AI STORIES – Ep. 1 #Belgium


Can a train drive autonomously?


Interview with Sam de Smet, Co-founder of OTIV.AI


Hi Sam, we are very happy to introduce you to the France Digitale community! Can you tell us something about yourself and OTIV.AI?

I’m Sam de Smet, co-founder of OTIV.AI, a young AI startup based in Ghent, Belgium. I am a former business consultant, while my co-founder is an automation engineer. OTIV’s mission is to teach rail vehicles to drive autonomously in complex environments. Concretely, we focus on trams in urban spaces and locomotives in industrial environments, like shunting yards and ports. Those settings, while radically different, have one thing in common: complexity. Unlike metros and high-speed trains, these rail vehicles interact with other actors, like bikes, cars, pedestrians, trucks, machinery and other trains. We start by assisting the driver in avoiding collisions, preventing overspeed, etc., then we go fully autonomous by removing the driver altogether- we delegate the control of the vehicle to algorithms.

How did you come to such a specific niche?

Our conviction is that in all the autonomous driving hype, rail vehicles have been overlooked, despite being the most appropriate use case. We’ve seen a lot of money going into self-driving cars but we see little added value in them- if you are in a car, even if autonomous, you are still in a traffic jam. Trains and trams are for us the right mobility solution to avoid congestion. We’ve also seen millions go into hyperloops, but with the rail infrastructure already in place, we don’t see the point in building them. We have trains, we have trams: all we need is making them better. And that’s where we come in. 

Do you have any use cases in place? 

We did one pilot with the Dutch national railway company NS on a shunting yard and we are currently carrying out pilots with industrial players. We chose to focus on the latter as their willingness to embrace automation is much higher than in the public sector. 80% of the work in industrial environments is already automated, so why don’t add self-driving trains.

What has been your fundraising strategy so far? 

We funded the company with a small pre-seed round and closed our seed round a few months ago. The latter amounted to 1 million EUR with a mix of grants, loans and equity from business angels. It wasn’t easy to find the right investors- as we are developing a truly deeptech product, not a SaaS platform that you can set up over a weekend. But we eventually came across two angels who believed in our story and introduced us to their angel fund.

Do you see any barriers to the growth of Otiv?

The main challenge is that we are active in a very slow and conservative industry. For the last few decades, the railway sector has been dominated by a small number of large players, which means that we don’t have many potential customers. Take Germany: there is only Deutsche Bahn. That’s why we turned to industrial clients. Automation is for them an investment decision guided by their will to cut costs and improve efficiency. Our first fully-automated pilot will also be in an industrial setting.

In your opinion, what should the EU do to make the railway industry more innovative?

I think there are three key actions the EU should take. First, liberalise the  industry. Railways are a natural monopoly and despite all the discussion about increasing competition, very little has changed. This is true not only for passengers but also for freight transport. 

Second, the European Commission should continue to look out for market unbalances, with dominant players growing increasingly larger, attracting public innovation subsidies and ultimately stifling innovation by young companies. Even today, many startups are more innovative than these established players, yet they struggle to secure public funding. 

This leads me to my third and last point: the EU should establish a preference for startups in public funding and harmonize existing schemes, while also increasing their visibility. There’s a billion European funding programs out there, but some are just hidden, so all startups end up applying for the same schemes while plenty of money is available in other facilities that simply go unnoticed. 

What are your plans for the future? 

Our industrial use case is global, whereas our public transport use case is very European. For our next round, we will certainly look for investors beyond Belgium, notably in the UK, Germany and France. 

Thank you for your time Sam, we look forward to seeing you at the France is AI conference in Paris on November 8th!