Chemistry-Native AI Startup Albert Raises $55 Million Series B Led by TCV

Albert, a startup building what it calls the first chemistry-native AI operating system for enterprise R&D, announced a $55 million Series B on Tuesday led by TCV, with new investors Delta-v Capital, Socium Ventures, HighSage Ventures, Coral Capital and Legion VC joining existing backers Coatue, J.P. Morgan Growth Equity Partners, Index Ventures and Applied Ventures.

The thesis is that general-purpose frontier models consistently fail at specialized scientific work. Albert's pitch rests on a companion research release: a benchmark for formulation chemistry called FormBench, on which the company says its specialized model Valence is 45 times more accurate than general-purpose AI at retrieving correct answers to scientific queries. Valence was trained to discriminate chemical composition and function while reading text — the way an experienced chemist would, per the company. Those figures come from Albert's own research and have not been independently verified.

The product surface is Ask Albert, an orchestrated harness of tools, agents and models that sits on top of Albert OS, the company's scientific ontology and data platform. Chemists query it in natural language; it surfaces institutional knowledge, suggests formulation candidates, designs experiments and runs analysis. Customers named in the announcement include AkzoNobel, Avient, BASF, Chemours, Clorox, Henkel, Kenvue and Nouryon — an unusually deep roster of chemical and consumer-goods majors for a company at Series B.

Growth metrics, again company-reported: project creation on Albert OS grew 26 times year over year, scientist users grew six times across 44 countries, and workflow creation quadrupled. In six months the company says Ask Albert evolved from an AI search assistant into a full R&D partner that reasons across all of Albert OS. Kenvue's head of R&D digital capabilities credited the platform with connecting data from raw materials to notebook entries in a global organization.

Albert is also selling governance alongside intelligence — zero data retention, fine-grained access controls and reasoning audit trails — aimed at regulated industries where a hallucinated formulation is not an embarrassing chatbot answer but a compliance event. CEO and co-founder Nick Talken, who comes from the chemical industry, argues enterprise AI for chemistry has to be built "from the ontology up" rather than wrapped around a general model.

The round fits a widening pattern: vertical AI companies raising at healthy multiples while horizontal model makers absorb scrutiny over ROI. Chemistry and materials R&D remains one of the last big enterprise functions run on spreadsheets and institutional memory, and the failure mode of general LLMs there is well documented. Whether Albert's 45x retrieval claim survives independent scrutiny matters less commercially than whether its customers keep expanding deployments — on that front, the named roster does most of the talking.