Startup profile · funding coverage
Daisytuner funding, valuation and investors
Multi-chip AI compiler tuning and enterprise self-hosted inference stack.
Verified company research · funding, investors and latest developments
Keep track of Daisytuner
Save this profile to your VCT watchlist for a quick return.
Funding, valuation & investors
Answer-first snapshotLatest funding
Seed (announcement)
€6M · November 16, 2026
Latest known valuation
Not publicly disclosed
Total disclosed equity funding
Not publicly disclosed
Excludes debt, grants, acquisitions, secondaries, and IPO proceeds.
Current status
private
seed · Darmstadt, Germany
Investors in latest funding
Lead: Elaia , LEA Ventures
Other: Amadeus APEX Technology Fund , Seedcamp
Overview
Daisytuner, founded in 2024 in Darmstadt by Lukas Trümper and Adrian Schmitz, combines compiler optimization, hardware benchmark verification, and an enterprise model-serving product called Daisyflow. Its November 16, 2026 €6M seed was co-led by Elaia and LEA Ventures with Amadeus APEX Technology Fund and Seedcamp participating.
Investor publicity quotes an up-to-18% Qwen 2.5 improvement over a vLLM setup on AMD Instinct, but no independent test or generalizable benchmark was located. Valuation, exact close and sales-led enterprise prices are not disclosed.
Why Daisytuner is interesting
€6M seed from Elaia, LEA Ventures and partners backs automation of performance engineering for AI models across chip architectures.
Product & use cases
Illustrative vendor-guided buyer workflow: assess models and available accelerators, deploy the Platform/Enterprise or Daisyflow stack inside private infrastructure, compile and benchmark workloads on actual chips, verify correctness, connect inference applications using standard/OpenAI-compatible APIs, and monitor throughput/operating costs. Daisyflow includes serving/routing/batching and can build on open vLLM components, not universally replace vLLM. No VCT test performed.
- Reduce manual model-to-chip kernel performance optimization.
- Operate private LLM, vision or speech inference on customer hardware.
- Use model routing and batching with enterprise SLAs.
Key facts
- €6M seed co-led by Elaia and LEA Ventures, announced November 16.
- Amadeus APEX Technology Fund and Seedcamp also participate.
- Up to 18% faster claim is vendor/investor-reported, not independently audited.
Funding history (newest first)
Seed (announcement)
2026-11-16 €6M- Elaia (lead)
- LEA Ventures (lead)
- Amadeus APEX Technology Fund (participant)
- Seedcamp (participant)
Source: Read source · amadeuscapital.com
Competitive landscape
Edge: Compiler autotuning and correctness checks plus managed enterprise serving; results depend on model, hardware and benchmark conditions.
-
vLLM adjacent
Open source inference runtime and test baseline, potentially also part of Daisyflow.
-
SGLang alternative
Open source model-serving option.
-
NVIDIA TensorRT-LLM alternative
NVIDIA-centric optimization infrastructure.
Market / IPO context
Editorial / static context — not a live quote.
Last updated:
Editorial note: AI tools assisted with research, structure, or drafting. Venture Capital Tracker retains human editorial responsibility for factual accuracy, relevance, and source quality before publication.