Anderon and IBM’s $2 Billion Quantum Foundry Is a Bet on the Shape of an Industry

A standalone IBM company has finalised $1 billion in federal money to make quantum wafers for customers across the industry. The announcement frames the award as accelerating research and development; what the structure implies is worth unpacking separately, and carefully.

What Happened

On 16 September 2026, Anderon LLC — a standalone IBM company operating a pure-play quantum foundry headquartered in Albany, New York — finalised a $1 billion CHIPS award with the United States Department of Commerce. The agreement was first announced as a letter of intent in May 2026. No milestones or conditions attached to the award are specified in the announcement.

A separate $1 billion from IBM backs the venture, bringing total capitalisation to $2 billion split evenly between public money and Anderon’s owner. The facility operates a 300-millimetre fabrication plant producing quantum wafers for external customers across the quantum industry ecosystem — not solely for IBM’s internal needs. Current products support superconducting qubit arrays, quantum input/output signalling, and readout signal chain components; plans exist to expand to other quantum modalities. First quantum wafers are already running through the facility. No external customer is named in the announcement, and no figures for jobs, capacity, wafer volumes, yield, timelines, or revenue are given.

Mukesh Khare, Anderon’s chief executive, said: “This agreement with the Department of Commerce strengthens our ability to deliver the manufacturing scale the quantum industry needs. We’re building a foundation that will let quantum move from research breakthroughs to real-world impact.” Jay Gambetta, Director of IBM Research and an IBM Fellow, put the dependency plainly: “IBM’s quantum roadmap requires a manufacturing foundation that can scale at the pace our work demands. Anderon’s focus on wafer fabrication gives us exactly that and strengthens our ability to execute and accelerate progress toward the next era of quantum computing.”

The Structural Read

The foundry model exists because of one specific structural fact about classical semiconductors: far more organisations can design a chip than can afford to build a fabrication plant. That asymmetry — many designers, few fabs — is what made the fabless model possible, and fabless is what allowed chip design to scale from a handful of vertically integrated giants to thousands of firms. The separation of design from manufacturing did not happen because it seemed elegant; it happened because the economics of the fab became so demanding that most design teams could not justify them.

Building a pure-play quantum foundry is therefore a wager that quantum will develop that same ratio. It is worth being precise that this is a wager and not an observation, because it is not how quantum has worked so far. The field has been dominated by vertically integrated efforts in which each laboratory builds its own stack from end to end. The bet embedded in Anderon’s facility is that such an arrangement stops being the norm — that quantum acquires a population of designers who want to produce devices without owning a fabrication line. Whether that population materialises, and how quickly, is not stated in the announcement and is not predicted here.

The detail that carries more structural information than the billion-dollar headline is “300-millimetre.” That is the standard wafer size of the classical semiconductor industry. Committing to it means committing to the existing toolchain, metrology, handling equipment, and supplier base built over decades for conventional chips — not to bespoke laboratory apparatus assembled for a single research group. That is the practical difference between a research programme, which is limited by how many devices a skilled team can produce, and a manufacturing industry, which is limited by how many times a process can repeat reliably.

It is a meaningful signal about intent and about the stage the technology is entering. It is not evidence of results. Adopting standard equipment does not establish that any process works at scale. No yield figure, volume figure, performance figure, or reliability figure has been reported, and none is offered here.

The structure of the arrangement is the genuinely unusual part. IBM is capitalising a supplier to its own competitors. Anderon is a standalone IBM company, funded with a billion dollars of public money and a billion of IBM’s own, and its stated purpose is to serve external customers across the quantum ecosystem — a description that necessarily includes organisations competing directly with IBM’s quantum programme. An arrangement of that shape only makes structural sense if the binding constraint is ecosystem-wide manufacturing capacity rather than IBM’s own internal capacity. IBM is judging, in effect, that the whole field is held back by the absence of a shared manufacturing layer, and that funding that layer — even for rivals — is a better use of capital than withholding it.

This is a structural reading of the arrangement, not a claim about IBM’s motives, which are not disclosed beyond the two quotations in the announcement. No claim is made that any particular competitor will or will not use the facility. The same reasoning is visible elsewhere — an incumbent funding a layer it does not fully control, on the view that the absence of that layer is what holds the whole field back. It is a recognisable pattern, not a unique one.

Three Implications

The Bottom Line

Anderon’s $2 billion quantum foundry is not, at its core, a manufacturing story — it is a structural prediction. The prediction is that quantum will disaggregate the way classical semiconductors did: many designers, few fabs, a shared supply layer that no one player can afford to own alone. The 300mm wafer choice says the intent is serious and the toolchain is real. IBM funding a supplier to its own competitors says the binding constraint is ecosystem-wide, not internal. Whether the designer population that would justify a pure-play foundry actually emerges is the only question that matters, and this announcement does not answer it.

91,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.

No milestones or conditions attached to the CHIPS award were specified in the announcement, and none are stated here. No external customer has been named, and none is named in this article. No figures for jobs, capacity, wafer volumes, yield, timelines or revenue were given, and none are asserted. Nothing here reports a performance, yield or reliability result. The adoption of standard 300-millimetre tooling is a signal about manufacturing intent and does not by itself establish that any process works at scale. This article offers no forecast of quantum computing timelines, capability, commercial usefulness or so-called quantum advantage, and makes no claim about what quantum computers can presently accomplish. It does not predict whether the venture succeeds or whether the award is disbursed in full, and takes no position on the CHIPS programme, its merits, its politics or any other award made under it. The observation that IBM is capitalising a facility serving external customers, including organisations competing with its own quantum programme, is this publication’s structural reading of the arrangement. IBM’s motives are not disclosed beyond the statements quoted, and no claim is made that any particular competitor will or will not use the foundry. Anderon is described throughout as a standalone IBM company, which is how the announcement describes it; it is not characterised here as a spin-off, joint venture or subsidiary. IBM is a publicly listed company. This is business analysis, not investment advice, no view is expressed on any security, and no recommendation is made.

Sources: prnewswire.com