Korean Investment in Quantum Computing, and the Five Machines It Has Bought

Korean investment in quantum computing is easier to measure by its yearly budgets than by its headline target. The figure most often quoted is ₩3 trillion by 2035. That is a floor for public and private money together, set by the government in 2023, and the private part of it is pledged only as far as 2027. The yearly budget is firmer ground. Its dedicated quantum programmes received ₩198 billion in 2025, about $148 million at today’s exchange rate.

The money has bought real machines. Yonsei University runs an IBM system and Chungbuk National University runs one from IQM, while the state computing institute KISTI is installing a 100-qubit machine from IonQ. Two Korean companies have bought machines too. None of the buyers or sellers has said what any of the five cost, which leaves the price of quantum hardware in Korea a matter of guesswork.

The ₩3 trillion target is a floor for public and private money together, not a state pledge

The number comes from the Korea Quantum Science and Technology Strategy of June 2023. The science ministry, MSIT, put it in its release as “’23~’35년 간 민•관 합동 최소 3조원 투자”, which translates as “at least ₩3 trillion of joint public and private investment from 2023 to 2035”. The same release splits the sum. The government’s share is ₩2.4 trillion over the full thirteen years, while the private ₩600 billion covers only the five years from 2023 to 2027, so the two parts are not alike.

The windows differ for a reason. KED Global reported at the time that private firms had not yet settled how much they would put in after 2027, so the strategy left those years out of the private pledge. In May 2023 press reports had put the total at ₩3.5 trillion. MSIT denied the report, saying the claim “사실이 아닙니다”, which means “is not true”.

The legal base came later. The National Assembly passed the quantum act on 6 October 2023, and it came into force on 1 November 2024. It set up a Quantum Strategy Committee chaired by the Prime Minister, though that committee did not first meet until 12 March 2025, more than four months after the law took effect. An amendment adding duties on quantum security passed the Assembly in April 2026 and the Cabinet in May. It takes effect six months after it is published.

Dedicated quantum budgets rose almost nineteenfold from 2019 to 2025, and later years use a wider measure

MSIT reports a dedicated quantum budget, which counts only the programmes set up for quantum. That series began at ₩10.6 billion in 2019 and reached ₩32.6 billion in 2021. The ministry secured ₩128.5 billion for 2024, up 32.7 per cent, and for 2025 it reported ₩198 billion spread across 24 programmes, a further rise of 54.1 per cent. That is almost 19 times the 2019 figure.

Growth still fell short of the plan. In April 2024 MSIT said it aimed to more than double the budget for 2025, and the rise it got was 54.1 per cent, large but well short of that.

The figures for 2026 and 2027 come from a different table. MSIT’s 2027 budget release lists quantum among its next-generation strategic fields, at ₩0.30 trillion for 2026 and ₩0.37 trillion proposed for 2027, a rise of 24.3 per cent. That category is wider. So the step from ₩198 billion in 2025 to ₩300 billion in 2026 cannot be read as growth, because the two numbers do not count the same things. The 2027 sum is still only a proposal before the National Assembly.

Three years are missing. No stated total was found for 2020, 2022 or 2023, and the table leaves those years blank rather than filling them in by working back from the growth rates.

The ₩645.4 billion flagship is the largest single quantum programme, and its hardware dates keep moving

The biggest single line is the Quantum Science and Technology Flagship Project. After a review of whether the plan was sound, MSIT set its cost at ₩645.4 billion over 2025 to 2032, of which ₩621.3 billion is national money. Phase one runs to 2029 at ₩360.4 billion. Phase two, worth ₩285.1 billion, will be worked out again once phase one has reported.

Its goals are specific. With neutral atoms, which are single atoms held in place by beams of laser light, it aims for 1,000 physical qubits with error correction running on them. On superconducting chips it aims for three logical qubits. A logical qubit is one reliable qubit built out of many error-prone ones, the idea behind quantum error correction. MSIT calls it the first national research programme to try for such a chip.

Part of the money is disputed. In August 2026 the JoongAng daily reported that a neutral-atom line had asked for ₩100 billion and received only about ₩30 billion, which it described as a cut. MSIT replied that “예산이 삭감된 것이 아닙니다”, meaning “the budget was not cut”, and said the scope of the work had changed instead. Both sides agree on the smaller sum.

