TechBio vs. Biotech: What’s Real, What’s Hype — with Kavya Sharman

In our latest episode of Digital Health Interviews, Kavya Sharman, PhD, Founder & Managing Partner at Phase Capital, unpacks what “techbio” really means, how it differs from traditional biotech, and why generative AI is already reshaping the way science gets done in the lab. Along the way, she shares her unconventional path from aspiring radiologist to data scientist and techbio investor, and gives candid advice for early-stage founders.

From Radiology Dream to TechBio Investor

Sharman jokes that her career started “completely by accident.” At seven, she decided she would become a radiologist: a safe, well-paid medical specialty where, as she puts it, “I didn’t have to touch people.” That plan held until her sophomore year of undergrad, when she shadowed in a hospital and realized the day-to-day clinical environment wasn’t for her.

At the same time, she had just started at Vanderbilt and heard that nanotechnology would be “the future.” Studying biology, she found a lab working on nanoparticles to deliver gene therapies for brain cancer. That experience pulled her into neuroscience, where she became fascinated by why different molecular mechanisms make people think and behave the way they do.

When she had to choose a neuroscience lab, she didn’t want to abandon the nanotechnology project and asked: can I just put these nanoparticles into the brain? To her surprise, the answer was yes. She designed a research study combining these two worlds, and that experience - taking one piece of breakthrough technology and bringing it into a new domain - turned out to be her first taste of entrepreneurship, even though she didn’t yet know the term.

After undergrad, a mentor invited her to work at Vanderbilt University Medical Center in business development. There, she saw how innovation ecosystems function: how academic informatics and early “big data” precision medicine work could be translated into real-world impact.

Later, at a startup, Sharman moved to the operator side and began teaching herself to code in MATLAB to handle growing amounts of data. That self-taught “ceiling” pushed her back to Vanderbilt for a PhD in data science, knowing all along she wanted to return to business and startups, but this time with a much stronger technical foundation.

During her PhD, she and a group of peers built a small student-run investment fund. Students acted as LPs, and they invested in publicly traded biotech companies, using their scientific expertise to try to “outcompete Wall Street.” Days were spent coding and analyzing data, nights were spent learning valuation techniques and DCFs. That experience became her gateway into investing, and eventually into venture capital focused on early-stage companies.

TechBio vs. Biotech: A Paradigm Shift

Ask three people in the industry what “techbio” is and you’ll probably get three different answers, Sharman admits. But to her, there’s a clear conceptual distinction.

Biotech, in its “truest” form, starts with a biological system and a biological question:
you take something from biology, engineer or modify it outside the body, then bring it back into the body to make things better. Think traditional therapeutics, wet-lab biology, and interventions that operate inside living systems.

Techbio, on the other hand, starts with data, infrastructure, and computational capabilities, often including AI. Instead of asking, what’s this one biological pathway doing?, techbio companies ask, what can we learn from all the data we have, and where can we go from here?

In practice, these are often technology companies whose customers are pharma or biotech firms. The business model might look like B2B SaaS, but the end users are drug discovery and development teams rather than typical enterprise software buyers.

Sharman points to one of Phase Capital’s portfolio companies as an example. Form Bio originated as a spin-out from Colossal Biosciences, the company known for its ambitious efforts to “bring back” the woolly mammoth. Colossal’s work involved inserting “woolly” genes into a mouse, and ultimately aims to edit elephant embryos to recreate mammoth-like traits.

Behind that scientific spectacle, however, is a powerful gene-editing and machine learning platform. This is where Form Bio comes in: instead of developing drugs themselves, they partner with biopharma companies to radically shrink the search space in early-stage drug discovery.

Historically, arriving at a viable early-stage candidate could take two to three years and several million dollars just to get a short list of molecules. Companies like Form Bio integrate large, heterogeneous datasets, from discovery experiments, manufacturing, safety, toxicity, and even mass spectrometry, to prioritize candidates that are not only promising scientifically, but also less likely to fail later in Phase III trials. That holistic, data-driven optimization across the entire drug development pipeline is what Sharman sees as emblematic of techbio.

Business-model wise, these companies typically mix B2B SaaS or fee-for-service contracts with milestone payments and revenue-sharing downstream if a partner’s drug succeeds.

GenAI in the Lab: Speed, Scale, and New Ideas

Sharman has lived through the pre-GenAI and post-GenAI eras in research. During her PhD, running a single model on an imaging dataset could take 24 hours. A huge amount of effort went into simply figuring out how to chunk the data into pieces computers could handle.

