Commerce Finalized a $1 Billion CHIPS Award to IBM's Anderon on 16 September 2026 for a 300 mm Quantum Wafer Foundry in Albany; the Release Calls It Pure-Play, and Names No Customer Other Than IBM
In 1987 Morris Chang founded Taiwan Semiconductor Manufacturing Company on one rule: the company would manufacture chips for others and never sell a chip under its own name. The rule was commercial before it was technical. Fabless design houses would hand a foundry their most valuable asset, the mask set, only if the foundry could not turn around and compete with them. Everything the phrase pure-play foundry has meant since then flows from that separation between the manufacturer and the product; TSMC's own dedicated-foundry page still states that the company pioneered the pure-play foundry business model in 1987.
On 16 September 2026 Anderon LLC, an IBM company, announced that its $1 billion award under the CHIPS and Science Act with the US Department of Commerce is final. The release uses the phrase pure-play in its headline and three times in its text. This post takes the phrase seriously and tests it against what the release, the letter of intent of 21 May 2026 and the Commerce notice of the same day actually state.
Practical takeaway. A $1 billion federal award plus $1 billion from IBM now backs a 300 mm quantum wafer fab in Albany that is already running wafers. The word pure-play in the release is a business-model promise. Anderon's website adds a pledge never to compete with a customer's products; the releases give no external customer, no wafer contract and no governance term that describes how customer designs are kept apart from IBM's own program. Read the award as federal support for superconducting-wafer R&D on a running line; read the pure-play claim as a question the public documents do not yet answer.
What the 16 September 2026 release states, and what it leaves out
The release places Anderon in Albany, New York, as a 300-millimeter pure-play quantum wafer foundry, backed by the $1 billion CHIPS award and an additional $1 billion from IBM. It says the first quantum wafers are now running through the manufacturing facility. It lists three wafer products: high-performance superconducting qubit arrays, quantum input/output signaling, and readout signal chain components, and says the company plans to expand to other quantum modalities. Two executives are quoted. Mukesh Khare, Anderon's chief executive, speaks of delivering the manufacturing scale the quantum industry needs. Jay Gambetta, director of IBM Research, says IBM's quantum roadmap requires a manufacturing foundation that can scale at the pace IBM's work demands, and that Anderon's focus on wafer fabrication gives IBM exactly that.
The release does not name a customer other than IBM. It does not state the award's disbursement schedule, milestones, clawback terms or any equity component. It does not say how much of the $2 billion is capital expenditure and how much is operating R&D. Those omissions are ordinary for a press release, and they are the omissions that matter for the claim under test.
What the 21 May 2026 letter of intent added
The May release from IBM and Commerce is more detailed. It says IBM contributes $1 billion of cash into Anderon along with significant intellectual property, assets and a skilled workforce, with additional investors expected as Anderon grows. It describes the 300 mm processes on offer: superconducting wiring, through-silicon vias and bumps, backed by dedicated process design kits and in-line wafer testing. It says Anderon will first support wafer fabrication for superconducting qubit and supporting electronics wafers, with the goal to expand into other modalities, and that the foundry is meant to offer wafer fabrication for multiple quantum technology vendors across the world. Secretary of Commerce Howard Lutnick and Bill Frauenhofer, executive director of Semiconductor Investment and Innovation, are quoted on jobs and leadership.
The Commerce notice published by NIST on 21 May 2026 frames the whole program: nine letters of intent worth $2.013 billion, two domestic quantum foundry companies and seven quantum computing companies. IBM's $1 billion is described as planned funding to establish a new quantum foundry subsidiary for quantum-grade superconducting wafers. GlobalFoundries' $375 million is described as planned funding for a secure domestic quantum foundry for multiple modalities: superconducting, trapped ion, photonic, topological and silicon spin. The seven computing companies, Atom Computing, Diraq, D-Wave, Infleqtion, PsiQuantum, Quantinuum and Rigetti, receive between $38 million and $100 million each for discrete engineering problems.
Test one: what the documents say about ownership and customer separation
The pure-play model TSMC pioneered has one defining part, the manufacturer sells no competing product, and one practical corollary that customers rely on, that their designs stay walled off from any product group. On the defining part the documents are thin. IBM sells quantum computers. Both releases call Anderon an IBM company and neither states the ownership structure; the May release says IBM contributes $1 billion of cash, intellectual property, assets and workforce and expects additional investors as Anderon grows, which is consistent with IBM as the sole disclosed investor today and with a wider shareholder base later. The Gambetta quotation presents IBM's own roadmap as the need Anderon meets. A foundry inside a company that also sells the end product is how Samsung Foundry and Intel Foundry operate in classical silicon, and both have spent years persuading customers that the wall between foundry and product group holds.
The corollary, customer separation, is addressed in one sentence outside the releases. Anderon's website tells prospective customers to innovate with confidence, knowing their manufacturing partner will never compete with their products. That is the right promise, and it is a promise, not a mechanism: neither the website nor the releases describe how customer designs and process data are kept apart from IBM's own program, or on what terms non-IBM customers are served. None of that is part of the pure-play definition itself; it is what a customer competing with IBM would want to know before handing over a mask set. An unnamed external customer does not prove a captive model, and a single public contract would not prove independent governance either. The honest reading is that the pure-play claim is an open question about customer protection, and the verified fact is that a 300 mm superconducting-wafer line at an IBM company is now running wafers with $1 billion of federal support.
