In a recent development that could reshape the conversation around the quantum vulnerability of major blockchain networks, a group of crypto researchers has published a paper indicating that the projected timeline for quantum attacks on Bitcoin and Ethereum may be considerably longer than previously thought. The study, which was shared with CoinDesk, demonstrates that both human analysts and advanced artificial intelligence agents have succeeded in surpassing the performance of Google’s March‑2024 benchmark on a crucial sub‑routine that underpins Shor’s algorithm—a quantum algorithm famed for its ability to factor large integers and break widely used public‑key cryptography. Shor’s algorithm has long been the centerpiece of the so‑called "quantum apocalypse" narrative, where the eventual arrival of sufficiently powerful quantum computers could render the elliptic‑curve digital signatures that secure Bitcoin, Ethereum, and countless other digital assets obsolete.
The prevailing concern among many in the cryptocurrency community has been that once a quantum computer can reliably execute the algorithm at scale, it would be able to derive private keys from publicly visible addresses, enabling attackers to siphon funds with alarming ease. The new research challenges that timeline by focusing on a specific computational step that is essential for Shor’s algorithm to function efficiently: the modular exponentiation operation.
This operation, while mathematically straightforward, becomes exponentially more demanding as the size of the numbers involved grows. Historically, estimates of quantum readiness have assumed that achieving a breakthrough in this step would be a relatively straightforward engineering hurdle for quantum hardware developers.
In the paper, the authors detail a series of experiments in which they tasked both seasoned mathematicians and state‑of‑the‑art AI models with optimizing the modular exponentiation process. The participants were given a set of parameters mirroring the cryptographic key sizes used by Bitcoin (256‑bit elliptic‑curve keys) and Ethereum (also 256‑bit). Their goal was to reduce the number of quantum gates and overall circuit depth required to complete the operation, thereby lowering the error tolerance and physical qubit count needed for a successful attack. Remarkably, the human experts—drawing on deep knowledge of number theory, algorithmic shortcuts, and circuit design—were able to propose novel decompositions that shaved off roughly 20 percent of the gate count compared with the baseline model that Google reported in its March results.
Simultaneously, the AI agents, trained on vast corpora of quantum circuit literature and reinforced through iterative self‑play, discovered alternative pathways that achieved a similar reduction, and in some cases even outperformed the human solutions by a marginal margin. When the researchers combined the most efficient human‑derived and AI‑derived techniques, the cumulative improvement amounted to a 50 percent reduction in the quantum resources required for the modular exponentiation step. This translates directly into a halving of the estimated number of logical qubits—and consequently the physical qubits—necessary to mount a successful Shor‑based attack on Bitcoin and Ethereum addresses. The implications of this finding are twofold.
First, it suggests that the quantum threat timeline may be more extended than many worst‑case scenarios have projected. If the hardware requirements are effectively doubled, the engineering challenges for building a quantum computer capable of breaking current blockchain cryptography become significantly steeper. Quantum error correction, a major bottleneck in scaling up quantum processors, would need to handle a larger error budget, pushing the required qubit count well beyond the thresholds that leading quantum labs anticipate achieving in the next five to ten years.
Second, the study underscores the importance of a multi‑disciplinary approach to quantum‑resistant security. By leveraging both human ingenuity and machine‑learning‑driven optimization, researchers can uncover efficiencies that neither side might discover in isolation.
This collaborative model could be extended beyond the modular exponentiation problem to other components of cryptographic algorithms, potentially leading to a broader reassessment of quantum risk across the digital finance ecosystem. Critics, however, caution that the results should not be interpreted as a guarantee that cryptocurrencies are safe from quantum attacks for the foreseeable future. The paper acknowledges that while the gate count has been reduced, the overall complexity of Shor’s algorithm remains formidable, and breakthroughs in quantum hardware—such as the development of error‑corrected logical qubits with higher fidelity—could still accelerate the timeline.
Moreover, the research highlights a new variable in the "quantum clock" that policymakers and industry stakeholders must consider: the speed at which algorithmic optimizations can be discovered and implemented. As AI continues to advance, it may become increasingly capable of autonomously refining quantum circuits, thereby compressing the gap between theoretical feasibility and practical execution. In response to these findings, several blockchain projects have begun to accelerate their migration plans toward quantum‑resistant cryptographic schemes.
Post‑quantum signatures based on lattice‑based constructions, hash‑based signatures, and multivariate polynomial approaches are being evaluated as potential replacements for the current elliptic‑curve mechanisms. The research community is also exploring hybrid models that combine classical and quantum‑resistant primitives, aiming to provide a layered defense that can adapt as quantum capabilities evolve. In summary, the paper shared with CoinDesk offers a nuanced perspective on the quantum threat landscape. By demonstrating that both human and AI agents can substantially cut the resource requirements for a core component of Shor’s algorithm, the study effectively halves the previously estimated quantum attack window for Bitcoin and Ethereum.
While this does not eliminate the risk, it does buy the cryptocurrency ecosystem valuable time to transition to more robust, quantum‑safe cryptographic standards. The findings also emphasize the growing role of AI in cryptographic research, suggesting that future security assessments will need to account for the rapid, collaborative advancements that such technology can enable.