RaiderChip, Spanish AI Startup, Secures €1M Seed Funding

RaiderChip, a company founded in 2024 by CTO Victor Lopez, is making waves in the tech industry with its innovative approach to accelerating AI on FPGA platforms. The company's flagship product, the GenAI v1 IP core, is designed to revolutionize AI inference at the edge and in the cloud.

The GenAI v1 implementation by RaiderChip supports a wide range of language models, catering to different needs in the AI space. From the Microsoft Phi-2 Small Language Model for simpler solutions to the Meta Llama-2 and Llama-3 Large Language Models for more complex applications, RaiderChip aims to provide a versatile solution for various AI tasks.

One of the key features of the RaiderChip GenAI v1 IP core is its massive floating-point multiplication parallelism and optimized memory bandwidth utilization. This parameterized IP core can be scaled for different applications by licensees, offering flexibility and efficiency in AI acceleration.

RaiderChip sets itself apart in the market by focusing on the normalized throughput metric, measuring tokens per second per unit of memory bandwidth. According to the company, the GenAI v1 IP core outperforms competitors such as Intel's Gaudi, Nvidia cloud GPUs, and Google's latest tensor processing unit, showcasing its superior performance in AI inference.

With a successful edge demo showcasing the execution of the Microsoft Phi-2 LLM 2.7B model on AMD Versal FPGA devices, RaiderChip proves its capabilities in achieving interactive speeds without quantization. The company's roots in Visengi, a Spanish electronic engineering firm, have provided a strong foundation for its current endeavors in AI acceleration.