【AI前沿】This young startup is taking on a fragrance industry that hasn’t changed in almost half a century

Fragrance tech companyPatinasays it has raised $2 million in funding investors, including Betaworks and True Ventures.The company focuses on creating new scent molecules using advanced molecular design, machine learning, and scent research. Today, most of the scent molecules used in consumer products are created by a small number of specialized labs, which then sell those molecules to fragrance houses or cosmetics companies — the brands that ultimately turn them into perfumes, candles, or flavored products. Patina is trying to shake that up, entering an area that has seen little innovation in the past half century.The company was founded by Sean Raspet and Laura Sisson. Raspet is an artist and perfumer who, over time, developed an obsession with human senses and began creating new scent and flavor molecules as a creative pursuit. Sisson, meanwhile, came from a background in food and software engineering and became obsessed with human senses after discovering an entire scientific field dedicated to modeling them. The two met, naturally, at a scent art gallery in New York in 2024, where Raspet was exhibiting new molecules and Sisson was an engineer building olfactory learning models.“We started collaborating on research, and it became clear that the timing was right to finally build the tools to understand scent at the biological level,” Raspet told TechCrunch. “That felt like a company.”They launched Patina last year and began working on a foundation model called Sense1, designed to replicate the scent receptors in the nose and create what they describe as “the first universal code of smell and taste.” Currently, researchers largely use words like “floral” or “woody” to describe smells, an imprecise system that leads to inconsistencies across regions and languages. Working on the receptor level, he said, allows them to create “never-before-smelled molecules and reconstruct the world’s rarest natural ingredients.”Patina said it is already in talks to work with top fragrance houses and with fashion brands about creating custom scents. The timing feels right. Customers increasingly want “newer, safer and more expressive perfumes,” Sisson said. There’s also supply-chain pressure. Many natural ingredients like rose oil are becoming harder to produce and more expensive — a problem that synthetic alternatives could help solve. Patina’s molecules can simulate the smell of rose oil at the biological level, mimicking the natural material without the need for plant extraction.“These replications are less carbon-intensive than the original plant extract, consuming significantly less water and petrochemicals,” Raspet said.Others in this space include startups like Osmo and legacy incumbents like Givaudan and Symrise, two of the largest flavor and fragrance giants in the world.For Patina, there is also an intellectual property angle worth noting. Right now, only fragrance molecules can be patented, not the formulas themselves, meaning that scents can easily be replicated. This benefits the large fragrance houses, the only players that could really afford to develop enough scent variations in a lab. AI has made this process cheaper and faster, letting smaller companies like Patina create custom scent ingredients in weeks, not years.“We think by expanding the palette, perfumers and flavorists at all scales will be able to develop and protect their signature style,” Raspet said.AI is also transforming other parts of the scent industry. It’s helping to phase out animal testing, since new models can predict human-skin reactions nearly as accurately, Raspet said. And while understanding how primary scents work at a molecular level seemed far-fetched to researchers even five years ago, the Patina team said AI is helping unlock breakthroughs in how the senses function at a molecular level.Raspet said the new funding has already allowed the team to move from his backyard into a proper office in Bushwick, Brooklyn, with a small group of chemists and will go toward launching new molecules and funding new partnerships.“All models need data to learn from, and we’ve been able to fund collaborations with startups and academic labs to gather this receptor activation data. At the same time, we believe more computationally detailed simulation of interactions of molecules with odor receptors will be a huge unlock for scaling,” he added.The long-term ambition is to create what Raspet calls a “Pantone for scent” — a reference to the universal color-matching system used across design and manufacturing industries — establishing the primary scent molecules from which any smell or flavor can be built. “The information has been there the whole time, waiting for the technology to catch up and a team with the right combination of expertise and obsession to unlock it,” Raspet said. “These ideas can now be made real, with Patina as the underlying intelligence layer.”Topicsbiotech,fragrance,Startups,VentureWhen you purchase through links in our article