The Challenge
As the world’s largest chemical producer, BASF continuously develops new materials for consumer goods, transportation, healthcare, agriculture, energy, and beyond. To accelerate innovation, BASF sought to understand how hybrid quantum-classical approaches could enhance classical machine-learning pipelines used to predict molecular properties.
The central question: Can quantum methods provide measurable value in materials discovery workflows today?
Our Approach
BASF partnered with Zapata to benchmark quantum and quantum-inspired methods against state-of-the-art classical techniques on real-world chemoinformatics tasks.
Together, we:
- Evaluated hybrid quantum-enhanced ML approaches for molecular-property prediction
- Explored feature-selection and classification routines where quantum subroutines may offer benefits
- Benchmarked performance, scalability, and practicality relative to existing classical pipelines
- Investigated how quantum-inspired algorithms could later be deployed across BASF’s global R&D and manufacturing operations
This structured evaluation leveraged Zapata's algorithmic breadth and deep benchmarking expertise to focus on practical utility, not academic novelty.
Results & Impact
The collaboration generated clear insights and timelines into where and when quantum approaches can meaningfully contribute to BASF’s modeling workflows. Key takeaways included:
- Identification of molecular-modeling subroutines most likely to benefit from quantum acceleration
- Establishment of baseline performance comparisons to guide future adoption decisions
- A validated framework for evaluating quantum utility across BASF’s broader materials pipeline
This work helped BASF develop a forward-looking strategy for integrating quantum and quantum-inspired methods into its R&D operations.
