As in semiconductors, we expect quantum's hardware layer will churn and commoditize, while the software layer on either side of it compounds upstream in design (EDA) and downstream in quantum-powered applications. Lorimer's actively identifying and investing in these compounding areas.
Our first investment in Lorimer Fund 2 is in an EDA for quantum (to-be-announced). We will selectively invest in quantum-enabled applications next, with an initial focus on sensing, where we see near-term commercial-grade applications emerging.
As with all of our theses: the declining cost and ubiquity of compute are force multipliers: they will make design tools like EDA more powerful - enabling cheaper, faster, more varied sensors that power enterprise-grade applications. Lorimer will invest in those applications that have a real path to a near-term market.
The core problems in quantum chip design: error correction, circuit compilation, qubit calibration, and routing, are combinatorial optimization problems that resisted human effort for decades. Recent research shows AI now beats expert humans at calibration - keeping real logical qubits stable on live hardware, autonomously and indefinitely, for the first time - and matches the best human work at compilation.
Net-net, ongoing breakthroughs demonstrate that quantum design is responsive to AI's increasing compute power - paving the way for an AI-native quantum EDA.
Ongoing breakthroughs demonstrate that quantum design is responsive to AI's increasing compute power - paving the way for an AI-native quantum EDA.
Within calibration: Google's pioneering paper Reinforcement Learning (RL) Control of Quantum Error Correction (Nature, 2026) showed an RL agent autonomously stabilizing logical qubits on the Willow processor, with a 3.5x improvement in logical stability under drift, plus a further 20% error suppression beyond exhaustive expert calibration.
Within compilation: DeepMind's AlphaTensor-Quantum (Nature Machine Intelligence, March 2025) matched or beat the best human-designed circuit optimizations fully automatically.
These two papers don't exist in a vacuum. Lorimer's thesis in this space has been informed by a large corpus of ongoing research, including by that of David Hyde at Vanderbilt, whose work fusing traditional physics simulations with AI informs our own thesis of how EDAs for quantums will develop. We were honored to feature David as a speaker at our Quantum Venture Studio The Dry Dock.
Classical semiconductors had this moment in the 1980s. Synopsys and Cadence turned chip design from a human exercise into a compiler problem. Quantum has no equivalent today - but it will - as design moves from notebooks, institutional memory, and legacy non AI-native platforms (e.g., Keysight's quantum suite, Qiskit Metal, KQCircuits) to an AI-native entrant. We're excited to announce our bet in this space soon.
Quantum sensing is commercially viable today. That sounds like a contradiction: how is this possible when quantum computers remain years away? This becomes logical when you consider the two both utilize quantum physics, but rearrange those properties differently.
A quantum computer needs its qubits isolated from the world. Every stray magnetic field, vibration, or temperature flicker corrupts the computation - which is why useful computing requires millions of coordinated qubits, error correction, and cryogenic isolation before it works.
A quantum sensor inverts this: its entire job is to couple to the environment, and the extreme sensitivity that destroys a computation is a quantum sensor's form of measurement. Decoherence is computing's enemy and sensing's product.
The extreme sensitivity that destroys a computation is a quantum sensor's form of measurement. Decoherence is computing's enemy and sensing's product...the result is that there are multiple commercially-viable quantum sensing applications today.
This inversion has a commercial consequence: compute is binary, hard, and unsolved today - and below the fault-tolerance threshold you have a science project. Sensing is incremental - a sensor that beats the classical alternative is sellable the day it works, frequently needs one quantum element rather than a million, runs at room temperature, and requires no error correction. The same quantum physics as quantum computers - applied very differently and with a very different fault tolerance threshold.
The result is that there are multiple commercially-viable quantum sensing applications today. Ones we're most excited about at Lorimer Ventures include:
These applications are only a small subset of commercially-viable quantum sensing applications. We're interested in meeting teams with a strong point of view in this space who share our views on current research and our excitement for quantum's capabilities.
If you are building in quantum, reach out.