Nobody knows when Q-Day will come. Published estimates run from 2029 (Gartner) to the mid-2030s, and the people making them are not all predicting the same thing. The tracker below lists who expects what, how far today's machines are from breaking RSA-2048, and every hard deadline, each with its source. The trend underneath is clear: each new estimate of the hardware needed to break RSA-2048 has come out lower than the one before.
Q-Day tracker
Last checked . We only list figures we checked at the original source.
When do experts expect Q-Day?
Forecasters do not all predict the same thing. Read the line under each date.
- 2029GartnerQuantum computers break public-key cryptography · Gartner
- April 14, 2030Cloud Security AllianceA quantum computer can break present-day cybersecurity infrastructure · CSA
- 2030ForresterQuantum computers can break mainstream public-key cryptography · Forrester
- Sep 2033Metaculus forecasters (26)The RSA-2048 challenge number is factored by a quantum computer. Median forecast; range Jun 2030 to Feb 2039 · Metaculus 30596
- Mar 2035Metaculus forecasters (136)One of the RSA challenge numbers is factored by a quantum algorithm (a much easier target than RSA-2048). Median forecast; range Oct 2029 to Jul 2047 · Metaculus 3684
- 28–49% by ≈ 2035Global Risk Institute expert survey (26 experts)A quantum computer breaks RSA-2048 within 24 hours. Chance within 10 years; 51–70% within 15 years (≈ 2040) · GRI report 2025
- no dateNSA“NSA does not know when there will be a CRQC. Expert assessments disagree significantly about timing.” · CNSA 2.0 FAQ v2.1
- no dateG7 Cybersecurity Working GroupCalls for preparation now, “although the exact timeline is uncertain” · G7
The qubit gap
8 of the 1,400 logical qubits needed
How the estimate for breaking RSA-2048 has fallen:
- 201920 million physical qubits · 8 hours · Gidney & Ekerå 2019
- 2025under 1 million physical qubits · under a week · Gidney 2025
- 2026under 100,000 physical qubits · on paper, with a different type of error-correcting code · Iceberg Quantum 2026
- 2026about 23,000 physical qubits · cat qubits, about 3 days, on paper; assumes almost no bit-flip errors (about 45,000 at one bit flip per hour) · Alice & Bob 2026
Deadlines for the switch to quantum-safe encryption
A deadline is not a prediction. It is the date by which governments and companies want to be ready.
- End of 2026EU (NIS Cooperation Group)All EU member states should start the transition to post-quantum cryptography · EU
- Jan 1, 2027NSA / CNSSP 15 (US)New purchases for US national security systems must be CNSA 2.0 (quantum-safe) compliant · CNSA 2.0 FAQ v2.1
- 2028NCSC (UK)UK organisations finish discovery and an initial migration plan (NCSC guidance, not law) · NCSC
- 2029GoogleGoogle's own deadline to finish its post-quantum migration (a company target, not a prediction) · Google
- After 2030NIST (US)RSA, elliptic-curve and other quantum-vulnerable public-key algorithms become “deprecated”: still allowed, but at a known risk (final guidance) · NIST IR 8547
- Dec 31, 2030White House, EO 14412 (US)High-value US government systems use post-quantum key exchange · EO 14412
- Dec 31, 2030NSA / CNSSP 15 (US)Equipment in US national security systems that cannot support CNSA 2.0 is phased out · CNSA 2.0 FAQ v2.1
- End of 2030EU (NIS Cooperation Group)Critical infrastructure in the EU should be protected with post-quantum cryptography · EU
- 2031NCSC (UK)UK organisations finish their highest-priority post-quantum migrations (NCSC guidance, not law) · NCSC
- Dec 31, 2031White House, EO 14412 (US)High-value US government systems use post-quantum signatures · EO 14412
- Dec 31, 2031NSA / CNSSP 15 (US)CNSA 2.0 algorithms mandatory for US national security systems · CNSA 2.0 FAQ v2.1
- 2035OMB (US)All remaining US federal systems are quantum-resistant · OMB M-26-15
- 2035NSA / NSM-10 (US)All US national security systems are quantum-resistant · CNSA 2.0 FAQ v2.1
- 2035NCSC (UK)UK organisations complete the migration of all systems to post-quantum cryptography (NCSC guidance, not law) · NCSC
- After 2035NIST (US)RSA, elliptic-curve and other quantum-vulnerable public-key algorithms are “disallowed” under NIST standards · NIST IR 8547
2016–2019: A distant problem
In 2016, NIST asked cryptographers worldwide for new algorithms that could withstand a quantum computer, the start of what became an eight-year effort. The goal was to replace RSA and elliptic-curve cryptography. A large enough quantum computer running Shor's algorithm could break both. Such a machine is called a cryptographically relevant quantum computer (CRQC).
In 2019, Craig Gidney and Martin Ekerå calculated that breaking RSA-2048 would take about 20 million physical qubits running for 8 hours. Measured in qubits multiplied by time, that was about a hundred times cheaper than earlier estimates. It was still far beyond anything that existed.
