In October 2024 the Nobel Prize in Chemistry went to David Baker for designing new proteins, and to Demis Hassabis and John Jumper of Google DeepMind for AlphaFold — an AI model that predicts a protein's 3D shape from its sequence, a problem scientists had chased for 50 years.
Working out one protein shape used to take a year or more in the lab. AlphaFold does it in minutes. By 2025 researchers in more than 190 countries were using it. AlphaFold 3 went further, predicting how proteins interact with drug molecules, and early-stage drug discovery that took five to seven years is now being done in roughly two.
The bigger lesson is the pattern: a clear problem, lots of good data, a model that learns from it, and experts who check the results. That pattern works in manufacturing, agriculture and energy too — and the companies that organised their data early are moving fastest.
What it means for your business
- Good, well-organised data is the foundation of every AI breakthrough — start building it now.
- Define the problem precisely; vague goals do not produce results.
- Always have experts check what the AI produces.
How MTH can help: Our free discovery session finds where your data could give you an AI advantage, and our Data Services make that data ready to use.
This briefing is original commentary by MTH based on public reporting; figures are as reported by the sources named above and may be revised over time.
