Machine learning analysis of resting electrocardiograms can predict submaximal VO2max with 0.61 correlation and body fat percentage with 0.65 correlation, using ECG data from 8,540 to 78,265 UK Biobank participants. The AI model also estimated grip strength, lung function, and blood pressure with moderate accuracy (correlations 0.31-0.55). Adding ECG features to basic demographics improved VO2max prediction by 8%. This represents a significant advance in accessible fitness assessment. Currently, measuring VO2max requires expensive equipment and maximal exertion testing, limiting its clinical utility despite being the gold standard for cardiorespiratory fitness. If validated, this approach could democratize fitness monitoring through routine medical visits, potentially enabling early detection of cardiovascular decline and personalized exercise prescriptions at population scale. The technology could prove especially valuable for elderly or mobility-limited patients who cannot perform traditional fitness tests. However, the moderate correlation values suggest the predictions remain estimates rather than precise measurements. As this is a preprint awaiting peer review, the methodology and results require independent validation before clinical implementation. The cross-cohort transferability challenge also needs addressing for broader applicability.