Data Science Investment Counter —Funding raised by UK data science companies in 2018.
£ 5.640 Million

“AI Companies, Be Vertical” Draper Esprit’s Take on London Startup Scene

Draper Esprit is one of Europe’s most dynamic Venture Capital firms.

Founded in 2006 and a member of the global Draper VC network, the company has so far invested over $1 billion in 200 startups—including nibbling breakout company Graze, IoT  firm Movidius, and online fashion portal Lyst— and it pegs the overall value of its exits at $4 billion. In June, Draper Esprit even took the peculiar step of listing on London Stock Exchange’s AIM market and on the Irish Stock Exchange, with the stated aim of becoming a $1 billion-worth source of “patient capital” for Western European startups.

The London-based fund has  shown a keen interest in machine learning, data science and Artificial Intelligence,  notably partaking in a $30 million investment into Graphcore—a company manufacturing chips that could dramatically speed up machine learning applications —in October.

But what sort of AI startup is most likely to catch Draper Esprit’s eye? Is there a formula the firm turns to when it comes to spotting budding companies? To discover this, we sat down with Draper Esprit’s Principal Vinoth Jayakumar.

Hello Vinoth, great to talk to you. Just to start, give us a quick run-down about Draper Esprit— and yourself. 

So, as a principal , my primary role is to find and originate deals, and one of the areas I am most looking into at the moment is Artificial Intelligence. I am not an academic— my job is to look at the commercial aspect of AI technology. Historically, Draper Esprit has done deals in certain specific sectors, which include consumer e-commerce, enterprise software, and hardware (which can sometimes be relevant to AI.) When it comes to AI, our view is that we are looking for AI companies that are doing something in a vertical rather than in a horizontal way.

What does it mean, precisely?
It means is that we are unlikely to back a company that is creating an AI tech—or a new machine learning mechanism— that is applicable to any kind of industry. We are unlikely to back a company that is working on a technology which it wants to apply to a lot of things— from cars, to language translation, to image recognition. That’s a horizontal view of AI.
We’re more into companies that develop AI software for  very specific purposes. One of the reasons for that is that we have to think exactly  how the company is going to make money, how it is going to monetise this tech.
For instance, Graphcore: even if it is a semiconductors manufacturer, its chips are designed with the aim of lowering the cost of accelerating AI applications in the cloud. That’s a very specific vertical.
In London, though, we  increasingly see more companies that are very horizontal, that develop technology that is applicable to anything. That is actually good, but it’s hard to see a clear path to the money from there.


Draper Esprit’s Vinoth Jayakumar likes AI Companies— provided they are vertical.

This is interesting. What do you make of London’s data science and AI scene, more in general?
London is an AI hotspot, because most of the AI tech has come out of Cambridge—and, more rarely, from Oxford— and London is not far from Cambridge. Often the technology gets developed at the university and then it becomes intellectual property, and then when they want to monetise that intellectual property they come to London, which is where the money is. In fact, there’s also a well-developed Cambridge Angel Network that makes a lot of AI deals, but London is the go-to place.
In a strange way, when I think about London, my general feeling is that AI companies that are based in London were not  originally designed to be AI companies.

Is that a bad thing?
Absolutely not. What I meant is that they were designed to solve specific problems—word recognition, image recognition, and so on— and then they developed a solution that was based on AI. And that’s a big plus for a startup, because it shows it can think ahead of the curve. For instance, I know of a lot of companies in FinTech, and a huge problems they have is preventing fraud. And even if they are not purely AI companies, more and more they are relying on AI or machine learning solutions.  This is great because it means they are going for  the best way to do it—which today is probably AI.

One of the things that come to mind when one thinks about British AI companies is how they often end up being bought by US giants—think of Google’s DeepMind or Apple’s VocalIQ. How can London, and the UK, retain their best AI startups?
The companies that we are talking about have been acquired by Google or Apple, but they’re still pretty much British companies. What this reflects is that US cash is moving to this part of the Atlantic: it means that there are more people bringing money into the UK.  I don’t see it as a problem. If anything, what we’d like to see more and more is British AI companies getting more access to US research and researchers. Cambridge is one of the leaders in AI, but it’d be good if UK companies could tap into ideas from  MIT, Harvard, and other top US universities.

If that is not a problem, what do you  see as a problem, instead? Is there any challenge, right now, to London’s AI startups?
I think something that is in need of a solution is horizontality. AI companies that are horizontal need to figure out more quickly what verticals they could be doing. They need to have a commercial team that looks at how to concretely apply their technology, and the faster they do it, the better.

Featured Image via


Co-working space and blog dedicated to all things data science.

Subscribe to our newsletter