Many are founded by scientists with deep domain expertise and a small engineering team, funded by investors who expect a credible route from laboratory result to revenue.
That route is usually paved with software. Lab instruments produce data that must be captured, annotated and analysed reproducibly. Algorithms written in research notebooks need to become reliable services. Pilot customers expect a secure interface, user accounts and support. And regulators or pharma partners want evidence of how data was handled. For a Cambridge start-up, the risk is spending scarce scientific talent on infrastructure, or shipping a fragile prototype at the moment investors and partners start looking closely.