In the 1980s, Shafi Goldwasser co-invented “zero knowledge proofs,” a cryptographic breakthrough that, improbably, enables someone to prove a fact as true without revealing any information about that fact. For example, an investor seeking to prove her status as an accredited investor could demonstrate that her salary exceeds a certain minimum threshold while withholding the exact amount.
Now Goldwasser, a professor at the Massachusetts Institute of Technology who also holds posts at University of California at Berkeley and the Weizmann Institute of Science in Israel, is helping build a business that combines cutting edge data privacy with machine learning techniques. She’s co-leading a new startup, Duality Technologies, which aims to take a buzzy technology called “homomorphic encryption” mainstream.
Despite the jargon and complex mathematics involved, homomorphic encryption refers to a concept one can easily grasp, at least at a high level: the ability to perform mathematical operations on data that have been encrypted. Normally, when data are rendered into gobbledygook-like code, processing them becomes a challenge.
Cryptographers first conceived of homomorphic encryption in the late aughts. The idea was to let people manipulate encrypted data—running programs, crunching analytics, and extracting value—while keeping raw data secret.
The technology has become practical with advancements in areas of computer science such as multiparty computation, which lets computers split tasks between machines. With homomorphic encryption, “you can compute on data and collaborate without actually showing each other the data,” said Goldwasser, Duality’s chief scientist and a 2012 recipient of the A.M. Turing Award, computer science’s very own Nobel Prize-like honor, on a call with Fortune.
Assembling all-stars
Goldwasser’s cofounders include Vinod Vaikuntanathan, a colleague at MIT who co-invented a foundational homomorphic encryption scheme dubbed “BGV,” and Kurt Rohloff, founder of a key open source library of homomorphic encryption software called PALISADE. Another cofounder, Duality’s chairwoman Rina Shainski, a former venture capitalist and board member at many a tech startup, helped connect the team to its now-CEO, Alon Kaufman.
Kaufman, a neuroscience PhD-holder, said he believes homomorphic encryption has matured enough to enter the business world. “It was clear we had crossed a barrier and it is now prime time to go and commercialize it,” said Kaufman, who was formerly the global director of data science and innovation at RSA, a cybersecurity subsidiary of Dell Technologies.
Duality was cofounded at the end of 2016. Originally, its earliest members referred to the company as Genome Safe, a short-lived moniker that evoked the business opportunity in health care, where health records and DNA data demand privacy.
The team quickly realized the technology had potential applications in industries well beyond medicine, especially ones bearing similarly strict data regulations, such as in finance and insurance.
Adding another teammate
Duality said Tuesday it raised $4 million in a round of fundraising led by Team8, a venture capital firm and startup foundry established in 2014 by former leaders of Unit 8200, a top Israeli military intelligence unit.
This is the first investment out of Team8’s recently raised second fund, worth a total of $85 million. Team8’s backers include a host of corporate giants in a variety of industries, including Walmart, Airbus, Microsoft, Softbank, Nokia, Barclays, Munich Re, Cisco, and AT&T.
Duality’s cofounders connected to Team8 through Yuval Shachar, a founding partner of Innovation Endeavors, a venture capital firm created by former Google CEO Eric Schmidt, Kaufman told Fortune. (Innovation Endeavors was an early investor in Team8, and Shachar has since become Team8’s executive chairman.)
Shachar expressed exasperation with the status quo in corporate data-sharing. “It’s about time companies are able to collaborate on data without having to trust that others won’t misuse it,” he said in a statement.
Duality started receiving venture capital from Team8 early on in its life, though the funding round officially closed only in May of this year. That same month, Duality also received a grant from the National Institutes of Health for a project involving privacy-oriented work on genome analysis.
With its second fund, Team8 plans to set up or invest in about eight cybersecurity or data-focused companies over the next five years. Last month Team8 celebrated its first exit, a sale of Sygnia, a cybersecurity startup incubated as part of the first fund, to Singaporean investment company Temasek for $250 million.
Duality is not the only company working on homomorphic encryption technology. The research arms of tech giants such as Microsoft and IBM have active projects, and rival startups such as Enveil and Inpher are pursuing their own versions of the technology.