Google DeepMind has announced AlphaGenome 2, a next-generation DNA sequence model that predicts how genetic variants change gene activity in specific human cell types. The original AlphaGenome predicted broad tissue averages for many experiments at once. The new model is conditional: it takes a target cell type or cell state alongside a one-million-letter DNA sequence and predicts effects for that condition alone. It was trained on more than 100 million individual cell profiles across 68 tissues from human and mouse samples. It can also adapt to new single-cell data without retraining, and it can predict missing measurements, such as estimating chromatin accessibility from gene activity data. DeepMind reports a substantial gain in variant prediction. On the TenK10K dataset, a single-cell map of 28 primary immune cell types, AlphaGenome 2 identified causal variants 30% of the time, compared with 16% for the original AlphaGenome fine-tuned on single-cell data and other specialised tools. It identified directional gene effects 43% of the time, compared with 23%. On regions outside its training data, the model also explains more of how gene expression differs between cell types within a tissue, such as during the maturation of blood stem cells. Its outputs cover gene expression, histone modifications, chromatin accessibility and DNA methylation. The model is available now for non-commercial research through the AlphaGenome API, with commercial research access in private preview. The published text is incomplete, so results beyond those cited should be treated as unverified. For Australian businesses and professionals, the practical impact is limited for now. Commercial use is restricted to a private preview, and most organisations have no direct use for a genomics model. Researchers in genomics may find the non-commercial API worth evaluating. Any project that feeds patient or participant DNA into a model should be treated with care, because genetic information is health information and therefore sensitive information under the Privacy Act 1988 (Cth). That requires clear consent, data governance and security controls before any pilot begins.