Shor’s Algorithm Closes In on Encryption (Image Credits: Images.newscientist.com)
Sydney, Australia – A new fault-tolerant quantum computing design has dramatically lowered the hardware threshold for breaking widely used RSA encryption, potentially accelerating the quantum threat to digital security.[1][2]
Shor’s Algorithm Closes In on Encryption
Quantum computers have long posed a theoretical risk to RSA encryption, which secures online banking, emails, and government communications by relying on the hardness of factoring large numbers into primes.
Peter Shor’s 1994 algorithm demonstrated how a sufficiently powerful quantum machine could solve this problem exponentially faster than classical computers. Early estimates pegged the requirement at tens or hundreds of millions of physical qubits, far beyond current capabilities. Progress in algorithms and error correction has steadily eroded that barrier.[1]
Researchers at Iceberg Quantum unveiled their Pinnacle Architecture this month, claiming it enables RSA-2048 factoring—a standard key size—with under 100,000 physical qubits under realistic hardware assumptions.
Evolution of Qubit Estimates
Prior work set high bars. In 2019, Craig Gidney at Google Quantum AI and colleagues cut the need from 170 million to 20 million qubits. Gidney refined this further in 2025 to below one million.[1]
The new Pinnacle design achieves a tenfold reduction from that benchmark. It projects about 97,000 superconducting qubits could crack RSA-2048 in roughly one month, or 471,000 qubits for a one-day run.
| Scenario | Physical Qubits | Runtime |
|---|---|---|
| Superconducting, fast cycles (1 μs) | ~97,000 | 1 month |
| Superconducting, fast cycles | 471,000 | 1 day |
| Lower error rate (10⁻⁴) | ~22,000–53,000 | Feasible runtime |
| Slower cycles (1 ms, neutral atoms) | 3–13 million | 1–3 weeks |
These figures assume a physical error rate of one in 1,000 operations and microsecond-scale cycle times, compatible with platforms from IBM and Google.[2][3]
Unveiling the Pinnacle Innovation
Iceberg Quantum’s approach hinges on quantum low-density parity-check (qLDPC) codes, which replace rigid surface codes. These allow qubits to connect beyond nearest neighbors, boosting information density and efficiency.
The architecture features modular code blocks, a “magic engine” for distilling magic states needed for universal computation, and techniques for parallel processing in Shor’s algorithm. Simulations validated these under depolarizing noise models.
- Generalized bicycle codes with distance 16–24 for fault tolerance.
- Quasi-local connectivity suits sparse hardware links.
- Optimizations for magic state injection and Clifford operations.
- Scales to other tasks, like quantum chemistry simulations with 20,000–60,000 qubits.
All authors hail from Iceberg Quantum in Sydney, which recently secured a $6 million seed round.[3]
Engineering Challenges Persist
Experts caution that theory outpaces hardware. “These stricter demands make the hardware harder to make, and making the hardware is already the hardest part,” Gidney noted.[1] Scott Aaronson of the University of Texas at Austin highlighted difficulties in wiring distant qubits for superconducting systems, though trapped ions or neutral atoms face fewer connectivity issues but slower speeds that inflate qubit counts to millions.
IBM views qLDPC codes as a “cornerstone” of its roadmap, yet real-world scaling to 100,000 qubits remains unproven. Current machines top out at low thousands amid fidelity and control hurdles.[4][1] Lawrence Cohen of Iceberg Quantum urged vigilance: “It’s always much better to err on the side of this could very much happen sooner rather than later.”[1]
Cybersecurity Implications Ahead
Several firms target hundreds of thousands of qubits this decade, narrowing the gap to RSA vulnerability. This benchmark also aids useful applications like materials simulation. Organizations already migrate to post-quantum cryptography, but the shrinking timeline underscores urgency for banks and governments.
Key Takeaways
- Pinnacle Architecture uses qLDPC codes to cut RSA-2048 qubits below 100,000—a 10x gain over recent estimates.
- Superconducting hardware could achieve it in weeks to months; slower platforms need millions.
- Hardware engineering, not theory, remains the primary obstacle.
While no machine cracks RSA today, Pinnacle signals accelerating progress. What steps should secure your data next? Share your thoughts in the comments.
