The computational resources required to decipher a standard data encryption technique have decreased tenfold, intensifying concerns about the vulnerability of this encryption method to quantum computers. Experts anticipate that the necessary computational scale could be achieved within the next ten years.
One of the most prevalent encryption algorithms in use is RSA. It underpins critical functions such as online banking and secure communications. The algorithm’s security relies on the inherent mathematical challenge of identifying the two prime numbers that were multiplied to produce a very large composite number. While researchers have understood since the 1990s that quantum computers could circumvent this difficulty, the prospect was largely theoretical due to the immense size such a quantum machine would need to be.
However, this landscape has gradually shifted. As researchers have succeeded in constructing increasingly complex quantum computers, the estimated size required for this specific task has diminished. In 2019, a significant development occurred when Craig Gidney from Google Quantum AI was a co-author on a paper that reduced the qubit requirement from 170 million to 20 million. Further progress was made in 2025 when Gidney developed a method to bring this figure down to under a million qubits. Most recently, Paul Webster and his team at Iceberg Quantum in Australia have managed an even more substantial reduction, bringing the requirement to approximately 100,000 qubits.
This latest study extends Gidney’s prior algorithmic advancements. A key distinction in their approach is the adoption of a different method for connecting and organizing qubits, known as qLDPC code. Unlike previous schemes where qubits could only interact with their immediate neighbors, qLDPC code allows for interactions with qubits at greater distances. This enhanced connectivity has the effect of increasing the density of information processing within the quantum computer.
With this improved connectivity, the research team estimates that using around 98,000 superconducting qubits—similar to those currently developed by companies like IBM and Google—would enable the breaking of a common RSA encryption form in approximately one month. To achieve the same outcome in a single day, the requirement would escalate to 471,000 qubits.
Several quantum computing enterprises are targeting the development of quantum computers with hundreds of thousands of qubits within the current decade. The new estimate is broadly applicable to various qubit technologies, depending primarily on their error rates and operational speed. Beyond the practicalities of executing a month-long computation, the feasibility of implementing Iceberg Quantum’s specific scheme remains a subject of discussion. The implications of such a capability are profound; any entity possessing a quantum computer of this power would gain access to a vast array of sensitive information, including emails, bank accounts, and confidential government files protected by RSA encryption.
Craig Gidney notes that these more stringent hardware demands present considerable manufacturing challenges, which are already the most difficult aspect of quantum computer development. Similarly, Scott Aaronson of the University of Texas at Austin has voiced reservations, highlighting the practical engineering difficulties in establishing the necessary interconnections between qubits that are not in close proximity.
IBM’s researchers have been strong advocates for qLDPC codes in recent years, and their quantum computing hardware has been adapted to better accommodate them. However, the ultimate success of this strategy is yet to be determined. A representative from IBM stated that qLDPC codes are expected to be a fundamental element of their quantum computers but did not offer specific comments on the viability of the new scheme.
Establishing connections between distant qubits is more readily achieved with systems utilizing extremely cold atoms or ions, two quantum computing modalities that have gained significant traction. Nevertheless, these quantum computers operate at slower speeds. According to the recent study, this slower operation could potentially drive the qubit requirements for breaking RSA encryption back into the millions.
Lawrence Cohen, also affiliated with Iceberg Quantum, emphasizes the importance of not being overly conservative with projections for such advancements. He points out that a breach of RSA encryption would have far-reaching consequences, and it is prudent to consider the possibility of it occurring sooner rather than later. Cohen also notes that breaking RSA encryption serves as a valuable benchmark for quantum computer development due to the extensive research dedicated to it. Beyond cryptography, his team’s approach could also be beneficial for running more sophisticated simulations in the fields of quantum materials and quantum chemistry.