Biodata Lessons from the Rise and Fall of 23andMe
An expensive case study in why business models matter for biodata companies
It’s been a rough few years for 23andMe. After going public in 2021 and peaking at a $6 billion valuation, the company’s value has plummeted to just $35 million. On March 24, 2025, they filed for Chapter 11 bankruptcy. What happens to their assets—most notably, the genetic data of 15 million customers—remains to be seen. An auction has been proposed for mid-May 2025. For many, the downfall wasn’t exactly a surprise. The warning signs were there: the company had never turned a profit since its founding in 2006, faced internal conflicts, and saw the resignation of all independent board directors in September 2024.
And yet, the mission resonated with so many. Who wouldn’t want to “help people access, understand, and benefit from the human genome”? It sounds empowering. If our genetic codes are the foundation of who we are, then unlocking that code could help us better manage health risks tied to our biology. But as with most things in science and medicine, it’s not that simple. Even 23andMe’s own disclaimers clarify that their reports aren’t a substitute for seeing a healthcare professional. Understanding health isn’t just about the genome—it’s also about how genes are transcribed, how proteins are regulated, how cells interact, and that’s a simplified version. These layers are not captured in a consumer-facing genetic test.
The Challenge of the Two-Sided Data Business Model
If you’ve chatted with us about 23andMe before, we’ve likely shared some skepticism—primarily about their business model. As a direct-to-consumer business, the one-time nature of the product puts a cap on revenue. Aside from rare situations—like specific cancers—your genome doesn’t change. So, once a customer receives their results, there’s little reason to come back. They did try to build a recurring revenue stream through memberships, but the value proposition just wasn’t compelling enough. It didn’t pass the “so what?” test. Customers get their health reports—so what?
Originally launched at $1,000 in 2007, the test price dropped to $299 by 2012 and now sells for $199. Even though they reported ~50% gross margins on around $200 million in annual revenue since 2021, their net income has always been deep in the red. The business model simply never added up.
Where things could have worked was on the pharmaceutical partnership side. About 80% of customers opted in to share their genetic data and self-reported health information. In 2015, 23andMe launched a drug discovery business, and in 2018, they signed a partnership with GSK to use this data for R&D. That was one of the big selling points behind their 2021 SPAC. Two drug candidates even made it into clinical trials—but by 2024, their entire therapeutics effort had shut down.
In hindsight, 23andMe’s pharma partnerships may have come too late. Their scale had already grown beyond what’s typical for an acquisition, and their lack of early alignment with clinical work made it hard to generate real traction. The underlying challenge? This was always a two-sided data platform model, and those are notoriously difficult to make work.
The Missing Clinical Link
Even beyond the business model, the lack of clinical integration was a major issue. While GSK and 23andMe cited a study showing that genetically supported targets have 2x the success rate in drug development, the truth is that their customer data—tied to self-reported conditions rather than clinical diagnoses—may not meet the bar for clinical quality. This highlights one of our core investment theses at Creative Ventures: if a healthcare solution doesn’t align incentives across patients, providers, and payers—and doesn’t plug into the clinical ecosystem—it’s extremely difficult to drive meaningful outcomes.
In another world, if 23andMe had started with sequencing patients within the healthcare system, their story might have played out very differently. Of course, that’s an uphill path—especially in the U.S., where insurance and reimbursement dynamics make it incredibly tough to pull off.
From DNA to Drugs: Easier Said Than Done
And let’s not forget the science. Even when a genetic variant is known to be associated with a disease, that’s just the beginning. Understanding how that gene impacts its associated protein, how the protein affects the mechanisms, and what molecules could potentially intervene—that’s an enormous lift. DNA data alone is not a silver bullet for drug discovery. In fact, many symptoms can’t even be traced to genetic roots.
What We Can Learn
It’s easy to pick apart a failed company in hindsight. But 23andMe offers real lessons—especially for those of us building and investing in biodata businesses. Most notably, the business model matters. How a company makes money and how it eventually exits are critical. Reinventing a business model is hard (though not impossible). In this case, it didn’t work out.
Equally important is clinical relevance. In health-related applications, a product must deliver meaningful and actionable outcomes, unless it’s going after the customer lifestyle market. And finally, the science has to hold up. If a company aims to bridge the gap between DNA data and therapeutics, its R&D should reflect that from day one. As we keep repeating ourselves here at Creative, deep tech startups don’t have the luxury of SaaS-style pivots.
Our Portfolio Companies in the News
AM Batteries announced a partnership with TDK to advance energy-efficient battery manufacturing.
AM Batteries’ proprietary dry coating technology eliminates the need for solvents and electrode drying – reducing energy consumption by 75 percent and capital expenditures by 30 percent in the battery electrode manufacturing process. Through this collaboration, AM Batteries and TDK will leverage their respective expertise to further optimize battery production and drive advancements in next-generation energy storage solutions.
NxLite named to Cleantech Group's prestigious "50 to Watch.”
NxLite was highlighted in the Materials & Chemicals category for their "thermally and optically smart Diamond-Like Carbon (DLC) coating for windows to control heat and light emission." The recognition specifically notes how their technology provides an innovative approach to energy efficiency in buildings and beyond.
Fort Robotics was highlighted as one of the key solutions to lighten the labor shortage challenge.
“We’re seeing remote operation from line-of-sight and tele-operation, as well as task automation for repetitive jobs, help skilled laborers focus on tasks where their skills are most needed,” said Doug Riffle, vice president of business development at FORT Robotics
And Other News
SXSW was a success! James was at SXSW in March, moderating an all-star panel on the future of AI, with around 200 people turning out—even with Peter Attia and other big names literally next door.
Reach out to catch up with us on his hot take and the latest from the conference. Listen to a recording of the panel here!
Catch Us In-Person!
We’d love to grab coffee and catch up if you’re in town. Here are our upcoming traveling schedules. Please do reach out!
Seattle: Apr 9 - 11
Japan & Asia: Apr 20 - 28
LA and Palm Springs: Apr 28 - May 1
Creative Ventures is a deep tech venture firm investing in early-stage companies for human betterment, bringing the advanced technologies to where the people are and maximizing the value they can generate for the world. Our focus within deep tech opportunities are in Real-World AI and Biodata; bringing the full promises of the advancement in Large Language Models to the physical and biological worlds.