Spanish AI startup RaiderChip raises €1 million seed funding
AI processor IP licensing startup RaiderChip Sl. (Solares, Spain) has raised €1 million in a seed funding round.
RaiderChip, founded in 2024 by CTO Victor Lopez, is developing an IP core aimed at accelerating AI on FPGA platforms.
The company states that GenAI v1 implementation supports various language models: from the Microsoft Phi-2 Small Language Model, for smaller and simpler solutions, to the Meta Llama-2 and Llama-3 Large Language Models for broader and more complex applications.
The company claims that its IP offers the fastest and most power efficient AI inference at the edge and in the cloud.
The RaiderChip GenAI v1 IP core features massive floating-point multiplication parallelism and optimized memory bandwidth utilization. The IP core is parameterized allowing it to be scaled for different applications by licensees.
Claimed to be better than Nvidia, Intel, Google
RaiderChip said the normalized throughput metric – tokens per second per unit of memory bandwidth – differentiates the quality of each accelerator design, independently of the memory technology and bandwidth selected by each vendor. The company states that under this metric GenAI v1 performs 37 percent better than Intel’s Gaudi, 28 percent better than Nvidia cloud GPUs and 25 percent better than Google’s latest tensor processing unit.
RaiderChip has an edge demo that executes the Microsoft Phi-2 LLM 2.7B model on AMD Versal FPGA devices, achieving interactive speeds on the vanilla LLM without quantization.
The compny was founded as a spin-off from Visengi, a Spanish electronic engineering company founded by Lopez in 2009. Among the achievements of the RaiderChip team while at Visengi was the development of an H.264 video encoder. Today, their goal is to bring Generative AI acceleration to Edge markets.
“Our goal is harnessing our two decades of experience in semiconductor design to achieve the maximum efficiency also in the field of AI, where the demand for computing resources makes the optimal resource usage absolutely crucial,” said Lopez, in a statement.
RaiderChip said it will use the funding will continue to deepen the pursuit of more efficient and versatile accelerators capable of meeting the growing market demand for local acceleration solutions in a field as open to innovation as Artificial Intelligence.
Related links and articles:
News articles:
Esperanto teams up with Rapidus for ‘post-GPU’ AI
ARM to compete with AI customers in 2025, says report
Nvidia, Softbank, Microsoft back UK AI driving startup Wayve with US$1 billion
Blaize raises US$106 million as its heads to AI SPAC deal
DeepX moves AI chips into video analytics, intelligent security
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
