en.Wedoany.com Reported - MagiQware, a deep-tech spin-off from Delft University of Technology (TU Delft), has completed a pre-seed funding round with a total investment of €575,000 (approximately $658,000). The round was led by early-stage investment management firm LUMO Labs through the TTT.AI program, with an initial close of €500,000. Subsequently, Graduate Ventures and Delft Enterprises B.V. co-invested, expanding the total funding to €575,000. The funds will be used to accelerate product development, validate the automated software compiler, and expand the technical engineering team.
MagiQware focuses on reducing the substantial physical resource overhead required to sustain fault-tolerant quantum computing (FTQC). Standard quantum algorithms rely on quantum error correction to shield operations from phase decoherence, but executing non-Clifford logic gates requires "magic state factories" composed of specialized algorithmic subroutines. These factories distill high-fidelity magic states by filtering out physical noise, yet the distillation process itself becomes a major bottleneck, typically consuming up to 90% of the physical qubits and circuit footprint in a full-stack quantum computer.
The company builds specialized optimization tools within the quantum compilation and software stack to lower these technical barriers. Led by CEO Dr. Arash Ahmadi, CTO Shakeeb Majid, Device Lead Dr. Sahar Hejazi, and Theory Lead Dr. Ali Moghaddam, the technical team deploys specialized reinforcement learning models to orchestrate magic state production. By dynamically discovering and optimizing distillation circuit architectures through automated AI agents, MagiQware's compiler has demonstrated up to a 40% reduction in circuit length within target factories, thereby lowering the overall hardware overhead required for full-stack system developers without modifying the physical hardware layer. The aforementioned funding parameters and technical progress have been disclosed by LUMO Labs, Graduate Ventures, and MagiQware, respectively.
