America bets ~$2 billion on its first quantum chip foundry — and hands IBM the keys
IBM and the US Commerce Department are building Anderon, America's first pure-play quantum wafer foundry, in Albany — $1B of IBM cash plus a proposed $1B CHIPS award. It's industrial policy for the post-GPU era: own the fab before the market exists.
IBM and the U.S. Department of Commerce are building Anderon — America's first pure-play quantum chip foundry. It's a standalone company headquartered in Albany, New York, capitalized with $1 billion of IBM cash plus IBM's intellectual property, equipment, and staff, and supported by a proposed $1 billion CHIPS Act award — roughly a $2 billion bet that the country that manufactures quantum wafers will hold an edge in whatever comes after the GPU era.
What it actually is
"Pure-play" is the load-bearing phrase. Anderon's entire purpose is producing quantum wafers — not adapting classical semiconductor lines for occasional quantum runs. It will operate as a 300-millimeter quantum wafer fab, starting with superconducting-qubit and supporting-electronics wafers and expanding to other quantum modalities over time. Crucially, it's meant to fabricate for the whole field: a contract foundry that multiple quantum vendors can use, the way TSMC builds chips it doesn't design.
That's the strategic move. Today every serious quantum effort builds its own fragile, low-yield fabrication in-house. A shared foundry turns a craft into an industry — standardized processes, higher yields, a supply chain other companies can build on without each reinventing the hardest part.
Our read
This is industrial policy aimed at the next compute era, and the framing is deliberate. Washington watched the leading edge of classical chipmaking migrate to Taiwan and spent the last few years and tens of billions trying to claw some of it back. Anderon is an attempt to not repeat that — to own the manufacturing base of quantum before the market exists, rather than after it's already concentrated somewhere else. It's the same sovereignty logic driving Europe's deeptech bet and the push to rethink chip design itself: whoever controls the fab controls the industry's gravity.
For IBM, the deal is shrewd. It converts a cost center — quantum fabrication few customers were paying for — into a government-subsidized, standalone business that also happens to entrench IBM's quantum stack as the reference platform. Put $1 billion in, get matched by the public purse, and own the foundry everyone else's quantum ambitions route through. That's a strong position if quantum pays off.
Here's the catch: it's a foundry for a market that barely generates revenue yet. Useful, fault-tolerant quantum computing is still measured in "years away," and the timeline has slipped before. Building 300mm quantum capacity now is a bet that demand shows up on roughly the schedule the field keeps promising — and the history of quantum is a history of promised schedules. If error correction and useful scale arrive late, Anderon is expensive capacity waiting for customers; if they arrive on time, it's the most important factory in American computing. Governments can afford that asymmetric bet in a way companies can't, which is exactly why it's structured as a CHIPS award rather than a pure private venture.
The number to watch isn't the $2 billion. It's yield and cadence: how quickly Anderon can turn out quantum wafers that work, for vendors other than IBM. A foundry only matters if the rest of the industry actually shows up to use it. That turnout, more than any ribbon-cutting, will tell you whether the U.S. just built the TSMC of quantum or a very advanced bet on a market that hasn't arrived.
