IBM Commits $10 Billion to Quantum, Launches America’s First Quantum Chip Foundry

June 2, 2026 – IBM announced plans to invest more than $10 billion in quantum computing over the next five years. The commitment spans research and development, capital expenditure, manufacturing, ecosystem partnerships, and mergers and acquisitions.

The bottom line: this is the largest single-company quantum computing investment ever disclosed, and it arrives alongside $2 billion in CHIPS Act funding that makes the U.S. government a direct equity holder in nine quantum companies (including IBM). Whether the money accelerates the path to a cryptographically relevant quantum computer (CRQC) depends on whether the technical milestones hold up under scrutiny.

The announcement consolidates several threads from recent weeks. On May 21, IBM and the U.S. Department of Commerce signed a Letter of Intent to establish Anderon, a standalone subsidiary that will operate as America’s first purpose-built quantum wafer foundry. The Commerce Department is contributing a proposed $1 billion in CHIPS Act incentives; IBM is matching that with $1 billion of its own cash, plus intellectual property, fabrication assets, and staff.

Anderon will be headquartered in Albany, New York, and will operate a 300-millimeter quantum wafer fabrication facility. The foundry will initially produce superconducting qubit wafers and supporting electronics, with stated plans to expand into other quantum modalities. IBM is positioning Anderon as an open foundry, offering fabrication services to outside quantum hardware developers.

The $10 billion figure encompasses the Anderon investment alongside broader spending on IBM’s fault-tolerant quantum roadmap, centered on IBM Quantum Starling, a system the company targets for 2029 that would execute 100 million quantum gates across 200 logical qubits. Starling is designed as the precursor to IBM Quantum Blue Jay, which would scale to 2,000 logical qubits and 1 billion quantum operations by 2033.

The Anderon foundry announcement was part of a broader Department of Commerce package distributing $2.013 billion in CHIPS and Science Act incentives across nine quantum companies. The breakdown: IBM received $1 billion; GlobalFoundries received $375 million for its own quantum foundry capabilities; D-Wave Quantum, Rigetti Computing, Infleqtion, Atom Computing, PsiQuantum, and Quantinuum each received $100 million; and Diraq received $38 million. In exchange for the funding, the government is taking minority, non-controlling equity stakes in all nine companies.

IBM CEO Arvind Krishna said the company expects its partners to demonstrate quantum advantage using IBM quantum computers in 2026, citing recent experiments including a 12,635-atom protein simulation with the Cleveland Clinic and RIKEN, and accurate simulation of magnetic materials reproducing national laboratory data.

IBM says it has deployed over 90 quantum systems globally and signed more than $1.1 billion in quantum contracts since 2017. Its Qiskit software framework is used by approximately 70% of quantum developers, with over 4 trillion quantum circuits executed on IBM hardware to date.

My Analysis

Three elements of this announcement deserve distinct treatment: the Anderon foundry, the $10 billion investment figure, and the Starling roadmap’s implications for the path to CRQC.

Anderon: The TSMC Analogy and Its Limits

The most interesting piece here is Anderon. IBM is positioning it directly as a TSMC analog for quantum: an independent foundry that fabricates chips for multiple customers rather than for a single vertically integrated manufacturer. The classical semiconductor industry’s separation of design and fabrication created one of the most successful industrial models of the past three decades. Applying that model to quantum is a reasonable strategic bet on how the industry might scale.

But the analogy has limits that IBM’s press materials don’t address. TSMC succeeded because it served an enormous addressable market of fabless chip designers across every computing domain. Anderon’s near-term customer base is far narrower. Google fabricates its own superconducting chips at its Santa Barbara facility. Quantinuum and IonQ use trapped-ion architectures with no process commonality with superconducting silicon. Microsoft’s topological qubit program sits on a different fabrication path entirely. As Tom’s Hardware noted, the realistic near-term customers are other superconducting companies: Rigetti, IQM, SEEQC, and a handful of smaller players, plus IBM itself.

The 300mm fabrication capability is real and significant, regardless of the customer question. According to IBM Research Director Jay Gambetta, the shift from 200mm to 300mm wafer processing produces device output 30 times faster by increasing complexity tenfold and tripling production rates. For any superconducting quantum company that has been running small-batch fabrication in university or national-lab cleanrooms, access to dedicated 300mm production with established process design kits could compress development timelines by years.

The sovereignty dimension matters here. I analyzed the strategic significance of domestic quantum fabrication capacity in Quantum Sovereignty, where the IBM-Fraunhofer partnership in Germany was the clearest example of a sovereignty-friendly quantum import model. Anderon goes further: it anchors quantum chip manufacturing on U.S. soil with explicit national security framing. The Commerce Department’s language about positioning the United States to “manufacture most of the world’s quantum wafers” is a direct articulation of supply chain sovereignty applied to quantum hardware, the kind of thinking that has been conspicuously absent from most national quantum strategies, which tend to focus on algorithms and applications while ignoring the physical manufacturing layer.

There is a tension in the open-foundry model, though, that IBM will need to resolve. Quantum hardware startups considering Anderon will have to weigh production access against the risk of sharing process knowledge with their largest competitor. IBM says Anderon will be a standalone company, but the intellectual property flowing into it comes from IBM, and IBM will remain its anchor customer. How much process isolation Anderon can credibly offer to, say, Rigetti will determine whether the open-foundry model is a real industry utility or an IBM cost-sharing arrangement with a broader label.