Test two: is it a foundry, or an R&D program?
The award is a CHIPS Research and Development award, administered by the CHIPS R&D Office rather than the manufacturing incentives program that funded the large logic and memory fabs. The May notice says so plainly, and the September release says the award will accelerate Anderon's R&D efforts. The May release describes what Anderon will do once definitive agreements are in place; the September release reports that the first quantum wafers are running, on a site with, in Anderon's own words, billions in prior investment. The Albany NanoTech complex has hosted IBM's 300 mm research line for two decades. How the $2 billion divides between expansion, equipment and operating R&D is not stated in either release. What the documents do establish is a running 300 mm line with federal R&D money attached, which is a different starting point from the greenfield 5,300-square-meter quantum chip factory that Quantum Foundry Copenhagen plans for 2027 with Novo Nordisk Foundation money, where the building itself is the deliverable.
Why the superconducting bet is narrower than it looks
Commerce split its foundry money by modality. GlobalFoundries' $375 million is meant to serve five architectures. Anderon's $1 billion is for superconducting wafers first, with other modalities as an intention. Superconducting qubits are the modality IBM, Google and Rigetti build, and the one NEC abandoned in March 2026, twenty-seven years after its Tsukuba lab made the first superconducting qubit. One reading of that split, and it is an interpretation rather than anything Commerce has said, is that the largest foundry award went where 300 mm wafer processes already exist. Of the seven computing companies in the portfolio, two (D-Wave and Rigetti) build superconducting machines; the other five build other modalities: Atom Computing and Infleqtion neutral atoms, Diraq silicon spin, PsiQuantum photonics, Quantinuum trapped ions. The Commerce notice describes funding objectives for each and assigns no company to either foundry. Some of the seven have existing foundry relationships of their own (PsiQuantum, for one, has long described GlobalFoundries as the manufacturer of its photonic chips); none of the Commerce or IBM documents examined here names any of the seven as an Anderon customer.
The seven computing awards carry a condition the foundry paragraphs of the Commerce notice do not mention. After listing the seven companies, the notice states that the Department will receive a minority, non-controlling equity stake in each company as a condition for receiving the funds, to enhance the return for the taxpayer. Rigetti and D-Wave disclosed in early September the equity terms on their $100 million awards. The notice's equity sentence sits in the computing-company section and is worded for those seven; it does not say whether the two foundry awards carry the same condition, and neither IBM release mentions equity. Silence in a press release is not a term, so the accurate statement is narrow: the equity condition is public for the seven computing companies and unstated in public for Anderon. Whether any of the seven will buy wafers from Anderon is also unstated.
How Quentir Reads It
The pure-play question is a governance question, and it is the one the public money makes legitimate to ask. A foundry is a trust arrangement. Its customers hand over designs that embody years of work, and in quantum hardware the design of a qubit array, its Josephson junctions and its wiring stack is the whole company. A customer who competes with IBM in superconducting machines has to weigh whether Anderon's process engineers, its yield data and its process design kits sit on IBM's side of a wall or on the customer's side. The releases do not say. Commerce, which is paying half the bill, has an interest in an answer, because a captive fab that serves one roadmap is a smaller return on $1 billion than a shared national wafer base.
The humane stake is local and concrete. Albany's 300 mm line sits in a research complex New York State has funded for two decades, and the award adds $1 billion of federal money to it. The people who will work there are hired for a foundry, and a foundry with one customer is a different employer from a foundry with ten; which of the two Albany gets is the same open question in a different form.
For readers following how these federal instruments accumulate, from the NEC exit to the Rigetti and D-Wave equity terms to this award, every post linked here sits in the same archive, and the All-access membership is the single subscription that covers all of it, with the Signature Report and Signature Brief editions included.
The fact to watch is a name, and after the name, the terms. The first publicly disclosed non-IBM wafer contract will show that Anderon serves outside customers. Whether the manufacturer sells no competing product is a separate matter, settled by what IBM does with its own quantum systems rather than by any single contract. And how customer designs and process data are kept apart from IBM's program is a third question, the one a customer competing with IBM would ask first, and the public documents do not answer it yet. Until then, the accurate description of what Commerce finalized on 16 September 2026 is a $2 billion superconducting-wafer R&D program at an IBM company's fab that is open, in principle, to others.
Sources: IBM Newsroom, "Anderon, an IBM Company, Finalizes Agreement with the U.S. Department of Commerce for a $1 Billion CHIPS Award to Accelerate R&D for U.S.-Based Pure-Play Quantum Foundry", 16 September 2026. IBM Newsroom, "IBM and U.S. Department of Commerce Announce America's First Purpose-Built Quantum Foundry, Supported by Proposed $1 Billion CHIPS Award", 21 May 2026. NIST / US Department of Commerce, "Department of Commerce Announces Letters of Intent With 9 Companies for $2 Billion to Accelerate U.S. Leadership in Quantum Computing", 21 May 2026. Earlier Quentir coverage is linked inline. All sources read on 17 September 2026.
Published intelligence, built to inform your own decisions. Published: September 17, 2026.