A partner at a $630 million fund reveals his firm's secret to scouting billion dollar companies - and says that the rest of venture capital is 20 years behind
- EQT Ventures has developed Motherbrain, an AI-powered software program to help it find new startups in which to invest.
- Motherbrain is treated as a partner at the firm, and even has her own e-mail address.
- EQT Ventures partner Alastair Mitchell believes that the venture capital industry is 20 years behind other kinds of investors because it doesn't embrace algorithms or data in making deals.
- Some other VCs believe that no machine can replace the human touch in dealmaking.
Stockholm-based investment firm EQT Ventures has 30 employees. Among them, the firm has one partner who stands out from the rest.
Like EQT Venture's other partners, she has a company email address, a short bio on the company's website, and a record of insightful market predictions.
But there's one small catch: She isn't actually human.
Her name is "Motherbrain" and she's a powerful, data-driven machine overseen by a team of data scientists and software developers. Her job description: Use all of that data to predict which startups will become billion dollar companies, faster than anyone else in venture capital.
"It's a 'she,' actually. We refer to her as a 'she,'" Alastair Mitchell, a founding partner at the firm said in a phone conversation, after I referred to Motherbrain multiple times as "it."
"She's a core part of our thesis," he continued. "When we founded EQT we wanted to disrupt European ventures. We had multiple approaches to disrupting venture capital, and one of our main interests was the way in which software and AI will disrupt venture capital itself."
Already, Mitchell said that Motherbrain has found two formerly undiscovered companies for the firm, one of which is currently unannounced. The other company is a remote desktop service called AnyDesk, which EQT Ventures led a €6.5 million investment in earlier this year.
"When we invested in them, they were a team of 15 people based out of a small town in Germany," said Mitchell. "They weren't on anyone's network. They weren't on Pitchbook or Crunchbase, they hadn't raised any angel money or talked to investors. But they had built an amazing bit of software."
With Motherbrain, Mitchell and his firm have taken a bold stance in an important debate at the center of the venture capital world: In an industry that drives technological innovation, how much should technology itself influence the way firms invest?
'The VC fund of the future will not be based purely on the old boy's club of internal partners'
For EQT Ventures and its team, data is the key to not only informing whether or not a startup will be a good investment in the future, it's a route to discovering great companies as well.
"Clearly the more startups you see, the more likely you are to find the better deal," said Mitchell. "It's a data-based problem. The more data you have, the more companies you see, and then you can make better decisions faster."
Typically, investors rely on their professional and social networks to hear about interesting companies and notable new technological innovations. But even the most well-connected investor will fall short of a software program that's tasked with the exclusive job of combing the internet in search of undiscovered startups.
"It seems totally ludicrous that this process of discovering companies could not be dramatically improved through the use of modern data and AI," said Mitchell. "It's a pretty obvious conclusion that the VC fund of the future will not be based purely on the old boy's club of internal partners, but on the data and software they use to source and pick those companies."
While Mitchell declined to disclose the exact specifics of how Motherbrain's proprietary technology works, he said that its AI-powered software discovered AnyDesk's software by tracking down the large number of web downloads it regularly received.
"We wouldn't have been able to discover this company without Motherbrain, because there's no way we would have known to look for it," said Mitchell.
Mitchell also said that Motherbrain's data revealed that AnyDesk was poised to outpace a $2 billion competitor in its field, meaning that EQT was going to get in on the deal early.
In addition to AnyDesk, Mitchell credits Motherbrain with one of EQT's best investments in a fast-growing Finnish gaming company called Small Giant Games.
"It's an unbelievable unicorn company. It's gone from taking in $50,000 in a month to $15 million a month in just over a year," he said.
Mitchell said that the decision to invest was almost entirely based on Motherbrain's insight.
"We entirely made the decision to invest in them based on data," "We looked at the trends within the data, and the potential monetization. From that information, we were able to make an investment."
Can a machine invest as well as a human?
While data typically plays a large role in traditional venture capital, firms differ in their approach.
In multiple conversations with Business Insider, many venture capitalists have said that their investing decisions are based on a number of contributing factors. Tangible data like metrics and analytics are big part of the process, they say, but there's also a fundamental human element that goes into calling a good deal, as well.
Some call it intuition. Others describe it as emotional intelligence. Still others describe it as instinct; a gut feeling that guides the final decision.
Or, as Deven Parekh, managing director at Insight Venture Partners puts it: "At the end of the day, for now at least, venture capital is still a human decision business. Determining a good deal still requires human judgement."
But does it?
According to Mitchell, venture capital's current reliance on human insight has placed the industry abysmally behind the data-driven approach traditionally used by institutional investors.
"Venture capital is probably 20 years behind given the fact that the stock market has been trading algorithmically for many years," said Mitchell. "Right now, the classic phrase that investors would use is that the industry is about 98% human pattern recognition. 2% is based on data."
This paradigm, said Mitchell, is deeply flawed. While Mitchell said that EQT Ventures currently relies on a 50/50 split of human insight and AI-powered data, he suggests that an ideal investing model would rely more heavily on algorithms and machine-driven analytics.
"I think we're going to end up around 80% data and 20% human insight and personal experience," he said. "In venture capital, we're investing in technology, but the industry itself is so far behind. The irony is never lost on me."
Not every VC loves having that much data
While most venture firms typically use at least some data to inform their deals, there's been debate over just how much of that information should be a part of the decision-making process. After Silicon Valley venture firm Social Capital announced plans to move to a more data-informed investment approach last year, several partners left the firm. In conversations with Business Insider, multiple people familiar with the firm suggested that the departures had resulted in part due to its move to data-driven investing.
"As soon as you move to algorithms and you're just purely data-driven, you're going to lose the sensibility of a company," one person familiar with the departures who asked not to be identified told Business Insider at the time. "I don't care if you're the founder of a tech startup or a VC. Being disconnected from the touch is damaging."
Social Capital is far from being the only venture firm to express a growing interest in data. GV (formerly Google Ventures) has long distinguished itself as leader in data driven investing. In a 2013 interview with the New York Times, a managing partner at the firm said that data was a critical component of its decision-making process.
"We have access to the world's largest data sets you can imagine, our cloud computer infrastructure is the biggest ever," he said. "It would be foolish to just go out and make gut investments."
This could be the future of VC
Mitchell predicts that data-driven investing is still only in its very nascent stages. In the upcoming years, venture firms will run with a deeper reliance on AI-sourced insights, he said.
"In the future, you're going to see the rise of small startup investment shops that will be run exclusively by engineers," he said. "They'll say, 'Oh, we can run this better than everyone else and raise money off the back of the software we make.'"
Not every investor seems sure about the algorithmic future of investing.
"Oftentimes, the best deals aren't clear," said Insight Venture Partner's managing director Deven Parekh. "When you put an algorithm against an investment, I worry that the algorithm will focus on the lowest common denominators in the investment process."
While Parekh said that data informs a large part of how his firm invests, he's speculative of the idea that AI-driven investing could be the future of venture capital.
"But who knows?" Parekh continued. "Will this evolve over the next thirty years? I don't know. But there's probably a lot of livelihoods that are at risk before we get to venture capital."
For Mitchell, running a venture firm that actively cultivates its own proprietary tech is essential not only to investing but to attracting startup founders as well.
"Our founders love it," he said. "They're on the cutting edge of data. They love that we have a team of 10 data scientists, and that we're iterating and testing our own software. We're not sitting in an ivy tower saying that we have the best insight based on our own gut feeling."