The hardware dates have slipped with each plan. The 2021 strategy promised a 50-qubit Korean system built between 2022 and 2024. The 2023 strategy moved that to 2027 and set a goal of 1,000 qubits for 2031, leaving four years to grow from the first machine to the second. In August 2026 the seven SEED projects set a new goal of a 100-qubit processor that can correct its own errors by 2029.

Three quantum clusters were named in September 2026, and their budget is a press estimate

On 1 September 2026 the Quantum Strategy Committee named three clusters, each a zone where colleges, labs and firms in one region work together on one field. Seoul takes computing. Jeonnam and Gwangju take communications, and the south-east, meaning Busan, Ulsan and Gyeongnam, takes sensing. Work begins in 2027.

MSIT’s release gives no sum. The figure in circulation is ZDNet Korea’s, which reported “각각 5년간 1,000억원 전후가 지원된다”, or “each will receive around ₩100 billion over five years”. It added “지자체 펀드 매칭은 현금 30%다”, meaning matching funds from local government are 30 per cent cash. Until the ministry confirms it, around ₩100 billion each is a press estimate rather than a budget line, and it should not be added to the national totals above.

Korean buyers have bought five quantum computers, and none has disclosed a price

Five machines have been bought or contracted by Korean organisations. Four came from foreign makers and one from a Korean start-up. None of the buyers or sellers has published a price. So any figure attached to these machines is either a project budget or a guess, and the guide to what quantum computers cost sets out the prices that are public elsewhere.

Yonsei University was first. IBM announced the deal in November 2021 and deployed the machine at Yonsei’s Songdo campus in Incheon on 19 November 2024, three years after that first announcement. It is an IBM Quantum System One with a 127-qubit Eagle processor, which IBM calls the first in Korea. The Korea Herald gave a budget of ₩17.8 billion, including ₩10 billion of state money. That sum covers the whole computing centre, not the machine. Yonsei plans to fit IBM’s newer Nighthawk processor into the same machine in November 2026, Seoul Economic Daily reports, which is an upgrade rather than a second computer.

Chungbuk National University in Cheongju bought a five-qubit IQM Spark from the Finnish maker, in a sale first announced in December 2024. IQM said at the time that it was “the first commercial quantum computer to be installed through the Korean government’s official procurement process”. The machine went online in May 2025. IQM says delivery and installation took four months, and in the same release it said it would open an office in Seoul.

KISTI, the state supercomputing institute in Daejeon, chose IonQ in July 2025, and IonQ described the funding only as a “multi-million dollar government award”. The two sides finalised the deal on 23 December 2025 for a 100-qubit Tempo system, linked to KISTI-6, which IonQ calls “the largest high-performance computing (HPC) cluster in Korea”. IonQ booked it as sold in its fourth-quarter results. Neither side gave a price. DongA Science reported in September 2026 that installation was under way and that service should start in the first half of 2027.

One of the five was made in Korea. QAI, a Seoul data-centre operator, said in January 2026 that it had bought the Kreo SC-20, a 20-qubit superconducting machine made by the Korean start-up SDT. SDT built it with design and manufacturing from its partner Anyon Technologies. It sits in a data centre in Cheongdam-dong, Gangnam. In the same announcement QAI said commercial service would start in the first quarter of 2026. SDT calls it the first commercial hybrid system in Korea.

SDT is now a buyer as well. In September 2026 IonQ said it would supply SDT with a Superion 256 system and a quantum memory module, for use with an unnamed SDT customer in Korea. IonQ’s chief executive called it “this purchase”. SDT also plans what IonQ’s release calls a “dedicated quantum manufacturing and system-integration facility in Gumi”, where IonQ systems will be assembled.

Korea’s home-built machines have reached 20 qubits, and the 50-qubit machine has been shown only as a model

The state also pays to build machines at home. KRISS, the national standards institute, leads a programme of about ₩49 billion over 2022 to 2026 to build a 50-qubit superconducting computer, and that programme is due to end this year. It has shown a 20-qubit machine, which the science minister was briefed on during a visit in August 2025. We found no demonstration of a working 50-qubit machine. When the President viewed it on 12 August 2026, Bloter’s photo caption called it a 50-qubit “모형”, a model.

Two other home builds exist. ETRI, the state electronics and telecoms lab, runs a machine built on a processor from the Dutch firm QuantWare, according to QuantWare, which gives no qubit count. We found no ETRI release of its own. On 1 October 2026 the first vice minister of science watched a demonstration of an ion-trap prototype at Seoul National University, built with Korean chips, controls and software.

