AI scent sensor takes on RFID and barcodes

US startup Osmo has launched an AI-powered scent sensor to authenticate products in different ways to barcodes and RFID (radio frequency identification) tags.

Counterfeiting costs retailers and consumers billions of dollars each year. Osmo’s scent sensor works differently than existing authentication technology by reading the unique scent signatures of authentic products. Chemical sensors are combined with AI trained on massive datasets to recognise subtle scent patterns, ignore background scents, and deliver clear yes-or-no answers in the field. Embedding these sensors in a supply chains makes authentication faster, simpler, and more reliable.

Osmo uses a variety of models in its work to digitise scent, including Graph Neural Networks (GNNs) to understand molecules, their properties and their smell. This technology allowed for the creation of a Principal Odour Map, which is the basis of the research the company is undertaking to be able to read, write and map scent across various fields.

• Indoor air quality platform uses AI for smart odour sensing
• Sensor plus AI to digitise scent
• Robotic AI hybrid smells sensor is 1000x more sensitive

“What’s in the air tells the truth,” said Alex Wiltschko, CEO and founder of Osmo. “Our AI sensors listen carefully, cutting through noise to confirm authenticity. They work when older methods fall short, helping businesses and customers get what the genuine products they’re paying for and deserve—every time.”

“Counterfeits are just the beginning,” said Rohinton Mehta, former Rapid Evaluation and Project Lead at Google X, now SVP of Hardware and Manufacturing at Osmo. “This same system can help ensure food stays fresh, protect the semiconductor supply chain, and even keep data centres running smoothly. The potential is enormous.”

Osmo’s scent sensor technology is trained to detect the right signals, cutting through background odours to focus on what matters, giving reliable yes-or-no readouts. Osmo offers the first bench of chemical sensors trained using AI, allowing Osmo’s sensors detect almost everything in the air, uncovering details others miss.

The sensors identify the unique chemical “fingerprints” of substances. This allows it to detect deviations from expected chemical compositions, making it invaluable for quality assurance and security applications.

The technology can be used for spotting fake products to ensuring food safety or monitoring air quality, guaranteeing products meet specific chemical standards for quality control and identifying counterfeit or adulterated materials.

Osmo was only founded in January 2023 with $60 million in Series A funding led by Lux Capital and Google Ventures along with angel investors including Henry Kravis, co-founder of private equity giant KKR, and the Bill and Melinda Gates Foundation.

Beyond fragrance, Osmo is applying its technology in commercial sectors for detecting counterfeit goods and in public health, where it is discovering new insect repellants. Osmo expects to expand into other industries as the potential for digitized scent grows.

Last week the company appointed Geoffrey Hinton, Professor Emeritus at the University of Toronto to its Scientific Advisory Board (SAB) to guide the company’s scientific progress at the intersection of deep learning and olfaction. Hinton, a Nobel Prize and Turing Award winner and Artificial Intelligence (AI) visionary, will lend his expertise on artificial neural networks.

“It is a great honour to continue to work with Geoff Hinton. He was a great supporter of our early work at Google Brain, and it’s fantastic to be able to draw on his expertise as we push the boundaries of what is possible in scent digitisation” said Wiltschko.