A German startup whose technology tunes AI models for different chips during inference has raised €6 million in seed funding, it said today.
Daisytuner has raised the funds in a round co-led by Elaia and LEA Ventures, with participation from Amadeus APEX Technology Fund and Seedcamp.
Daisytuner is part of a new breed of startups trying to make it easier to run AI models efficiently on different chips, amid a challenge from rivals to Nvidia. Its AI tech, called Daisyflow, tunes AI models during AI inference, the computing process of running an AI model, as opposed to training an AI model.
Daisytuner's technology combines an AI-powered compiler, which automatically optimises software to run on different chips, with Daisyflow, a platform that enables companies to run AI models during inference.
Companies can deploy Daisyflow in their own data centres rather than relying on an external AI cloud provider.
In a company benchmark, Daisytuner claims that Daisyflow ran Alibaba's Qwen 2.5 model up to 18 per cent faster than vLLM, a widely used open-source inference engine, on AMD Instinct GPUs.
Daisytuner says it will use the funding to expand its enterprise customer base, develop its compiler technology and grow its team, which includes researchers from institutions such as ETH Zurich, TU Munich, RWTH Aachen, TU Darmstadt and Lawrence Livermore National Laboratory.
Would you like to write the first comment?
Login to post comments