The development of tabular foundation models (TFMs) has accelerated in recent years, showing strong potential to outperform traditional ML methods for structured data. A key finding is that TFMs can be pretrained entirely on synthetic datasets, opening opportunities to design data generators that encourage desirable model properties. Prior work has mainly focused on crafting high-quality priors over generators to improve overall pretraining performance. Our insight is that parameterizing the generator distribution enables an adversarial robustness perspective: during training, we can adapt theResearch goal: How does adversarial generator pretraining affect the out-of-distribution robustness of tabular foundation models compared to standard pretraining on synthetic data?Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 7.6/10.