IBM Is Building Quantums TSMC — If the Government Lets It

IBM says it will spend $10 billion over five years to build what it calls the world's first large-scale, fault-tolerant quantum computer by 2029 IBM Quantum blog. The press release has the number. The more interesting question is what IBM is actually building with the other half of that money.

IBM Quantum Starling, arriving in 2029, will run 100 million quantum operations on 200 qubits IBM Quantum blog. IBM Quantum Blue Jay, further out, targets one billion operations on 2,000 qubits. Those are the headline numbers. The infrastructure underneath them is Anderon — a 300-millimeter quantum wafer foundry in Albany, New York, backed by $1 billion in CHIPS Act incentives and another $1 billion in IBM cash, intellectual property, and infrastructure The Quantum Insider. IBM calls Anderon "America's first pure-play quantum foundry," positioning it as a TSMC for quantum: a manufacturing substrate that external chip designers can access, not just a proprietary IBM research facility.

That framing is the story. Whether it is true is a different matter.

The press release says Anderon will be "modality-agnostic" — meaning it would process quantum chips based on superconducting qubits, trapped ions, photonics, spin, or neutral atom architectures from any vendor willing to pay for wafer access IBM Newsroom press release. Quantum Computing Report, which has covered the quantum foundry space longer than most, calls the Anderon model "replicating the exact open-foundry model that made TSMC indispensable to the classical computing ecosystem." If that holds, Anderon does for quantum hardware what TSMC did for classical chips: it separates design from manufacturing, lets startups go fabless, and shifts competition to who can design the best chips rather than who can afford to build a fab.

The SEC filing from May 28 contains the terms that will answer whether that framing survives contact with reality SEC Filing. The equity splits between IBM and the federal government, the governance rights each party retained, and the technology-access conditions attached to the CHIPS Act money are the documents that matter. They determine whether Anderon is a genuine open foundry or a federally subsidized IBM subsidiary with an open-access marketing line.

IBM has deployed more than 90 quantum systems globally, more than the rest of the industry combined, according to its SEC disclosure Reuters. Its quantum network includes more than 325 Fortune 500 companies, startups, universities, and government agencies. Qiskit, IBM's open-source software development kit, claims 70 percent of quantum developers. Since 2017, IBM's quantum program has signed more than $1.1 billion in contracts IBM Newsroom press release. These are real numbers from verifiable filings. They describe a company with the largest deployed fleet, the most used software, and a substantial contracted backlog — all in the NISQ era, before fault tolerance.

The question is what comes next. IBM's CEO Arvind Krishna said in the press release that "the quantum era has started." Google CEO Sundar Pichai said last year that practical quantum computers are five to 10 years away Reuters. These statements are not compatible, and they cannot both be right on the timelines they imply. Quantum computing has been five to 10 years away since the 1990s. IBM has a financial interest in the era arriving on schedule. Google has a financial interest in the transition being orderly. Neither data point is independent.

The investment total warrants scrutiny too. $10 billion over five years sounds large until you account for what it actually funds: R&D, capital expenditure, ecosystem partnerships, manufacturing scaling, and mergers and acquisitions. The CHIPS Act award for Anderon, at $1 billion, is more concrete: it is new money, it is tied to a specific facility, and it is matched by IBM's own $1 billion contribution The Quantum Insider. That is a real investment in a real facility. The $10 billion figure that leads every story is less an investment decision than a communications strategy.

What IBM is actually building is infrastructure for the quantum ecosystem, not just for itself. The Anderon fab in Albany is the more consequential bet. If Anderon is genuinely open — if external chip designers can access 300-millimeter wafer capacity on the same terms as IBM's internal team — it changes the competitive dynamics of quantum hardware. Startups that could not afford to build fabs can now design chips and send them to Albany. The moat shifts from manufacturing to design. IBM positions itself as the TSMC of quantum, collecting manufacturing margins instead of competing on chip architecture.

If Anderon is not genuinely open — if the federal money came with IP-sharing obligations, preferential access terms, or governance rights that favor IBM's own roadmap — then Anderon is an IBM facility that happens to have received government funding. The press release language about openness does not settle that question. The SEC filing terms do.

IBM filed its 8-K on May 28. The equity, governance, and IP terms are in that document. They are the next thing to read.