What was shown
The authors did not build or run anything. They calculated what a quantum computer would need to find the lowest energy of the active centre of cytochrome P450, an enzyme involved in how the body processes drugs. Think of an architect's materials list for a skyscraper: it says what the building would take, not that it stands. Their list names about 4.6 million physical qubits, the machine's raw and error-prone building blocks, running for 73 hours, assuming each one errs at a rate of 0.1%. Many raw qubits are bundled so that they act as fewer, reliable ones. The paper also sets this against a classical method.
“We demonstrate assuming physical qubit error rates of 0.1% that the ground state energy of the largest model Hamiltonian for the active space ‘X’ of Cpd I can be assessed with approximately 4.6 million physical qubits in 73 hours of run time.”
— Section III, Quantum Computing Resource Estimates, in: Reliably assessing the electronic structure of cytochrome P450 on today's classical computers and tomorrow's quantum computers1
- How it was done: a calculation on paper, not run on any machine
- Qubits: About 4.6 million physical qubits, the machine's raw error-prone units; this is a paper estimate, not a device.
- Error correction: A fault-tolerant design, which bundles raw qubits so that errors are caught and corrected, was assumed, not built.
- Compared against classical computing: Run times of a classical method called DMRG were set against the estimated quantum run times.
Who it affects
This concerns drug discovery and materials chemistry, where it matters how an enzyme such as cytochrome P450 handles a molecule. For someone in pharma today, nothing changes in the lab: no machine exists that could do this work. The result serves as a planning yardstick. It tells a strategy team how large a quantum machine must be before this task becomes a practical option.
Pharmaceuticals: McKinsey & Company estimates up to about $400B by 2035 — value at stake for the industry, not vendor revenue (Quantum Technology Monitor, April 2026, p. 20).
This figure is an estimate, not a measurement: McKinsey & Company is a consultancy with its own quantum advisory business. It covers a whole use case, not the result of this paper.
What is still missing
The task needs a fault-tolerant machine with about 4.6 million physical qubits and 73 hours of run time, assuming an error rate of 0.1%. That covers only the lowest energy of a model of the active region. Effects outside that region need new quantum methods, because simulating the whole system directly is judged infeasible. Extrapolating to models with 500 orbitals, the paper finds about 9000 reliable qubits and 1.5 trillion basic computing steps.
Against published roadmaps: the task needs 4.6 million physical qubits. No announcement we have on file reaches that many physical qubits. The furthest any of them goes:
An announcement is not a delivered machine, and this one would still fall short of what the task needs.
Who is working on it
The paper (February 2022) lists its authors' affiliations as: Google Quantum AI; Quantum Simulation Technologies, Inc; Department of Chemistry, Columbia University; Boehringer Ingelheim Pharma GmbH & Co KG; and 2 further institutions. That is who signed this work — it is not a ranking of who is ahead.
It strengthens qubitized phase estimation, a way to read off a molecule's energy. The paper reports that one way of splitting up the problem, tensor hypercontraction factorization, consistently beat single and double factorization.
How this brief was made
This brief was written by software, not by a person: a language model read the full text of the paper, and code then checked every quotation, number and name in this brief against that text before it was published. The rules are set out under Method. Responsible for the rules and for this page: Georg Singer. The paper is a preprint that has not been peer-reviewed. Georg Singer holds no shares in any company or institution named in this brief (as of October 2026).