Researchers have reduced the quantum computing power required to break the widely used RSA encryption algorithm by a factor of ten, to about 100,000 qubits. This advancement builds on prior work and highlights growing vulnerabilities in current security systems. However, significant engineering challenges persist in building such machines.
The RSA algorithm, a cornerstone of secure online banking and communications since the 1990s, relies on the difficulty of factoring large numbers into their prime components. Quantum computers have long posed a theoretical threat to this method, but practical implementation seemed distant due to the immense resources required.
Progress has accelerated in recent years. In 2019, Craig Gidney at Google Quantum AI lowered the qubit threshold from 170 million to 20 million. By 2025, Gidney further reduced it to under a million. Now, Paul Webster and colleagues at Iceberg Quantum in Australia have pushed it down to approximately 100,000 qubits, using algorithmic improvements and a qubit arrangement called qLDPC code. This code allows qubits to connect over longer distances, boosting information density.
For superconducting qubits, like those developed by IBM and Google, the team estimates that 98,000 qubits could crack common RSA encryption in about a month. Achieving the same in a day would demand 471,000 qubits. Several firms target hundreds of thousands of qubits within the decade, though error rates and speed matter more than the underlying technology.
Experts caution on feasibility. "These stricter demands make the hardware harder to make, and making the hardware is already the hardest part," Gidney noted. Scott Aaronson at the University of Texas at Austin expressed reservations about engineering distant qubit connections. IBM views qLDPC codes as a "cornerstone" of its systems but offered no specifics on this scheme.
Alternative approaches using cold atoms or ions enable easier long-range links but operate slower, potentially requiring millions of qubits for RSA breaking. Lawrence Cohen of Iceberg Quantum urged vigilance: "I think it’s important to never be conservative with the timelines... it’s always much, much better to err on the side of this could very much happen sooner rather than later."
Beyond encryption, the method could enhance simulations of quantum materials and chemistry. The findings appear in arXiv DOI: 10.48550/arXiv.2602.11457.