The non-profit research institution Biohub, supported and founded by Mark Zuckerberg and Priscilla Chan, announced the expansion of its 'Virtual Biology Initiative.' This initiative, in collaboration with multiple institutions including the U.S. Department of Energy, NIH, Google DeepMind, Isomorphic Labs, and Meta, has collectively raised approximately $1.8 billion in funding, data, computing power, and experimental measurement resources. The goal is to build an open data infrastructure for next-generation biology AI models. The project aims to generate multimodal standardized data that reflects cellular changes before and after interventions, supporting the training of AI virtual cell models to accelerate the exploration of disease mechanisms and the discovery of potential treatments. The project will focus on investing in cryo-electron tomography, large-scale microscopic imaging, and high-throughput biological intervention and measurement technologies. It is expected to produce the first large-scale new datasets within a year and achieve substantial progress in cell behavior prediction models within five years. Data will be subject to a differentiated open access strategy, with commercial sponsors enjoy (enjoying) priority time-limited usage rights for the data they fund, while government-funded data will be directly open. Currently, the field of AI biology is undergoing a transition from utilizing existing scientific research data to train models to directly participating in biological experiments and data production. However, key questions remain unresolved, such as whether large-scale models can be effectively applied to cell biology and how to design experiments to generate high-quality data that supports models in understanding the causal relationships of cellular operations.