PsiQuantum's $1B bet: A useful quantum computer by 2027?

PsiQuantum isn’t building a slightly better quantum chip. The Palo Alto-based startup, founded by four physicists in 2016, is skipping the prototype stage entirely and betting the house on a million-qubit, fault-tolerant machine the size of a data center. That machine, as CEO Victor Peng describes it, will use photons—particles of light—guided through a maze of optical switches and beam splitters inside stainless-steel cabinets cooled by liquid helium to near absolute zero.

It’s an audacious plan in a field littered with overpromises. But PsiQuantum has the receipts to make you pay attention. Last year, it raised a cool $1 billion and broke ground on a Chicago facility backed by local governments. A second site in Australia is slated to be hardware-ready by 2027. The company is also one of just two—alongside Microsoft—to reach the final evaluation stage of a rigorous Pentagon program meant to separate the quantum contenders from the pretenders. And they’ve convinced a major, unnamed chipmaker to fabricate their components in existing semiconductor fabs, a detail that separates them from competitors building bespoke, artisanal hardware.

The company’s promises are bold. Vice President of Quantum Applications Philipp Ernst claims they can slash the time needed to predict how cytochrome P450 enzymes break down drugs from over a decade to just four minutes. But here’s the reality check: quantum advances are notoriously opaque and incremental. What works in a peer-reviewed paper often fizzles in a real-world system. PsiQuantum is entering its prove-it phase, and the clock is ticking. Hardware readiness by 2027 doesn’t mean a useful computer will be online by then—it means the building will be ready for the machine. There’s a big difference, and the next year will reveal which side of that gap the company lands on.