
How to read this finding
Predicted structures and database usage do not measure discoveries or clinical benefits. Plasma control in one device does not establish net-energy fusion. Data leakage can invalidate apparently strong evaluations.
They greatly expanded the set of predicted protein structures available to researchers. They were computational predictions, and the report expected downstream scientific benefits to emerge over time.

Predicted structures and database usage do not measure discoveries or clinical benefits. Plasma control in one device does not establish net-energy fusion. Data leakage can invalidate apparently strong evaluations.
Evidence you can use
Historical snapshot: October 2022. Dates and populations are specified per row.
| Measure | Reported value | Definition and source |
|---|---|---|
| AlphaFold predicted structures | Approximately 200M | Predicted structures in the expanded database, not experimental determinations.2022 report, slide 16 (PDF page 16) |
| Reported database users | More than 500,000 | Researchers who had used the database in the report’s account.2022 report, slide 16 (PDF page 16) |
| Plasma-control deployment | TCV tokamak, Lausanne | Specific experimental device used for the reinforcement-learning demonstration.2022 report, slide 15 (PDF page 15) |
Predicted structures and database usage do not measure discoveries or clinical benefits. Plasma control in one device does not establish net-energy fusion. Data leakage can invalidate apparently strong evaluations.
Historical snapshot published October 11, 2022. This web edition was prepared on 2026-10-11 from the online deck and original launch posts. Findings and forecasts retain their original time frame.
Benaich, Nathan, and Ian Hogarth. “What did the 200 million AlphaFold structures mean?” State of AI Report 2022. Historical report snapshot; web edition prepared 2026-10-11.