The same year, Google reported that its 53-qubit Sycamore processor finished a sampling task in about 200 seconds. Google estimated that a classical supercomputer would need about 10,000 years for the same task. The task had no practical use and nothing to do with codebreaking. It showed that quantum hardware could do something ordinary computers struggle with.
2022–2023: Governments set dates, errors start to shrink
In September 2022, the NSA announced CNSA 2.0: the quantum-resistant algorithms that US national security systems will have to use. The schedule runs in steps. New purchases must comply from 1 January 2027, and all national security systems should be quantum-resistant by 2035. The NSA does not present these dates as a forecast. In the same document it writes that it "does not know when there will be a CRQC".
In February 2023, Google showed for the first time that a larger error-corrected qubit can make fewer mistakes than a smaller one. A logical qubit is one reliable qubit built from many unreliable physical qubits. Think of a message repeated several times over a noisy phone line: more repetitions make errors easier to catch, but only if each repetition is clear enough. Google's larger version ("distance 5") modestly outperformed the smaller one ("distance 3").
2024: Standards arrive, error correction scales
On 13 August 2024, NIST published its first three post-quantum standards: ML-KEM (FIPS 203) for encryption, and ML-DSA (FIPS 204) and SLH-DSA (FIPS 205) for digital signatures. NIST urged administrators to start switching "as soon as possible". More on how these work in our guide to post-quantum cryptography.
In December 2024, Google unveiled Willow, a 105-qubit chip whose errors fell exponentially as it used more qubits for error correction (our explainer). Google also reported a benchmark completed in under five minutes that it estimated would take one of today's fastest supercomputers 1025 years. Like Sycamore's task, that benchmark breaks no code.
2025: The estimate drops below one million
In March 2025, NIST selected HQC as a backup encryption algorithm. It rests on different mathematics than ML-KEM, in case ML-KEM ever turns out to be weak.
In May 2025, Gidney revised his own estimate: RSA-2048 could be broken with fewer than one million physical qubits in under a week. That is at least twenty times fewer qubits than in 2019, under the same assumptions about the hardware. The gain came from better algorithms and error correction, not from better machines. Three papers that moved Q-Day to 2029 explains why that matters.
2026: Lower estimates, firmer deadlines, a sober measurement
In February 2026, Iceberg Quantum put the requirement below 100,000 physical qubits, using a different family of error-correcting codes and the same hardware assumptions as Gidney. That is a calculation on paper. Caltech’s array of 6,100 atoms, a record when it was reported in September 2025, does not yet compute.
In March 2026, the Global Risk Institute published its yearly expert survey. The 26 experts put the chance of a CRQC within 10 years at 28–49%, and within 15 years at 51–70%. Later that month, Google set 2029 as the deadline for its own post-quantum migration. It pointed to progress in hardware, error correction and factoring estimates. That is a company deadline, not a prediction of Q-Day.
In June 2026, the US government turned plans into dates. Executive Order 14412 requires high-value federal systems to use post-quantum key exchange by the end of 2030 and post-quantum signatures by the end of 2031. OMB memo M-26-15 gives every federal agency until 22 October 2026 to submit its migration plan, and spreads the whole move over five phases up to 2035.
In September 2026, researchers at Sandia National Laboratories, with Quantinuum and NVIDIA, measured how much error-corrected computing today's machines actually deliver. The best result was 8 logical qubits on Quantinuum's Helios-1. By their measure, breaking RSA-2048 needs more than 100,000 times the computing capacity of the best machine they tested.
The latest estimates, the qubit gap and every hard deadline are in the Q-Day tracker at the top of this page.
Why the range matters
The forecasts in the tracker run from 2029 to the mid-2030s. Part of the spread comes from measuring different things. Gartner speaks of breaking public-key cryptography. The GRI experts mean breaking RSA-2048 within 24 hours. One Metaculus question counts any of the RSA challenge numbers, including ones far smaller than RSA-2048. Crowd forecasts also move: that question's median slipped a year later between August and September 2026.
What nobody knows is how fast the hardware will scale. The estimates on paper keep falling, while the best measured machine is still far short. A new algorithm could close the gap faster. An engineering barrier could stall it for years.
For planning, the exact year matters less than it seems. Moving a large organisation to new cryptography takes years, which is why the NSA and the US government set deadlines without claiming to know when Q-Day will come. If a migration takes five years and a CRQC arrives in 2031, the work had to start in 2026.
What the timeline means for data at risk today
The timeline also matters for data sent today. In a harvest now, decrypt later attack, someone records encrypted traffic now and decrypts it once a quantum computer exists. That is why Google calls the threat to encryption "relevant today". Anything that must stay secret for more than a few years is exposed, whatever the exact date of Q-Day turns out to be. The countdown on our homepage follows the earliest confirmed estimate.