There is a quiet, expensive war being fought in thousands of industrial labs. On one side is the old guard: materials scientists and chemical engineers who have spent careers developing new polymers, battery cathodes, or industrial coatings through intuition, trial, and error. On the other side is a stack of disconnected spreadsheets, proprietary databases, and decades of published but unsearchable papers. The cost of winning, or losing, is measured in years of development time and billions in potential revenue.
Citrine Informatics has spent the last eleven years building a platform to end that war by turning materials R&D into a data science. The Redwood City company sells an enterprise AI suite that ingests a lab's internal test data, patents, research papers, and supplier specs, then uses machine learning to predict how new material formulations will behave. It is a bet that the slow, artisanal process of inventing new physical stuff can be systematized.
The data refinery for physical things
At its core, Citrine is selling a data refinery. Its platform, built around products called DataManager, VirtualLab, and Catalyst, acts as a unified repository for the chaotic information streams of industrial R&D [Citrine Informatics]. DataManager records and displays processing parameters and measured outcomes. VirtualLab applies machine learning to that data to propose new material candidates or optimize existing ones. Catalyst, a newer addition, is a private digital assistant that uses large language models to search and synthesize knowledge from journal papers and internal documents [Citrine Informatics].
A bet on the industrial slow lane
Citrine's bet is fundamentally a timing one. Founded in 2013, the company predates the current generative AI boom by nearly a decade. It has raised an estimated $78.7 million over nine rounds from investors including Innovation Endeavors, Prelude Ventures, and Next47, with the most recent a $2.6 million Series C3 in February 2025 [Tracxn, PitchBook].
Citrine is not chasing tech startups. Its wedge is the massive, established chemical, coatings, battery, and consumer products companies where R&D budgets are large but processes are often sclerotic. Panasonic has used the platform to save time and grow intellectual property in materials development [Citrine Informatics]. German chemical giant Lanxess is launching a pilot project with Citrine to develop new materials more efficiently [Citrine Informatics]. Another partnership with Econic Technologies aims to accelerate development of CO₂-based materials [Citrine Informatics].
The team built to speak two languages
Building for this market requires fluency in both machine learning and materials science. The founding team appears structured to bridge that gap. CEO Gregory Mulholland and Chief Strategy Officer Bryce Meredig are both graduates of the Stanford Graduate School of Business and founded the company in 2013 [Crunchbase]. Co-founder Kyle Michel served as CTO, bringing the technical depth [Tracxn]. The company now lists 68 employees.
Where the model meets the molecule
The most credible risk for Citrine isn't technological overreach, but commercial friction. The platform's value is inextricably linked to the quality and quantity of data a customer feeds it. Citrine's answer seems to be a focus on specific, high-value use cases where the ROI is blindingly obvious. The recent $3.1 million award from the U.S. Department of Energy to develop data-driven methods for nuclear waste storage materials is a case in point.
The next twelve months of proof
For a company founded in 2013, the next phase is about transitioning from promising pioneer to scaled enterprise. The key milestones to watch are not technological, but commercial. The conversion of pilots like Lanxess into enterprise-wide deployments would be a powerful signal. Landing a first major contract with an automotive OEM or a top-tier battery manufacturer would validate the platform in the most demanding supply chains.