The numbers behind the AI-for-science movement.
Landscape2026-02-10AI x Quantum Research Team
AI x Quantum: The Data Behind the Hype
Funding tables, government programs, and a curated reading list for researchers
referencefundingFROsgovernmentreading listdata
This post is a living reference — the data tables and reading lists behind our analysis of AI-accelerated science and the AI x quantum intersection. Bookmark it.
AI-for-Science Startup Funding
The scale of investment tells the story of a field that went from speculative to operational in 2025.
| Company | Funding | Focus | Key Detail |
|---|---|---|---|
| Xaira Therapeutics | $1.3B | Drug design | Co-founded by David Baker (2024 Nobel). First drug entering human testing 2026. |
| PsiQuantum | $1B+ | Photonic quantum | Largest quantum hardware raise. Partnership with GlobalFoundries. |
| Isomorphic Labs | $600M | Drug discovery | DeepMind spinoff. Novartis + Eli Lilly partnerships (~$3B). Clinical trials late 2026. |
| Lila Sciences | $550M | Drug discovery | "AI Science Factories" — 235,500 sq ft Cambridge lab. George Church as Chief Scientist. |
| Arcadia Science | $500M | Open biology | Funded by Jed McCaleb + Sam Altman. Studies understudied organisms. |
| Insilico Medicine | $400M+, IPO'd | Drug discovery | First AI-discovered drug (rentosertib) in Phase II. Hong Kong IPO ($293M). |
| Periodic Labs | $300M seed | Materials & chemistry | Largest AI-for-science seed. Founded by Liam Fedus (ex-OpenAI VP) + Ekin Cubuk (ex-DeepMind). |
| Sakana AI | $2.65B valuation | AI research agents | Founded by Llion Jones (Transformer co-author). "AI Scientist v2" accepted at ICLR workshop. |
| CuspAI | $100M | Materials design | AI-guided materials discovery. |
| Edison Scientific | $70M seed | Drug repurposing | Spun out of FutureHouse. Robin agent identified ripasudil for macular degeneration. |
CZI (Chan Zuckerberg Initiative) committed $10B over the next decade to AI-for-biology, absorbing the EvolutionaryScale team (ESM protein language models). This is the single largest commitment in the field.
Focused Research Organizations (FROs)
The FRO model — championed by Sam Rodriques, Adam Marblestone, and Eric Schmidt through Schmidt Sciences — addresses a gap: scientific problems that need $20-50M over 5-7 years and produce public goods rather than products. Convergent Research has incubated 10 FROs:
| FRO | Focus | Notable Result |
|---|---|---|
| E11 Bio | Large-scale connectomics | Brain wiring maps at industrial scale |
| Lean FRO | Math theorem formalization | 210,000+ theorems in Mathlib; 2025 ACM SIGPLAN award |
| Cultivarium | Photosynthetic organisms | Engineering novel photosynthetic pathways |
| [C]Worthy | Ocean carbon | Carbon removal verification systems |
| Bind, EvE Bio, Forest Neurotech, Imprint | Various | Neurotechnology, bioengineering |
No FRO currently targets quantum computing, but the model fits: AI-accelerated QEC or autonomous quantum experiment design needs multi-year, multi-million-dollar focused effort that's too big for a single lab and too "public good" for a startup.
Government Initiatives
| Program | Country | Investment | Key Detail |
|---|---|---|---|
| Genesis Mission | US | $320M+ | 17 DOE labs + 24 industry partners. Google AI co-scientist across all labs. |
| AI for Science | UK | £137M | Multiple research councils funding AI integration. |
| ARIA "AI Scientist" | UK | £6M total | 12 projects at ~£500K each, selected from 245 proposals. Partners: Lila Sciences/MIT, UCL, Liverpool. |
| FROST UK | UK | TBD | ARIA + Convergent Research bringing FRO model to UK. |
| Quantum + AI Strategy | Japan | $135B | Combined quantum + AI national strategy, one of the largest globally. |
| EU Quantum Act | EU | Multi-billion | Major funding framework. TU Delft / QuTech centrally positioned. |
| Schmidt Sciences AI2050 | Global | $125M | 99 fellows at 42 institutions working on hard AI problems. |
Five Methods of AI-Accelerated Science
Across the landscape, five distinct approaches have emerged:
| Method | How It Works | Best Example | Quantum Application |
|---|---|---|---|
| Foundation Models | Train large models on domain data to learn a field's "language" | AlphaFold (protein), GNoME (materials, 2.2M crystals) | QUASAR (quantum circuits), domain-specific quantum models |
| Autonomous Agents | AI plans, executes, and analyzes experiments | Sakana AI Scientist, Google AI co-scientist, FutureHouse Kosmos | k-agents (superconducting calibration), our MCP pipeline |
| Self-Driving Labs | Closed-loop: AI + robotic execution + feedback | Ginkgo/GPT-5 (36,000 experiments), DeepMind UK materials lab | Quantum processors are already digital — no robot needed |
| LLM Code Generation | Generate working scientific code from descriptions | QCoder (78%), our benchmark (68–71% with RAG) | Circuit synthesis, SDK translation, error analysis |
| AI-Guided Search | RL/evolutionary methods for combinatorial optimization | AlphaTensor-Quantum (halved T-gates), AlphaEvolve | Circuit optimization, error code discovery |
Reading List: AI x Quantum Papers
Neural Error Correction
- AlphaQubit — Transformer QEC decoder, Nature 2024
- Mamba-based QEC decoder — O(d²) scaling follow-up
- NVIDIA CUDA-Q QEC — <4μs GPU decoding
AI Code Generation for Quantum
- QUASAR — 99.31% circuit validity via agentic RL
- QCoder — 78% functional accuracy with o3
- Qiskit HumanEval — The benchmark we used
Autonomous Quantum Agents
- k-agents — Self-driving quantum lab, Patterns 2025
- QCopilot — Autonomous atom cooling, 100x speedup
- AlphaTensor-Quantum — RL circuit optimization, Nature Machine Intelligence 2025
Circuit Optimization
- IBM AI transpiler — 42% two-qubit gate reduction
- AlphaEvolve — Evolutionary algorithm discovery
AI + Science Landscape
- AI for quantum computing — Nature Communications survey, 3 tiers
- AI expands impact, contracts focus — Nature 2026, 41.3M papers analyzed
- Andrew White on Latent Space — Scientific taste, Kosmos system
Hardware Milestones
- Google Willow — 105q, exponential error suppression, Nature 2024
- QuTech 10-spin germanium array — >99% fidelity
- 300mm silicon spin qubits — Industrial wafers, >99% fidelity (Diraq/imec, Nature 2025)
Sources & References
- Convergent Research (FRO incubator)https://www.convergentresearch.org/
- Schmidt Scienceshttps://www.schmidtsciences.org/
- DOE Genesis Missionhttps://www.energy.gov/articles/energy-department-announces-collaboration-agreements-24-organizations-advance-genesis
- ARIA AI Scientist programmehttps://www.aria.org.uk/
- AlphaQubit (Nature)https://www.nature.com/articles/s41586-024-08148-8
- AI for quantum computing (Nature Communications)https://www.nature.com/articles/s41467-025-65836-3
- AI contracts science focus (Nature)https://www.nature.com/articles/s41586-025-09922-y