Now, with modern compute and tools like code-assistants, that same work can be done in seconds or minutes. She’s of two minds about it: grateful to have learned the fundamentals “by hand,” but also very aware of how many more experiments she could have run if today’s tools had been available.

When asked whether AI can go beyond just executing tasks and actually generate new, non-obvious ideas, suggesting novel experimental sequences or hypotheses a human never thought of, Sharman’s answer is clear: we’re already there.

Large language models can digest enormous bodies of literature, including obscure publications in other languages or journals outside a researcher’s usual orbit. That allows them to surface previously overlooked connections and relationships. Combined with multi-dimensional pattern recognition across complex datasets, GenAI can propose hypotheses and experimental directions that might have been invisible with traditional methods.

The real challenge now, she says, is figuring out which AI-generated signals are meaningful and which are hallucinations, essentially building human + machine workflows that filter and validate the most promising ideas.

Will AI Cure Most Diseases in 10–20 Years?

On the big question—are we heading toward a world where AI helps cure most chronic conditions within the next two decades? Sharman is optimistic but grounded.

She highlights the shift from single-target thinking to systems biology. Early in her work on glioblastoma, the focus was on one pathway, one gene, one DNA mismatch. But real disease is a dynamic, systemic phenomenon: stress hormones, immune responses, medications, and more interact in a complex “war inside the body.”

Historically, we simply didn’t have the compute power to model that complexity in a meaningful way. Now, with AI and high-performance computing, we’re getting closer to understanding those multi-layered systems.

Sharman believes we will see meaningful cures and breakthroughs in the next 10–20 years. but the how is still hazy. Some of the future solutions don’t exist yet, and others may be impossible for us to imagine with today’s mental models.

At the same time, she’s very aware that these powerful capabilities can be misused. Asked directly whether AI could also increase the risk of a new, human-made pandemic, she doesn’t elaborate, but her answer is an unambiguous yes.

Phase Capital, Frontiers, and the Power of Ecosystems

At Phase Capital, Sharman and her co-founder Kristin Johns invest in early-stage techbio founders, primarily pre-seed and seed, with some opportunistic Series A. The fund is based in Nashville but invests across the United States, leaning heavily into ecosystem building.

Different regions have different strengths, she notes: a brilliant founder in Chicago might need to meet an investor in New York and a world-class engineer in Austin. Much of her energy goes into connecting those dots.

Today, she spends roughly 50% of her time fundraising and 50% working with founders, but she expects that ratio to tilt toward founders over time. In-person gatherings are a key channel for sourcing and supporting startups: they partner with large events like JPMorgan conferences, host invite-only meet-ups, and run their own Frontiers Summit in Nashville.

Frontiers Popups, smaller events in cities like San Francisco, are designed as “micro moments” of that summit: carefully curated groups of people, good spaces, and plenty of room for serendipitous conversations.

Gut Feelings, Data, and Founder “Signals”

When everything in a deck looks good on paper but a decision still feels uncertain, Sharman doesn’t pretend it’s all about numbers. She openly admits she trusts her gut.

For her, gut feeling isn’t mystical; it’s a machine-learning metaphor. The human brain, like a model, is constantly ingesting data and building internal representations. That “first instinct” often reflects a complex, multi-dimensional pattern that’s hard to articulate: small red flags, subtle signals, and previous experiences converging.

You might not be able to point to a single reason, she says, but your internal “model” is still flagging something important. The work is to give that feeling time and energy, and then back it up with explicit reasoning where possible. “Once a data scientist, always a data scientist,” she laughs.

Her Advice to TechBio Founders: Get Out and Meet People

Sharman closes with one simple but powerful recommendation for techbio founders:

Meet as many people as you can.

Every job she’s had and every person who has invested in her started with a human connection, not a pitch deck. That doesn’t mean being random, founders should be intentional about where they spend their time, but it does mean carving out space for real relationships, both inside and outside their immediate ecosystem.

Whether it’s Phase Capital’s Frontiers events, other techbio meet-ups, or local healthcare gatherings, she urges founders to put down the remote, step away from Netflix, and walk into the room where things are happening.

Because in techbio, as in most of life, breakthroughs rarely emerge from isolation. They’re born at the intersection of ideas, disciplines, and people who are excited about what they’re building and why it matters.