Tech Startup Uses AI to Create Scents in 48 Hours — But Critics Say It Misses the Point

Creating a new fragrance has always been a slow and subtle art. Traditional perfumers might spend six to 18 months carefully blending ingredients, testing formulations, and revising scent profiles — often after years of harvesting and aging the raw materials. But one U.S. tech startup says it can skip all that.

Meet Osmo, a company that’s turning heads and wrinkling noses with its claim to develop AI-generated fragrances in just 48 hours. Their secret? Something they call “artificial olfactory intelligence.”

According to The Verge, Osmo’s system works by digitizing scent molecules into code, then using algorithms to predict what various combinations will smell like. In theory, that means a client can prompt Osmo with a scent idea — say, “rain on cedar” or “espresso in Rome” — and receive a full fragrance compound just two days later.

Their first demo fragrance? A digitized plum. According to The Glossy, testers said it smelled almost exactly like the real thing — though perhaps a bit too perfect, lacking the subtle variability of natural scent oils.

Osmo says that’s exactly the point. The company’s vision is to “democratize scent” by making fragrance creation cheaper, faster, and more accessible, especially to small brands or independent creators. Their technology also promises to cut through red tape by producing shelf-stable and legally compliant formulas without the long delays typical in the perfume industry.

That could be a game-changer in a market dominated by just four major conglomerates, which manufacture the majority of global fragrance ingredients for everything from perfumes to cleaning sprays. Those companies already use AI in their workflow, but the actual creative process is still human-led — and fiercely guarded.

That’s where Osmo hits a wall.

Traditional perfumers and luxury retailers have reacted with skepticism. While some acknowledge that AI could improve manufacturing efficiency, they argue that fragrance is still an art, not just chemistry. High-end perfume, they say, is about time, storytelling, and sensory surprise — things that algorithms can’t replicate.

“AI-created fragrance doesn’t make the cut,” said Matt Belanger, co-owner of the luxury fragrance retailer Stele, in an interview with The Verge. Belanger’s team manually audits every brand they carry, and he says Osmo’s digitally generated scents don’t meet the bar for craftsmanship.

Many perfumers agree that Osmo’s approach could lower the barrier to entry — especially for businesses that want a branded scent without a full creative team. But they argue that making perfume isn’t just about matching molecules. It’s about evoking memories, telling stories, and sometimes even capturing emotions that are hard to explain, let alone code.

And there’s another concern: If AI can replicate natural scents in a lab within 48 hours, what happens to the traditional growers and harvesters — the farmers, distillers, and artisans who have supplied raw materials for centuries?

For now, Osmo’s ambitions are still in the early stages. The company plans to sell both custom scents and off-the-shelf fragrance formulations. Whether that model will disrupt an industry built on exclusivity and expertise — or simply create a new category of mass-market “smart scents” — remains to be seen.

But one thing’s clear: For an industry that prides itself on intuition, tradition, and artistry, the arrival of AI is forcing the perfume world to ask a tough question — can a machine smell the difference?