Pitching the future at Constructor Start Demo Day, part 3: five startups going deep (tech) for $100k

One pitch. One shot at $100,000. These are the stakes for the 14 finalist early-stage tech startups in the Constructor Start accelerator program, when they take the stage at Constructor University’s Demo Day on February 24.  The past two weeks have featured nine startups with future-facing solutions for both people and businesses. In this final feature, we dive into five deep-tech startups looking to push the frontiers of quantum and complex technologies in the business world. 

After honing their business plans through an eight-week intensive curriculum with mentorship from Constructor University’s global network of business leaders, Demo Day will see each team pitch their vision live to investors, venture capital firms, media and the public for a shot at securing life-changing investment. The top team will receive $100,000 in equity-free seed funding from Constructor Capital, part of a funding pool of up to $1 million available to Constructor Start participants.

These last five startups are tackling highly complex problems in specialized spaces like quantum computing, which is often described as the next great leap in computational power. For those interested in learning more about this fascinating technology, this year’s Demo Day will include a special panel discussion with leading experts from across the world about the current and future state of quantum computing. Sign up to attend Demo Day live in-person or remotely on February 24, 2026 to learn more.

MagiQware (Netherlands)

This Amsterdam-based startup is focused on solving what it sees as a critical bottleneck in quantum computing technology that could be the key to bridging the divide between today’s experimental machines and tomorrow’s real-world breakthrough applications in areas like climate modeling, fluid dynamics and nuclear fission. Rather than joining the race to build more qubits, MagiQware is targeting a less visible but equally important constraint: the production of so-called “magic states.” These specially prepared quantum resources are essential for running complex, fault-tolerant quantum algorithms – the kind of algorithms that are key to performing meaningful large-scale and long-term computations.

A helpful analogy is to think of magic states in quantum computing like crude oil. While oil holds the potential to propel an Airbus A380 from Bremen to Singapore, it can only do so when refined into high-performance jet fuel. In much the same way, raw quantum systems require “distillation” processes to produce high-quality or “high-fidelity” magic states. Currently, that distillation process is very resource-intensive, consuming the majority of a quantum computer’s resources before it even begins actual computation.

MagiQware is working to develop an AI-driven optimization engine that is designed to generate high-fidelity magic states much more efficiently, using fewer resources and less time. By reducing this hidden overhead, the company believes it can help produce enough ‘jet fuel’ to make universal quantum computing a reality.  

Quantum Logic (Delft)

Quantum Logic is another Dutch startup targeting bottlenecks in the quantum space. This Delft-based company specializes in advanced hardware that could make quantum computers far more scalable, focusing on electronics that operate at cryogenic temperatures, close to absolute zero. While most public attention focuses on increasing qubit counts (qubits being quantum bits, a basic unit of quantum information), every qubit must also be precisely controlled using electronic signals. As systems scale, the number of required cables, components and control pathways grows rapidly, creating heat, noise and engineering complexity that can overwhelm the system.

Quantum Logic is developing cryogenic and radio-frequency (RF) electronics designed to function inside ultra-cold quantum environments. Their technology converts densely-encoded input signals into highly precise control pulses using an approach called multiplexing, which allows multiple analog and digital signals to be combined over shared channels, dramatically reducing the hardware overhead required to control large numbers of qubits.

In today’s architecture, the control layer of quantum machines can become a scaling bottleneck long before theoretical qubit limits are reached. By developing scalable signal processing optimized with cryogenic performance, the founders of Quantum Logic believe they could break the bottleneck and unlock architectures capable of supporting millions of qubits.

Quasi AI (Germany)

Quasi AI is a Berlin-based company that is combining both quantum and classical computing technologies to develop high-performance simulation tools designed for some of the most computationally demanding tasks in engineering: modeling fluid flow, heat transfer and complex multi-physics systems. These types of Computational Fluid Dynamics (CFD) simulations hold tremendous value for industries like aerospace, semiconductor manufacturing, energy infrastructure and chemical production. However, running high-fidelity CFD simulations can be tremendously costly and resource-intensive, especially for turbulent or nonlinear systems.

Rather than replacing traditional simulation systems entirely, Quasi AI is pursuing a hybrid approach that integrates quantum machine learning with classical high-performance computing to dramatically accelerate simulations while maintaining or improving predictive accuracy. If successful, their technology could lead to dramatic efficiency gains and accelerated R&D cycles in high-tech manufacturing like semiconductor crystal growth, perovskite and silicon wafer production, cryogenic vessels and green hydrogen systems. Quasi AI is also planning to leverage their proprietary technology to offer “simulation-as-a-service" and synthetic data generation solutions.

Qendra (Switzerland)

Zurich-based Qendra describes its mission succinctly: “classical control for quantum computers.” While quantum processors rely on fragile atomic systems like trapped ions or neutral atoms, they depend on highly precise classical electronics to function. These control systems generate laser and radio-frequency (RF) signals that manipulate qubits with extreme timing accuracy.

Yet today’s setups are often fragmented and stitched together from multiple specialized subsystems with different software stacks and timing models. Add in a patchwork of specialized components and DIY solutions from experimental teams, and this can quickly lead to compounding complexity, with small adjustments in one subsystem introducing instability elsewhere, leading to fragile and unscalable architecture.

Qendra is building a unified, high-performance control platform tailored specifically for trapped-ion and neutral-atom quantum computers. Its system tightly integrates hardware, firmware and real-time control logic to manage complex multi-channel analog and RF workloads while maintaining deterministic timing, which is a critical requirement for quantum operations.

If quantum computing hardware is the engine, Qendra is building the control dashboard by transforming bespoke laboratory setups into reliable, scalable infrastructure for the next generation of quantum machines.

Tetractys (Netherlands/Germany)

Tetractys is applying advanced AI to one of the most complex industrial environments in existence: biomanufacturing. In biopharma and industrial biotechnology, production processes rely on living systems, often comprised of trillions of microorganisms interacting in tightly controlled reactors. Small deviations in temperature, nutrient balance or metabolic activity can and often do trigger cascading failures, leading to batch losses that can exceed €3 million for a single bioreactor in a single day. In fact, some estimates suggest the industry loses as much as €50 billion each year due to biomanufacturing failure and waste.

A truly international startup anchored at the University of Tübingen in Germany and the Eindhoven University of Technology in the Netherlands, Tetractys has developed what it describes as a graph-based “world model” AI. Essentially, it is an AI-powered system that can map and continuously analyze relationships between thousands of interconnected biological and process variables in real time. Instead of monitoring isolated parameters, this startup’s tailor-made AI solutions model entire production ecosystems as a dynamic network.

In the high-stakes game of biomanufacturing, the founders of Tetractys believe their solutions can reduce risk by providing AI-driven oversight that predicts and catches deviations early, optimizes yields and improves operational safety. They are betting on Tetractys becoming a standard enterprise layer for managing complex biosystems within the next five years, positioning AI not just as an analytical tool, but as an active manager of living production environments.

Join us live on Demo Day

Will one of these final five startups walk away from Demo Day with a fresh injection of capital to fuel their visions for tomorrow’s technologies? Register to attend Demo Day live in-person or remotely on February 24, 2026 to find out and learn more! 

Learn more and register to attend Demo Day live in-person or remotely on February 24, 2026.