and an important validation of its approach to edge AI computing.

“Today, we deliver a cutting-edge hardware and software platform for accelerating computer vision on edge devices at a fraction of the cost and energy consumption of current solutions, and Titania builds upon this unique product suite,” he insisted.

Looking forward, Axelera AI said it will use the bulk of the funds from today’s round to expand its research and development teams, which are spread across Belgium, Italy and the Netherlands.

Image: News/Dreamina