DY0-001 launched in July 2024 as CompTIA DataX, the data science entry in CompTIA's expert-level series. In early 2026 CompTIA renamed it DataAI. The exam code and content stayed the same. It is a 165 minute exam with up to 90 questions mixing multiple choice and performance-based items, scored pass/fail with no published scaled score.
CompTIA recommends five or more years in data science or a similar role, and its own page estimates the exam will retire around three years after launch. There are five domains: mathematics and statistics (17%), modelling, analysis and outcomes (24%), machine learning (24%), operations and processes (22%) and specialised applications of data science (13%). The machine learning domain covers supervised, tree-based and unsupervised methods plus deep learning, including neural networks, backpropagation, dropout and batch normalisation. Specialised applications cover NLP, computer vision and reinforcement learning. Credit where it is due. This is a hard, broad exam, and it does not shy away from the maths.
A lot of AI certifications in 2026 are really product literacy tests, and DataAI is the opposite: it tests whether you understand what the models are doing. The operations domain is also a sensible inclusion, because most data science work that fails, fails in deployment and monitoring rather than in the model. My problems are commercial. The rename to DataAI with no change to the exam feels like keyword chasing, and it splits search results and job ads between two names for one credential. CompTIA does not show the price on the certification page, and the figure I found from third parties, around $529 for a single attempt, would make it the most expensive exam CompTIA sells.
That is a lot to pay for a credential in a field where hiring managers mostly look at GitHub, Kaggle and past work. Search interest for the DataX name has also fallen by more than half over the past year. Who should take it? People in public sector or defence-adjacent roles, where CompTIA certifications slot into formal workforce frameworks and having one can genuinely matter. For a data scientist in a commercial company, I would spend the money on compute and build something.