AI Skin Model Trained on Living Human Tissue
Michael Polansky, the co-founder of Outer Biosciences and partner to Lady Gaga, is developing an AI skin model trained on living human tissue that stays alive outside the body for over a month. His startup has built a closed-loop system that combines extended-life tissue samples with machine learning to discover new cosmetic ingredients faster than traditional methods.
From Hedge Funds to Living Tissue
Polansky’s path to biotech wasn’t direct. After graduating from Harvard in 2006 with degrees in applied mathematics and computer science, he spent three years at Bridgewater Associates. A chance connection through Sean Parker’s assistant led him to Silicon Valley, where he joined Founders Fund and later ran Parker’s family office.
His work in cancer immunotherapy through the Parker Institute for Cancer Immunotherapy sparked his interest in life sciences. That work eventually connected him to Cynthia Germanotta, who insisted on setting him up with her daughter — Lady Gaga. The couple now collaborates professionally and personally, with Germanotta serving on Outer Biosciences’ board.
How the AI Skin Model Works
Outer Biosciences, co-founded in 2022, sources human skin discarded after surgeries — primarily plastic surgery — through nonprofit biobanks with documented donor consent. The tissue arrives de-identified within hours of surgery and remains viable for up to a month, far exceeding the industry norm of just days.
This extended window allows researchers to study slower biological processes like collagen remodeling and barrier repair that typically take weeks to unfold. The company’s proprietary support system feeds nutrients to the tissue while removing metabolic waste.
The AI skin model operates through a continuous feedback loop:
The AI predicts which untested chemicals will benefit specific skin functions
Those chemicals undergo testing on the living tissue
Results — whether correct or incorrect — feed back into the model
Each cycle improves the next round of predictions
Discovery at Unprecedented Speed
Early on, Outer Biosciences relied on a “brute force” approach, mining scientific literature and partnering with the National Cancer Institute. That method produced a couple of leads over 18 months.
With the AI skin model integrated, the company now generates a new candidate roughly every six weeks. Currently, six leads are active in the pipeline, with several dozen additional “hits” logged.
The discovery potential is significant. Only about 120 to 130 active ingredients are approved by the FDA for over-the-counter skin drugs, and adding cosmetic ingredients backed by actual research brings the total closer to 200. This relatively small universe presents enormous opportunities for new discoveries.
Commercial Strategy and Partnerships
Outer Biosciences doesn’t plan to launch its own consumer brand. Instead, it will license or sell finished ingredients to beauty and pharmaceutical companies that formulate them into products like serums and creams. Four of six current leads appear likely to reach commercialization.
The company generates revenue through collaborative research partnerships, including:
A pharmaceutical partner studying why certain cancer drugs cause severe skin rashes
Consumer beauty brands validating Outer Biosciences’ data against their product development needs
Competitive Landscape and Data Advantages
The company has raised approximately $23 million to date from investors including Wing Ventures, Initialized, and Polansky’s own Hawktail. It employs 19 people, with most based near Cambridge, Massachusetts.
Competitors like Vivodyne are pursuing related approaches, but Outer Biosciences’ use of real tissue rather than synthetic alternatives distinguishes it. Unlike AI companies scraping the internet, there’s no “biology internet” to scrape — the data Outer Biosciences generates simply doesn’t exist anywhere else.
This makes the company’s position more defensible, though slower to build than traditional software startups. All AI work runs on-premise rather than in the cloud, because “we don’t want the data in the cloud,” Polansky says.
Polansky envisions a predictive model accurate enough to spot promising directions in skin biology without physically running every experiment first. This would open up a rate of discovery in dermatology that doesn’t currently exist.
Building a product-development team with experienced talent is next on the roadmap. After years of operating in near-total silence, Polansky is ready for the world to know what Outer Biosciences has been building.
“I think it’s going to be fun,” he says, “to have people know that this is what we’ve been doing.”
