This is groundbreaking. Predicting how a complex protein molecule will fold in 3D space can speed up drug discovery for one thing. Using AlphaFold, a database containing 200 million proteins and their folding patterns has been created.
Protein folding prediction is also an important basic tool in understanding molecular biology in general. Its benefits will be felt across the entire field.
DeepMind first made headlines in 2016 when it's AlphaGo AI beat the world's highest ranked professional go player Lee Sedol using its system relying on reinforcement learning and having slightly different versions of it play millions of games against each other.
AlphaFold is open source, which means there will not be any immediate economic effects. Biotech stocks are generally quite speculative. In my limited experience, investing in biotech resembles investing in random shitcoins. That's because few people have the sort of specialized knowledge required to pick the winners who can benefit the most from something like this.
https://www.deepmind.com/research/highlighted-research/alphafold