dsm-firmenich and Boltz partner on AI for ingredient discovery

Key highlights
  • Partnership announced September 14, 2026.
  • Boltz’s proprietary foundation models will be integrated into dsm-firmenich’s ingredient-discovery for scent and flavor molecules.
  • dsm-firmenich gains full access to Boltz Lab, the Boltz API, and an LLM-agent integration for natural-language orchestration of workflows.
  • Embedded Boltz researchers will work with dsm-firmenich teams to build exclusive models using the company’s taste and odor receptor data.

Overview

dsm-firmenich has partnered with Boltz to apply Boltz’s proprietary biomolecular foundation models to the company’s ingredient-discovery processes for scent and flavor molecules. The collaboration aims to produce exclusive models that encode decades of sensory-science know‑how and to accelerate the discovery and design of new molecules.

Technical integration

The engagement gives dsm-firmenich full access to the Boltz platform, including the Boltz Lab scientist interface, the Boltz API, and a new integration with LLM agents that enables natural-language orchestration of complex design and prediction workflows. Models will be used to predict and design molecular structures and to model how molecules interact with taste and odor receptors.

Operational model

Boltz researchers will be embedded with dsm-firmenich teams to pair frontier AI models with hands-on expertise, allowing perfumers, flavorists, and scientists to explore molecular space faster and with greater confidence. dsm-firmenich notes this continues its prior use of AI, including launching the world’s first AI-created flavor in 2020.

Leadership comments

Sarah Reisinger, Chief Science & Research Officer at dsm-firmenich, cited Boltz’s scientific pedigree and model capability as reasons for the partnership. Boltz CEO Gabriele Corso highlighted the broader potential of biomolecular interaction and design models beyond pharmaceuticals and called taste and odor receptors “some of the most interesting and understudied proteins” to tackle with their models.

Source: dsm-firmenich

chemXplore Weekly

The week’s most important chemical-industry moves — analysed and summarised — delivered free every Wednesday.

Free. One email a week. Unsubscribe any time.