Why Biological Data Centers Could Revolutionize AI-Driven Drug Discovery
The field of artificial intelligence (AI) in science is rapidly advancing, with systems now generating hypotheses and running experiments in automated labs. However, most AI systems still lack real-world adaptability, as they primarily learn from papers and proxies. To overcome this, biology needs a substrate or infrastructure to enable AI to discover and adapt, much like how data centers advanced computation.
Infrastructure plays a crucial role in shaping possibilities, making intelligence inevitable by providing a substrate to run, adapt, and scale. In the realm of AI for science, biological data centers could serve as this infrastructure. These centers would offer an environment for AI to test molecules in human systems, thereby improving drug discovery and reducing late-stage failures.
Currently, drug development heavily relies on animal models and small patient cohorts, leading to a high failure rate in human clinical trials. Biological data centers could help overcome this challenge. However, the United States is not at the forefront of this development. China is building biological infrastructure at a scale and speed that surpasses U.S. efforts, potentially putting the U.S. at a disadvantage in leading drug discovery and defining medicine for the next century.
Policy and industry are increasingly favoring human-relevant evidence in drug discovery, driven by ethical and economic pressures. To stay competitive, the U.S. must act now to build its own biological data centers. By doing so, it can lead in drug discovery and ensure that its discoveries bear an American signature. The time to invest in this infrastructure is now, as the country that builds it first will likely lead in drug discovery and shape the future of medicine.
