UK researchers gain access to Google's AI and quantum tech
The UK government has struck two agreements with Google-run companies to expand scientists' access to quantum hardware and AI for research. Announced on 12 December, the deals open Google Quantum AI's Willow processor to UK researchers and give priority access to Google DeepMind's AI for Science models.
Science minister Patrick Vallance said access will "help keep Britain's quantum innovators at the cutting edge" and support work in new medicines and clean, affordable energy.
Quantum access via the NQCC and Google Quantum AI
The National Quantum Computing Centre (NQCC) will partner with Google Quantum AI to provide UK researchers with access to the Willow quantum processor. The NQCC and Google will jointly invite proposals from UK researchers and consortia.
"UK researchers will be able to submit proposals describing their intended experiments or applications, the scientific rationale behind them, and how access to Willow could enable transformative results," said Michael Cuthbert, director at the NQCC.
- State your experiment or application
- Explain the scientific rationale
- Show how Willow access could enable results that current methods can't
For context on the NQCC's mission and calls, see the official site at nqcc.ac.uk.
Priority access to DeepMind's AI for Science
On 10 December, the government extended its partnership with Google DeepMind, giving UK scientists priority access to four AI for Science models. Science secretary Liz Kendall called DeepMind "a firm with roots on both sides of the Atlantic, backing British innovators to shape the curve of technological progress".
- AlphaEvolve: a coding agent powered by Google's Gemini
- AI Co-Scientist: a "virtual scientific collaborator" for research workflows
- WeatherNext: a family of weather forecasting models
- AlphaGenome: focused on DNA
Background on DeepMind's research programs is available at deepmind.google.
Automated materials lab launching in 2026
Google DeepMind will establish the UK's first "automated laboratory" for materials science in partnership with the government. A multidisciplinary team will run the lab, including work on superconductor materials that carry electricity with zero resistance.
Demis Hassabis, Google DeepMind co-founder and chief executive, said the company was "excited to deepen our collaboration with the UK government and build on the country's rich heritage of innovation to advance science".
What this means for your research
- Start proposal prep now: Draft a concise concept note covering your target problem, algorithmic approach, expected outcomes, and why quantum access matters. Refine it into the formal proposal once the call opens.
- Map use-cases: Link your work to near-term areas mentioned by officials-drug discovery, clean energy, materials, weather, and genomics-and select the matching tool (Willow, AI Co-Scientist, WeatherNext, etc.).
- Plan for data, IP, and compute: Define datasets, security needs, and publication strategy. Check access terms, quotas, and any co-authorship or attribution requirements as they're published.
- Build a validation loop: Set baselines, design small-but-clear benchmarks, and document workflows so results are reproducible across classical and quantum runs.
- Upskill your team: Align skills on generative coding, model tooling, and experimental automation. If you need structured training, see AI courses sorted by leading AI companies.
How to get involved
- Monitor the NQCC and Google Quantum AI channels for the call for proposals.
- Assemble a cross-functional team (domain lead, methods lead, engineering support, and data steward).
- Scope a pilot that can show measurable scientific value within current hardware limits.
- Document risks and ethics considerations early, especially for biomedical and climate applications.
These agreements open practical pathways for UK scientists to test ideas that were out of reach last year. If you're ready with a sharp question, clean data, and a lean plan, this is your window.
Your membership also unlocks:
