Quantum Computing Qubit Counts: 2026 Status Report

Error Correction: The Barriers Are Falling

• QLDPC codes reduce hardware threshold by 10x (Iceberg Quantum "Pinnacle Architecture," February 2026). Using generalized bicycle codes instead of surface codes, RSA-2048 can be broken with under 100,000 physical qubits - down from ~1 million with surface codes. Iceberg is partnering with PsiQuantum, Diraq, and IonQ, all projecting systems of this scale within 3-5 years. These are simulation-based results, not experimental, but they fundamentally reset the hardware target.
• Below-threshold QEC now confirmed by four independent teams (Google, Quantinuum, Harvard/QuEra, USTC). This means the fundamental physics of quantum error correction works: adding more qubits makes the system more reliable, not less. This was the single biggest open question in quantum computing, and it has been answered.
• ETH Zurich demonstrated lattice surgery on superconducting qubits (February 2026, Nature Physics). Lattice surgery is the fundamental operation for fault-tolerant computing - all other logical operations can be built from it. This was the first demonstration on the superconducting architecture used by IBM, Google, and USTC.
• Reed-Muller codes enable full Clifford group without ancilla qubits (Osaka/Oxford/Tokyo, February 2026). Another pathway to reducing fault-tolerance overhead - fewer physical qubits needed per logical operation.
• Alice & Bob's "Elevator Codes" achieve 10,000x lower error rates for only 3x more qubits (January 2026). Their cat qubits are naturally protected against bit-flips; the elevator codes multiply that protection at minimal cost.
• IonQ's Beam Search decoder runs in <1ms on a standard CPU (January 2026). Real-time decoding was identified by the QEC Report 2025 as the critical remaining bottleneck. IonQ estimates three 32-core CPUs could correct 1,000 logical qubits.
• IonQ achieves 99.99% two-qubit gate fidelity - world record "four nines" (October 2025). Using EQC technology on mass-manufacturable semiconductor chips. Error rate of 8.4×10⁻⁵ per gate. At this fidelity, physical-to-logical ratio drops to as low as 13:1 (vs 500:1-1000:1 for typical superconducting systems).
• Infleqtion demonstrates first Shor's algorithm on logical qubits (September 2025). 12 logical qubits with error detection and loss correction on 1,600 physical qubits. Roadmap accelerated to 30 logical qubits in 2026, 1,000 by 2030.

Scaling: The Path to Millions of Qubits

• QuTech QARPET chip benchmarks 1,058 spin qubits at 2 million qubits/mm² (February 2026, Nature Electronics). Crossbar-tiled architecture requires only 53 control lines for 23×23 tiles. Compatible with existing CMOS fabrication. This brings semiconductor qubit testing in line with traditional chip industry practices.
• First-ever readout of Majorana qubits (QuTech, February 2026, Nature). Single-shot parity measurement via quantum capacitance with >1ms coherence. Solves a decade-old experimental challenge for Microsoft's topological qubit approach.
• Stanford's cavity-array microscope enables parallel qubit readout (February 2026, Nature). Demonstrated a 40-cavity array with a 500+ cavity prototype and a clear path to tens of thousands. This solves one of the biggest barriers to million-qubit systems: reading out qubit states fast enough.
• PsiQuantum appoints AMD/Xilinx veteran as CEO (February 2026). Signals shift from R&D to deployment. Sites under construction in Australia and Chicago. $1B+ Series E funding.
• Tsinghua demonstrated 78,400 optical tweezers using a single metasurface (December 2025). Optical tweezers are used to trap atoms in neutral-atom quantum computers. This is nearly 10x the current limit and shows the path to 100,000+ qubit systems.
• QuantWare announced the VIO-40K: 10,000 physical qubits via 3D chiplet architecture with NVIDIA integration, shipping 2028 at ~EUR50 million per chip (December 2025).

Attack Algorithms: Getting More Efficient

• Kim et al. (ePrint 2026/106) revised ECDSA attack estimates (February 2026). Optimized quantum circuits for Shor's algorithm on elliptic curves achieve 40% improvement in the qubit-count x depth product over all previous work. A practical attack on Bitcoin's secp256k1 requires ~6,500 logical qubits completing in ~2 hours.
• Shor's algorithm reliability reached 99.999% across over one million test cases (December 2025). One execution now suffices where thousands were previously needed.
• Tsinghua factored N=35 on real quantum hardware using optimized Regev's algorithm with space complexity at the theoretical minimum (November 2025). Small numbers, but a direct demonstration of quantum factoring on actual hardware.

The Gap Is Large but Closing Fast

The largest commercial quantum computers today have 1,600 physical qubits (Infleqtion Sqale) with the highest fidelity at 99.99% (IonQ, lab). Breaking Bitcoin's ECDSA requires roughly 8 million physical qubits using traditional surface codes - but the Pinnacle Architecture (Iceberg Quantum, February 2026) demonstrated that QLDPC codes can reduce the physical qubit requirement for RSA-2048 by 10x, to under 100,000. If similar techniques apply to ECDSA (plausible but not yet demonstrated), the gap narrows dramatically.

1. The gap is shrinking on multiple fronts simultaneously. It is not just qubit counts increasing - error rates are falling (IonQ's 99.99% reduces physical-to-logical ratios to as low as 13:1), algorithms are getting more efficient (Kim et al. 40% improvement), error correction codes are getting better (QLDPC 10x overhead reduction, Reed-Muller ancilla-free Clifford gates), networking allows combining multiple machines, and manufacturing is scaling up. Each of these independently compresses the timeline.

2. Company roadmaps project rapid scaling. IonQ targets 256 qubits at 99.99% fidelity in 2026 and 1,600 logical qubits by 2028. Infleqtion targets 30 logical qubits in 2026 and 1,000 by 2030. IBM targets 2,000 logical qubits by 2033. Google aims for a useful error-corrected machine by 2029. If any of these roadmaps come close to delivery, the CRQC threshold could be reached within a decade.

Why "Decades Away" Is No Longer a Safe Assumption

Nature (February 2026) reported a "vibe shift" among quantum researchers: the consensus is moving from "decades" to "within a decade" for useful quantum computers. Four independent teams have proven the physics of error correction works. The remaining challenge is engineering and manufacturing - a challenge backed by over $54 billion in government commitments and billions more in private investment.

Conservative estimates (Adam Back: 20-40 years) are increasingly outliers. The expert range now clusters around 2030-2035 for the first cryptographically relevant systems, with some projections as early as 2028.

What Should You Do?

• Never reuse Bitcoin addresses. Each spend reveals your public key. Once revealed, it is permanently vulnerable to future quantum attack.
• Monitor migration proposals like BIP-360 (Bitcoin) and the Glamsterdam/Hegota upgrades (Ethereum). These are the mechanisms that will eventually protect the ecosystems.
• Consider quantum-resistant alternatives. QRL / QRL 2.0 (Zond) has been operating with post-quantum cryptography since 2018. QRL 2.0 (Zond) adds EVM-compatible smart contracts with quantum-safe signatures.
• Take HNDL seriously. Your transactions today are being recorded by adversaries for future decryption. The Federal Reserve has confirmed these attacks are happening now.
• Stay informed. The Quantum News page tracks every major development as it happens. Quantum News