Google Ventures is on the hunt for the next DeepMind in Europe

Google Ventures had five partners in Europe when it announced Google Ventures Europe but today there are only two left: Tom Hulme and Avid Larizadeh.

Business Insider met Hulme and Larizadeh at a lunch in London with The Financial Times, and the BBC. During the lunch at Dean Street Townhouse, the duo were asked about what happened to the dedicated Google Ventures Europe fund; what their future investment strategy looks like, and whether they're really using artificial intelligence to identify the next companies to invest in.

Doubling down on DeepMinds

Fairly early on in the conversation, Hulme said he's keeping an eye out for companies like DeepMind — an AI research lab in London's King's Cross neighbourhood that Google bought for a reported £400 million in January 2014. "We would absolutely happily invest in the next DeepMind," he said. "Happily."

DeepMind is creating self-learning algorithms that can learn from raw experience or data. These algorithms are "general" in that they can perform well across a wide variety of tasks straight out of the box. The company is quite unlike the majority of the startups that Google Ventures has invested in across Europe in that it's more research-focused and less consumer-facing.

Conversely, neither Hulme nor Larizadeh have been able to find an AI company in Europe that's worthy of Google Ventures' millions.

In order to build a "really interesting general artificial intelligence project, you need some of the best PhDs on the planet," according to Hulme. "There will be people who find an elegant way to build a business while growing their team but I’ve not found it yet. DeepMind were really the example but Google bought them. I would hope that were we around when DeepMind was going we would have looked at them."

Hulme said he is interested in a specific type of AI known as AGI, or artificial general intelligence. AGI— the area DeepMind operates in — is an emerging field focused on the building of "thinking machines." These machines can be thought of as general-purpose systems with intelligence comparable to that of the human mind. While this was the original goal of artificial intelligence (AI), the mainstream of AI research has turned towards more niche applications and is often referred to as applied AI.

"Applied [AI] is injected in some way into most businesses now," said Hulme. "Two years ago at accelerator demo, every pitch deck said big data in it. This year every pitch deck says artificial intelligence."

The fact that Google already has DeepMind isn't an issue for Google Ventures, according to Larizadeh. "We don’t stop because Google is doing something unless we truly think that Google has a massive advantage," she said.

Previous investments

Google Ventures has made eight investments in Europe. The investments — ranging from £200,000 up to tens of millions of pounds, but averaging out at around £8 million — include:

1. Kobalt: Offers artists, songwriters, and publishers an alternative to the traditional music business model.
2. Lost My Name: Personalised children’s books.
3. Oxford Sciences Innovation: Dedicated £300 million to boost development of science and technology businesses in Oxford’s math, life sciences and medical science divisions.
4. Resolution Games: VR gaming studio based in Stockholm.
5. Secret Escapes: UK-based members-only luxury travel site.
6. Yieldify: Predictive marketing technology.
7. Cambridge Epigenetix: Biosciences company that aims to be the leading provider of high quality, easy to use epigenetic tools.
8. Weaveworks: Virtual private networks (containers). Enable customers to network, visualize and manage microservices and cloud native applications on Docker containers. GV’s newest portfolio company in the European region. 

Hulme spoke proudly of the eight investments Google Ventures has made in Europe but says that Google Ventures has the resources to go after more technical startups that are developing revolutionary technologies, like quantum computing, general AI, and futuristic travel concepts like Elon Musk's Hyperloop.

"Sometimes it’s easy just to do the stuff that you know," he said. "But we as a firm should be doing some of that stuff."

Forrays into fintech

Hulme said he's also interested in fintech, which is an area he focused on during his time as an angel investor.

"As an angel, I did 22 investments and probably about eight of those were in fintech," said Hulme. "So when I started [at Google Ventures Europe] I was convinced that half of what I do would be in fintech in London."

Like AI, Google Ventures is yet to back any fintech companies in Europe. Hulme puts this down to the fact the valuations are too high. "We looked at a lot of businesses and our concern collectively was around valuation," he said, adding that Google Ventures assessed most of the peer-to-peer payments startups, and most of the crowdfunding startups .

