As quantum computing progresses toward fault tolerance, the conversation around responsible disclosure is becoming more urgent and more nuanced. Our recent work analyzes the resources required to run Shor's algorithm on large-scale neutral atom quantum computers, with implications for industry-standard cryptographic protocols such as RSA and elliptic curve cryptography. Given these potential implications, we believe it is important to clearly explain why we are publishing these results, and how we approached this decision.

Over the past several years there has been a global push to develop large-scale neutral atom arrays. As it became clear that neutral atoms are well-suited to quantum error correction, research groups and companies began scaling these systems with fault tolerance explicitly in mind. This effort is now accelerating. Public awareness, however, has not kept pace. While it is widely understood that large-scale quantum computers are a future prospect, it is less appreciated that systems under development are already approaching the scale where fault-tolerant architectures become relevant. This gap between perception and reality is itself a source of risk.

Our paper is not a blueprint for breaking cryptography. It is a resource analysis: a theoretical existence proof that machines of a certain scale and capability could, in principle, be used to run Shor's algorithm. It identifies the resource requirements, but it does not provide a pathway to construct such a system, nor does it resolve the many open challenges that remain. The goal is to inform, not to enable.

The primary innovation in our work is not an improvement to factoring or discrete logarithm algorithms. It is an advance in the efficiency of fault-tolerant quantum computing itself. This brings us closer not only to running Shor's algorithm, but to running any quantum application that requires large numbers of logical qubits and extremely low logical error rates.

Responsible disclosure in the fault-tolerant quantum era differs from traditional security contexts. There is no single vulnerability to patch. Risk emerges gradually as hardware, error correction, and algorithms converge, and progress may not be externally visible. Reaching a cryptographically relevant machine may not require a dramatic, observable scale-up, but rather incremental improvements in architecture, control, and compilation. This makes the technology both powerful and easy to underestimate.

We are publishing because transparency serves the public interest. Withholding our results would slow progress across the entire field, not just for cryptanalytic applications. Avoiding its cryptographic implications would neglect a responsibility to communicate the urgency of safely migrating to quantum-secure standards. Clear, credible resource estimates ground discussions that might otherwise be driven by speculation, and provide a shared reference point for researchers, industry, and governments.