Busan’s IBM System Two is still only a plan, and memoranda and cloud deals are not purchases

Several announcements read like sales and are not. In January 2024 IBM and Korea Quantum Computing (KQC) said they planned to put an IBM Quantum System Two in Busan by 2028. KQC’s Nasdaq merger announcement of 7 October 2026 does not mention IBM at all, and says instead that KQC uses D-Wave’s machines through the D-Wave Leap cloud service. Until a contract appears, it is a plan.

Memoranda fall short too. Yonsei and Incheon signed one with D-Wave in June 2025 that covers “efforts towards the acquisition” of an Advantage2 system, and no purchase has been announced in the sixteen months since. QAI’s deal with the German firm QUDORA for an ion-trap machine is a feasibility study. IonQ’s links with Busan and Sungkyunkwan University give cloud access, not hardware on site.

SK Telecom’s ID Quantique stake is the largest documented corporate bet, and it is now held as IonQ shares

Korean firms put money into quantum before the state scaled up. In 2018 SK Telecom raised its stake in the Swiss encryption firm ID Quantique from 4.6 per cent to 65.6 per cent, taking control of the company. Its annual report puts the cost at ₩55.2 billion in cash and ₩5.7 billion in kind. In 2025 IonQ bought ID Quantique and paid the sellers in IonQ shares, and an IonQ prospectus lists 2,864,501 of them for SK Square and 859,778 for an SK Telecom fund.

Korea’s carmakers and chipmakers joined foreign rounds instead. Hyundai and Kia signed subscription agreements in the $350 million private placement that took IonQ public in 2021, but the filed agreements leave their own share counts blank. Samsung’s Catalyst Fund co-led IonQ’s $55 million round in 2019 with Mubadala Capital. That is the whole round. Samsung’s part was not disclosed.

Others followed the same pattern. Korea Investment Partners joined Quantinuum’s raise of about $600 million in 2025, but that is the size of the whole round and its own cheque was not disclosed. Samsung Next is among the named investors in Classiq’s $110 million round of 2025. LG Electronics is listed among Pasqal’s investors, with no amount given.

State funds hold listed shares too. At 30 June 2026 the National Pension Service reported 73,376 IonQ shares worth about $3.9 million, and the Korea Investment Corporation reported 132,395 shares worth about $7.1 million. For the pension fund that is about 0.003 per cent of the roughly $155 billion market value of all the US shares in the same filing.

Korean start-ups raise smaller sums. SDT took a ₩20 billion pre-IPO round in 2024, led by ₩10 billion from Shinhan Venture Investment. In January 2026 it announced “총 300억 원 규모의 투자”, or “investment totalling ₩30 billion”. KQC is taking another route. Its merger with Charlton Aria, a shell company listed on Nasdaq, values KQC at $80 million before the deal and the combined company at about $215 million. Closing is expected in the first half of 2027.

Defence has a quantum council and a memorandum, but no published budget

The armed forces are part of the effort. In June 2024 MSIT launched a civil-military council whose members include the defence ministry, the procurement agency DAPA, the Agency for Defense Development and all three services. Firms such as Hanwha Systems and LIG Nex1 sit on it too. In October 2025 MSIT signed a memorandum with the Air Force. No defence quantum budget line has been found, so this spending cannot be sized from public sources.

Machine prices, three budget years and a working 50-qubit computer are missing from the record

The record has three large holes. Not one machine price is public, so nobody outside the deals can say what Korea pays for quantum hardware or how that compares with buyers elsewhere. Budget totals for 2020, 2022 and 2023 were not found. The first five-year quantum plan of January 2026 targets 2,000 quantum companies and 10,000 quantum workers by 2035, but its release gives no total budget.

The hardware also lags the plans, since a 50-qubit machine first promised for 2024 has so far appeared only as a model. The pace of the plan can be tested once with the numbers here. The government’s ₩2.4 trillion over thirteen years averages about ₩185 billion a year, and the dedicated budget for 2025 alone was ₩198 billion. On its face, the state is ahead. But the 2023 strategy does not say which programmes its ₩2.4 trillion counts, so the two figures may not measure the same thing. The pace cannot be judged from public sources. The firm facts remain the budgets already voted and the five machines already bought or contracted.