The $10 Billion: Parsing the Number

The $10 billion figure is large, but it requires context. This is a five-year commitment spanning R&D, capital expenditure, manufacturing, ecosystem partnerships, and M&A. That breadth means the number captures spending that IBM would have made regardless of this announcement: ongoing research salaries, existing facility maintenance, Qiskit development. The marginal new investment above IBM’s existing quantum run rate is not disclosed.

For comparison, IBM’s total R&D spending was approximately $7.4 billion in 2025 across all business lines. A $10 billion quantum commitment over five years ($2 billion per year) would represent a significant share of IBM’s total R&D budget, but only if the spending is actually incremental rather than a relabeling of existing quantum expenditures under a larger umbrella.

The announcement also bundles the $1 billion CHIPS Act incentive and the $1 billion IBM cash commitment to Anderon within the $10 billion total. So $2 billion of the headline figure is the foundry, and the remaining $8 billion covers everything else over five years. That is still a large number, but it lands differently than a $10 billion research investment would.

The market’s reaction (IBM shares rose approximately 3% on the June 2 announcement, after climbing about 12% on the May 21 CHIPS Act news) suggests investors had already priced in most of the signal. The smaller pure-play quantum stocks moved far more violently: D-Wave, Rigetti, and Infleqtion each surged over 30% on the CHIPS Act day, collectively adding nearly $5 billion in market capitalization on $300 million in proposed awards. Those ratios speak to the speculative premium the market places on any quantum validation signal, regardless of the underlying revenue fundamentals. D-Wave reported $2.9 million in Q1 2026 revenue.

The Government as Quantum Shareholder

The equity-stake structure of the CHIPS awards is new and consequential. The U.S. government is not issuing grants; it is taking minority, non-controlling equity positions in all nine quantum companies. This is an industrial policy approach that blurs the line between public R&D funding and venture investing. The government becomes financially aligned with the commercial success of these companies. This arrangement creates incentives but also potential conflicts if policy decisions (export controls, procurement preferences, classification decisions) later affect the value of those stakes.

The $2 billion CHIPS quantum package also represents a portfolio bet across modalities: superconducting (IBM, Rigetti), trapped ion (Quantinuum, Infleqtion), neutral atom (Atom Computing), photonic (PsiQuantum), annealing and gate-model (D-Wave), and silicon spin (Diraq). The Commerce Department framed this as addressing “multiple modalities at once,” which is a sensible hedging strategy given the real uncertainty about which qubit technology will reach fault tolerance first.

What This Means for the Path to CRQC

For PostQuantum.com readers, the question that matters most is whether this investment materially changes the timeline toward a CRQC, a quantum computer capable of breaking current cryptographic systems using Shor’s algorithm.

The answer: probably not directly, but the structural moves matter.

Starling’s 200 logical qubits at 100 million gates by 2029 would be a major milestone for fault-tolerant quantum computing. It would demonstrate that error-corrected quantum computation can scale beyond proof-of-concept demonstrations. But 200 logical qubits is far from the thousands required for cryptographic attacks. Even with the most optimized resource estimates (Craig Gidney’s 2025 paper brought the estimate for factoring RSA-2048 down to approximately 1 million physical qubits), the gap between Starling and a CRQC remains enormous in terms of my CRQC Quantum Capability Framework dimensions.

What Starling does affect is the engineering credibility of the path. If IBM delivers a working 200-logical-qubit system using qLDPC codes (specifically the bivariate bicycle codes described in their arXiv papers), that would validate several capabilities in the Framework simultaneously: quantum error correction at scale, below-threshold operation in a modular architecture, magic state production across modules, and real-time decoding on classical hardware. Each of those is currently undemonstrated at the scale IBM is targeting. Delivering them together by 2029 would compress the capability gap between where the field is today and where it needs to be for a CRQC.

The qLDPC approach is particularly relevant. IBM claims these codes reduce physical qubit overhead by up to 90% compared to surface codes, encoding 12 logical qubits into 144 physical qubits with their gross code. If those numbers hold at scale, it would meaningfully reduce the total physical qubit count needed for cryptographic attacks, though the engineering challenges of implementing qLDPC codes (specifically the non-local connectivity requirements) are substantial and not yet demonstrated at large scale.

IBM’s claim that partners will achieve quantum advantage in 2026 also requires careful parsing. The protein simulation and magnetic materials experiments IBM cites are genuine scientific contributions. But “quantum advantage” as IBM uses the term here means outperforming classical methods on specific scientific problems, which is a very different claim from cryptographic relevance. The former could happen in 2026. The latter remains years to decades away.

The Real Takeaway

None of this changes the core argument that Q-Day predictions are less relevant than the deadlines already set by regulators, insurers, and procurement standards. Organizations should not read IBM’s announcement and either panic or relax about their PQC migration timelines.

What the announcement does confirm is that quantum computing’s transition from laboratory science to industrial infrastructure is accelerating. When the U.S. government takes equity stakes in nine quantum companies and a $300 billion-revenue technology conglomerate commits $10 billion over five years, the “quantum computing will never work” camp loses another talking point. The engineering challenges are immense, and I will continue to track each capability dimension as milestones emerge. But the institutional commitment to solving them has never been larger.