Google Launches $10 Million 'REPLIQA' Quantum Biology Research Initiative

Level 3 — Intermediate

Google announced on Thursday a five-year, $10 million research initiative called REPLIQA — short for Research Programme for Quantum Biology and Adjacent Sciences. The company will fund five US universities, lend them time on Willow-class superconducting quantum processors, and provide bespoke AI tools that combine large language models with quantum-circuit simulators.

The programme zeroes in on a long-running scientific puzzle: how, and to what extent, do uniquely quantum phenomena — superposition, entanglement, coherent tunnelling — show up inside warm, wet biological systems? Photosynthesis, enzyme catalysis, magnetoreception in migratory birds, and the radical-pair model of olfaction are all suspected to lean on quantum effects, but firm experimental evidence has been hard to assemble.

REPLIQA's first awards will fund projects on protein folding using variational quantum eigensolvers, on photosynthetic light-harvesting complexes, and on cryptochrome-based magnetoreception. Google will also sponsor open-source releases of the underlying simulation libraries — a deliberate contrast with the more proprietary stance taken by IBM and Quantinuum in their own biology efforts.

Catherine Vollgraff Heidweiller, Google Research's head of basic science partnerships, said the company expects the programme to be 'a long-term scientific effort, not a near-term commercialisation play.' That framing matters: Wall Street has tried for two years to push every Big Tech firm to monetise its quantum hardware, and Google's deliberate refusal to attach a product roadmap to REPLIQA signals that, at least for now, the company is willing to spend on curiosity.

initiative: a new plan or programme started for a specific purpose

bespoke: specially made for a particular customer or use

superposition: a quantum state in which a particle is in multiple positions or states at once

entanglement: a quantum link in which two particles behave as if connected over distance

tunnelling: a quantum effect where particles pass through barriers they classically couldn't

catalysis: speeding up a chemical reaction with the help of a catalyst

proprietary: owned and kept secret by a company

monetise: to turn something into a source of money or revenue

Level 4 — Advanced

Alphabet's Google Research division unveiled REPLIQA on Thursday — a five-year, $10 million initiative the company is positioning as the most concrete bet yet on the still-contested field of quantum biology. Funding will flow to five US universities, each of which will receive on-prem cloud access to Willow-class superconducting processors, a quota of runs on Google's classical-hybrid simulator stack, and a tailored fine-tune of Gemini-Bio, a derivative large language model trained on protein, biochemistry and quantum-circuit corpora.

The scientific premise is older than the funding cycle. Since the 2007 Fleming-Engel result showing apparently coherent excitation transfer in the FMO complex of green sulphur bacteria, a small but stubborn community has argued that warm, wet biology exploits quantum effects — superposition, vibration-assisted tunnelling, radical-pair spin chemistry — far more than the textbook decoherence story would predict. The evidence has matured unevenly: claims for cryptochrome-based avian magnetoreception have hardened, while early excitement around purely coherent photosynthetic transfer has been partly revised by the recognition of vibronic mechanisms.

REPLIQA's first round of awards is reported to target three classical bottlenecks. At Berkeley, a team will use variational quantum eigensolvers to compute electronic structures of folding intermediates that are too entangled for density functional theory. At MIT, a group will simulate light-harvesting complexes with sufficient bath fidelity to discriminate between Förster, Redfield, and exact open-quantum-system kinetics. And at Caltech, a third effort will model the avian cryptochrome radical pair under realistic geomagnetic field strengths, an experiment that has so far refused to converge on conventional hardware. Google will mirror all derived libraries under Apache 2.0 — a calculated contrast with the more proprietary platforms of IBM Quantum and Quantinuum.

Catherine Vollgraff Heidweiller of Google Research said REPLIQA is 'a long-horizon scientific effort, not a near-term commercialisation project,' echoing language Google has used since at least 2019 to insulate its quantum work from quarterly product reviews. The market is unlikely to disagree quickly: there is no obvious revenue line in funding quantum-chemical models of how a robin senses magnetic north. But Google's bet is structural rather than tactical — the same wager Bell Labs and PARC made in earlier decades, that the value of foundational science compounds, even when it cannot be priced.

on-prem: delivered or hosted on the customer's own premises rather than the public cloud

fine-tune: to further train a large model on a narrow domain for better performance

decoherence: the loss of quantum coherence due to interaction with the environment

vibronic: involving coupled vibrational and electronic energy levels

eigensolver: an algorithm that finds the eigenvalues — the allowed energies — of a system

bath fidelity: how accurately a simulation models the surrounding environment of a quantum system

geomagnetic: relating to the Earth's magnetic field

foundational: providing the base for later development, even without immediate use

Level 1 — Absolute Beginner

Level 2 — Elementary

Level 3 — Intermediate

Level 4 — Advanced

Google Launches $10 Million 'REPLIQA' Quantum Biology Research Initiative

Multiple Choice

True or False

Fill in the Blank