Increasing numbers of New Zealanders are being diagnosed with diabetes due to rising obesity and widespread screening. Diabetes is a major cause of cardiovascular disease (CVD) but not all people with diabetes have similar high risks of suffering CVD. Many expensive new drugs are becoming available for reducing CVD risk in people with diabetes, and their benefits are greatest in those at highest predicted CVD risk. Current CVD risk prediction tools are based on indirect measures of CVD and only have modest accuracy. The retina offers a direct measure of arterial health and disease. We have retinal photographs from over 50,000 people with diabetes and have developed artificial intelligence (AI) methods for detecting eye complications. This study will determine whether novel retinal AI-based methods for CVD risk prediction can significantly improve current CVD risk prediction tools. This tool could revolutionise CVD risk assessment, facilitating rapid and more cost-effective treatment.