"I would still be bullish about fintech in London because of the talent, the proximity, the number of potential acquirers," said Hulme. "If you look at media companies, there’s probably three or four companies that could buy you for a $1 billion in London. In fintech there’s probably 20. That’s going to play out in the market over time. We haven’t even seen that effect yet."

Larizadeh, who said she recently had dinner with the CEO of Barclays and the chairman of Aviva, added: "There’s so much still to happen. From partnerships and small acquisitions. There’s still a lot of time for Tom to invest in fintech."

Hulme went on to say that Google Ventures will make a fintech investment at some point, adding that he thinks valuations are already starting to fall, particularly at the later stage.

GV's rocky road in Europe

Last December, the Google Ventures Europe fund — initially overseen by a team of five investors — was folded into the main US Google Ventures fund, with numerous reports suggesting that tensions had developed between Google Ventures investors in the US and Google Ventures investors in the UK over a lack of autonomy.

"The strategy hasn’t changed," said Larizadeh early on in the lunch, trying to dispel what's been reported by the media about how Google's European investors struggled to get sign off on deals from Bill Maris, the California-based Google Ventures boss.

"What changed was we just pulled the money together. When we initially started, we were investing on the same strategy as the US was, but in Europe. So stage-agnostic, mostly starting with series A and then B, C, D. And looking across technology. So whether it’s containers, robotics, AI, consumer, enterprise, and life sciences as well. That hasn’t changed in any way shape or form. We merged."

Larizadeh added: "We went from $300 million (£205 million) [in the US] and $125 million (£85 million) [in Europe] to $425 million (£290 million) globally."

The merge occurred for a number of reasons, according to Larizadeh and Hulme.

It "didn't make sense" to pay for two sets of lawyers and two sets of "finance costs," said Larizadeh. In addition, all of the resources that Google Ventures can offer to portfolio companies in terms of PR, marketing, design, engineering, product are based out of the US. The meant European companies only get access to them three or four times a year. "The incentives [to use those resources] lay in the US fund not the European fund," she said. "It makes sense to have a unified fund for everybody to be incentivised in the same way."

Interestingly, investors working at Google Ventures in the US were co-investing in European startups alongside Google Ventures European investors. "I think as we sort of think about the structure of the fund, that was the biggest driver," said Hulme. "It’s like we now understand them and like the US investments; they understand us, and like the European investments, so wrap the two together."

Venture capitalist Mark Tluszcz, cofounder and CEO of Mangrove Capital Partners, said the significance of the folding of Google Ventures Europe into the Google Ventures global fund was overhyped.

"They just had an operating model and it wasn’t working for them so they changed it. I think Google is — like most American companies — very focused on its home market."

He added that five partners for the Google Ventures Europe fund may have been a bit excessive.

How Google Ventures finds companies to invest in

From looking at random emails to attending startup demo days, there are a range of different methods that investors use to find companies to invest in.

Reports suggested that Google Ventures had created a robotic algorithm to find the most promising companies to back.

When asked about these reports, Hulme said: "Look, we have 12 world-class machine learning and engineering guys who spend a lot of time helping the portfolio and when they’re not helping the portfolio companies they will look at the way the market is shifting and help us with a point of view.

"So they will have a point of view on stuff like whether late stage is more interesting than early stage but fundamentally it’s a human business. We’re making the final decisions."

Larizadeh added: "Do we use data to inform our decisions, like in any other business? Yes we do. Do we have a little robot? No not yet."

Fortunately for Google Ventures, it has access to more data than any venture capital company on the planet. It knows exactly what people are searching for and how popular certain apps and websites are before it invests in them.

Tluszcz says that venture capital does rely on data, but there's also something else needed as well. "It’s also a lot about the feeling you have as individuals. I think they go past that. I think they just go f*** it, I just care about the data, data, data, data, data.

"At the end of the day, the gut feeling doesn’t exist any more at Google. What exists is show me the data, show me the traffic that these guys are having. That’s my view of them. Having said, no matter how, if I could get them in I would."