IBM's new Anderon subsidiary aims to be a quantum chip foundry for the whole industry

[Maisie84]

Forbes has an analysis piece from TIRIAS Research on Anderon, a new IBM subsidiary that wants to bring the semiconductor foundry model to quantum computing. It is backed by a $2 billion effort. The Commerce Department finalised a $1 billion CHIPS Act award, and IBM is matching it with $1 billion plus IP, manufacturing assets and staff. The award itself was announced in September, so the news here is the deeper look at the plan

The idea is that Anderon will run a specialised 300mm foundry making quantum wafers for IBM and outside customers. That way, quantum developers would not need to build their own fabs. It will use the Albany NanoTech Complex in New York rather than build a new site. That complex covers 1.65 million square feet, is run by the nonprofit NY CREATES, represents more than $25 billion of investment and also hosts a $10 billion High-NA EUV initiative

Anderon will start with superconducting qubit wafers, along with the input, output and readout circuitry that surrounds them. Other qubit types may follow later. Analyst Jim McGregor calls it only the second commercial quantum foundry, after GlobalFoundries, which focuses on photonics. He also notes IBM's move into spin qubits through its HRL acquisition

McGregor is realistic about the limits. Commercial volumes in quantum are tiny, so a foundry will not fill its lines the way a smartphone chip maker does. His argument is that quantum computing cannot become a large commercial industry without this kind of shared manufacturing. The same model that helped classical chips grow could speed up quantum too

It is worth noting that TIRIAS has consulted for both IBM and GlobalFoundries, which the article discloses. That does not make the analysis wrong, but it is good context. The foundry idea itself makes a lot of sense to me. Startups spend fortunes on clean rooms before they can even test a design, and a shared fab could lower that barrier a lot

Would quantum startups trust a foundry owned by IBM, one of their biggest competitors? And is a shared foundry model the key to growing the industry?

Anderon Extends Ibms Semiconductor Legacy To Quantum Foundry Serviceswww.forbes.com

[Leo68]

Shared manufacturing is how every big chip industry grew. Fabless companies like Nvidia and AMD would not exist without foundries. Quantum needs the same split between designers and makers

Someone had to go first. Quantum is still waiting for its fabless generation

[DistantLeopard]

Starting with superconducting wafers makes sense, since that is IBM's own strength. The question is whether they will be any good at spin qubits or other types. Each one needs quite different processes. HRL might help there. Each qubit type is almost a separate industry in terms of manufacturing

[InferenceLoop]

Trust is the big question. Handing your chip designs to a competitor's subsidiary is a huge leap for a startup

TSMC works because it does not make its own branded chips. Anderon will need very strong firewalls to win outside customers. Intel struggled with exactly this when it tried to sell foundry services

[PulsarElk]

Having a $10 billion High-NA EUV programme next door is a nice bonus. Quantum chips do not need cutting edge lithography today, but being near that expertise helps

Shared tools and staff make the whole site stronger. Albany is turning into something special. Clusters like that attract talent from all over the world. Engineers want to be where the action is