Korva is a scroll of the internet’s best ideas.
Every subscription software company's reports share the same lines: gross revenue added, churn, and net revenue added. Benchling had absolutely crushed it, and for the first six years the investor on its board worked with the company, there was never a churn line to report, because no customers left.
Then biotech crashed, and the company turned grindy. The goalposts moved, and the team had to do something different. They kept working on how to apply AI models to the world of their biotech and pharma customers, whom they know closely, and kept iterating until those customers were excited about it.
Why did biotech funding crash after 2021?
Drug developers spend years without revenue, so they depend on venture and public-market money. When interest rates rose in 2022, investors pulled back, and many biotech companies cut staff and spending. Software tools for labs are among the expenses that get trimmed. The SPDR S&P Biotech ETF, known as XBI, fell by more than half between its February 2021 peak and mid-2022.
Why does lab software rarely lose customers?
A lab's experiment records live inside its software, so leaving means moving years of structured data and retraining scientists. Those records also matter for patents and regulatory filings, so labs treat them carefully. In a hiring boom, every new scientist adds another seat for the software. Electronic lab notebooks can serve as dated evidence of an invention.
How are AI models used in biotech?
AlphaFold, from Google DeepMind, predicts the three-dimensional shape of a protein from its amino acid sequence. Its 2020 version solved a decades-old challenge in the field. Researchers use such predictions to guide experiments instead of waiting on months of lab work. Demis Hassabis and John Jumper shared the 2024 Nobel Prize in Chemistry for it, along with David Baker.
More from Invest Like the Best with Patrick O'Shaughnessy
