A “To-Let” sign in the window of a shop is not unusual these days as businesses reel from the impact of the pandemic. But an abandoned Humain Health lab in HSR Layout, Bengaluru’s newest startup hub, has a more complicated backstory—its parent SigTuple, a promising startup, coming undone.
Founded in 2015, Bengaluru-based medtech startup SigTuple has raised more than $40 million on the promise of using artificial intelligence (AI) to analyse blood, urine and semen in pathology labs. The goal was to slash the cost of medical testing and solve India’s chronic shortage of pathologists by creating an affordable device that could automate the visual analysis of medical samples.
But, in 2019, things took a puzzling turn when the company launched a chain of diagnostics centres across India designed to be “the Apple store for diagnostic labs”. By November, the company had opened four labs in Bengaluru, one each in Delhi, Hyderabad, Chennai and Mumbai; it also installed its equipment and branding in four other partner labs in its home city. A spree of mergers, acquisitions and expansions gave birth to the “SigTuple group of companies” with interests in everything from home sample collection to corporate wellness programmes and lab equipment distribution.
According to several former employees—who requested anonymity due to fears of retribution—and industry veterans, none of SigTuple’s verticals had an obvious path to profitability and some were haemorrhaging money. SigTuple refutes this claim. While it saw a marked rise in revenue from Rs 83.5 lakh ($110,000) in the year ended March 2018 to Rs 5.65 crore ($744,000) in the next year, its losses widened to Rs 9.56 crore ($1.2 million) in the year ended March 2019, as opposed to Rs 6.41 crore ($844,000) the year prior.
Despite the company touting the AI technology in its labs, insiders say the company’s device was largely unused in everyday operations. It’s not easy to develop and deploy a device so fast.
Yossi Pollak, CEO of Israeli start-up Sight Diagnostics, which has commercialised an AI-powered device for carrying out the same kind of blood analysis as SigTuple, says it took them 10 years to develop their product. Beyond proprietary algorithms, the company uses fluorescent microscopy and a specially designed cartridge that makes sample collection foolproof. “We worked for a decade to perfect the technology because we did not want to release a product that could possibly misdiagnose or underdiagnose our patients,” says Pollak.
While people with knowledge of SigTuple’s technology agree it held genuine promise, it soon became a forgotten child in an overambitious company.
Several former SigTuple employees say problems appear to have come to a head towards the end of 2019 when investors told the company further funding would not be forthcoming.