Revenues during 2020 tripled at the UK-based AI unit.

DeepMind, the UK AI unit which Google (NASDAQ: GOOGL) acquired in 2014 for £400 million, has reported a maiden pre-tax profit of £46 million in 2020. Commercial revenues at DeepMind emanate from parent Alphabet, and those more than tripled to £826 million. Staffing and other costs surged to £780 million, up from £717 million in 2019. (FT)

Though the DeepMind’s pivot to profitability is encouraging, Google wrote off £1.1 billion of its debt in 2019. Since its acquisition, cumulative losses at the unit amount to nearly £2 billion.

Former video game developer and child chess prodigy Demis Hassabis (pictured above) founded DeepMind in 2010 and the startup soon earned a reputation for its cutting edge research in AI and machine learning. That led to Google acquiring the company.

In recent years, it has started to earn income streams from applications of its research.

DeepMind said that it was improving the lives of billions of people with products and infrastructure derived from its efforts.

DeepMind: Path-breaking research on proteins

“During this reporting period we made significant progress in our mission of solving intelligence to accelerate scientific discovery,” DeepMind said. “Our groundbreaking results in protein structure prediction were heralded as one of the most significant contributions AI has made to advancing scientific knowledge.”

Last year, DeepMind evolved AlphaFold, an AI solution to the 50-year “protein folding problem.”

“AlphaFold is a once in a generation advance, predicting protein structures with incredible speed and precision,” said Arthur D. Levinson, PhD, Founder & CEO Calico, Former Chairman & CEO, Genentech. “This leap forward demonstrates how computational methods are poised to transform research in biology and hold much promise for accelerating the drug discovery process.”

Rain “nowcasting”

Earlier this month, DeepMind announced DGMR (deep generative model of rainfall), its ‘nowcasting’ system for predicting, accurately, the likelihood of rain over the next 90 minutes, developed in collaboration with the UK Met Office.

DGMR correctly predicted various rainfall parameters – such as location, extent, movement, and intensity – over 89% of the time.

Related Story: DeepMind’s AI Solves 50-Year-Old Biology Mystery; Medical Breakthroughs Could Follow

Image Credit: DeepMind