Quantum Dice is pioneering a new approach to tackling notoriously difficult optimization problems with its research into “p-bits,” a novel form of probabilistic computing. The company’s 2026 research explores how networks of these fluctuating bits – which randomly shift between 0 and 1 with controllable bias – can efficiently navigate complex landscapes mirroring real-world challenges like network design and financial modeling. Unlike classical computers that search for solutions sequentially, Quantum Dice’s system operates like “releasing thousands of hikers simultaneously,” naturally converging on optimal solutions. As Quantum Dice explains, “networks of p-bits act as natural optimisation systems: as they interact and flip, the network tends toward low-energy states that correspond to good solutions,” potentially offering a scalable path to solving problems currently considered computationally intractable.
Spin Glass Models Map to Real-World Optimisation Problems
Finding the lowest energy state of a three-dimensional spin glass is NP-hard, a challenge that extends far beyond the realm of condensed matter physics and into the heart of real-world optimisation. “Spin glasses” – systems of interacting magnetic spins with competing alignments – aren’t merely physics curiosities; their complex, “rugged energy landscape filled with many local minima” directly mirrors the challenges found in fields like network design and portfolio management. The ability to solve these systems hinges on locating the lowest energy configuration, a task that escalates exponentially in difficulty with increasing system size.
This inherent difficulty motivates the development of specialised computational approaches, as exact algorithms simply aren’t feasible for realistically sized problems. Quantum Dice’s research highlights how these spin glass models serve as analogues for a broad range of NP-hard combinatorial problems, where minimising cost, risk, or time are paramount. By translating variables into spins and constraints into couplings, complex logistical, financial, and defence questions become tractable as physics problems. Examples include Max-Cut in graphs, scheduling, and even “weapon-target assignment (WTA).” The physical “ground state” then directly corresponds to the optimal solution for the applied problem.
Furthermore, investigations into three-dimensional cubic lattices and biclique graphs reveal that “for fixed target solution qualities, the number of iterations required to reach a solution saturates once the system reaches a certain size.” This suggests a potential for constant-time scaling, a significant advantage for applications like analysing a 1,000-drone swarm in the same timeframe as a 100-drone swarm, allowing for neutralisation of command structures.
P-bits and Network Topologies for Probabilistic Computing
Beyond the established realms of classical bits and quantum qubits, Quantum Dice is pioneering a novel computational approach utilising probabilistic bits, or p-bits. Unlike their deterministic counterparts, p-bits inherently fluctuate between 0 and 1, but crucially, this fluctuation is not random; it’s “controllable and biased influenced by its neighbours’ values.” This characteristic allows networks of p-bits to function as natural optimisation systems, gravitating towards low-energy states that represent viable solutions to complex problems. The company’s research centres on leveraging specific network topologies to enhance this process, focusing on three-dimensional cubic lattices and biclique graphs.
Edwards–Anderson models, representing cubic lattices, are particularly suited to problems involving spatial correlation, such as “identifying leaders in a 3D drone swarm.” Biclique graphs, conversely, excel at modelling complete bipartite interactions, applicable to matching problems like interceptor-threat assignment. By mirroring these structures with interconnected p-bits, the system efficiently samples configurations based on their energy levels. Furthermore, Quantum Dice’s prototype demonstrably matches the performance of advanced quantum annealers, like the D-Wave Advantage, on large three-dimensional lattice instances, achieving this “using room-temperature devices.” This circumvents the need for the costly and complex cryogenic infrastructure required by quantum computing, offering a scalable, robust, and affordable alternative for tackling NP-hard optimisation challenges.
This research demonstrates that probabilistic computing offers a scalable, robust, and affordable approximation alternative for NP-hard optimisation. It shifts the paradigm from requiring faster clocks to simply requiring “more coin flips” in parallel.
Quantum Dice
Constant-Time Scaling Achieved with Room-Temperature Prototype
Quantum Dice is demonstrating a significant leap in computational efficiency with a room-temperature prototype utilising probabilistic computing, a novel approach beyond classical and quantum paradigms. This is particularly impactful for tackling NP-hard problems, where computational time increases exponentially with size. In practical applications, such as integrated air defence, this translates to maintaining consistent processing speeds even with escalating threat volumes; “In drone swarm analysis, the system can identify leader nodes in a 1,000-drone swarm in approximately the same time as a 100-drone swarm.” Quantum Dice asserts this offers “a scalable, robust, and affordable approximation alternative for NP-hard optimisation.”
