Startup

Daisytuner

Optimization platform for modern compute: more performance from the chips you run.

  • AI · Infrastructure & Models
  • Darmstadt, Germany
  • Seed

About Daisytuner

Daisytuner is a Darmstadt-based startup building a self-learning compiler and optimization platform for AI and HPC workloads. Its Daisyflow platform tunes models at inference time across CPUs, GPUs and accelerators, so companies can run models efficiently on hardware of their choice, including their own data centres. The company grew out of Lukas Truemper's compiler research at ETH Zurich and raised a €6M seed round in November 2026.

Founders

  • Adrian Schmitz

    Co-founder

  • Lukas Truemper

    Founder & CEO

    Founded Daisytuner based on his research on AI-driven compilers at ETH Zurich.

Daisytuner logo
Industry
AI · Infrastructure & Models
Based in
Darmstadt, Germany
Stage
Seed
Founded
2024

Funding rounds

DateRoundAmountLead investors
Seed€6.0mElaia, LEA Ventures
Pre-seedUndisclosed—