Father of five, artist, economist, self-thought programmer, and as founder of AImotive, László is the most succesful Hungarian startup founder to date. He believes that the region would be better of without subtitles, obsessed with graphic and data, both of which accompanied him throughout his journey.
Since late last year you might have read few articles about AImotive's and László's story, but none was as thorough and candid as the recent interview done by György Jaksity and Áron Vidovszky from Concorde Group.
László Kishonti and Csaba Kákosy gave a full account of from which you can gain invaluable insights. Not only the story of AImotive, but how he started out as an entrepreneur, how did the US investors missed the deal early on and you can also have a glimpse on how you should think about building your product, wishing to replicate AImotive's success.
Disclaimer: The article below is not a transcript but a summary of the conversation.If you speak Hungarian we'd recommend to listen to the whole episode on Spotify or Youtube.
From Dunakorzó to the Tech Industry
László’s entrepreneurial journey began when he was just 14, sketching portraits at Dunakorzó. The competition was so intense that he was even barred from showcasing his work in the bustling Vörösmarty square, where most of the tourist traffic took place.
Economist by trade László built quite a career early on. He developed and named the first government bond index called MAX Index. As another first, he managed the first hedge fund and index tracking fund in Hungary in the late 90's.
In the mid 2000's at the age of 26 as a Chief Investment Officer at a large bank he thought that nothing exciting is going to happen in banking in the next 20 years so he decided to shift his focus on the thing that got him excited the most. Programmable mobile phones. In 2005. Bear in mind that the iPhone debuted in 2006.
He founded his first company, which was dedicated to the monitoring of Graphics Processing Units. The company had grown decently for 10 years, catering to industry giants like Nokia, Nvidia, and Tesla.
Adas Works and AImotive
As the clientele and László's focus shifted to the automotive industry they spun off some of their developments and 15 people into a new company called Adas Works in 2015. A year later their merged with the initial company forming the now well-known AImotive.
We hope László will soon show his initial pitch deck to the public because it resembles nothing we have seen before or even since. It was a hand-drawn comic book, showing that his early years as a Street Artists were well spent.
Lot of Silicon Valley investors called him right away, as they haven't seen anything like this before either.
Luckily for us, they had back then the so called biking distance policy (only invest into a company which they can visit via bike), and László didn't wanted to move to the US as he understood early on that he could have an edge building a deeptech company in the CEE region.
At Day One we understood László's well-articulated vision of the future. Namely that in the next decades the automotive industry is not going to be about the hardware anymore, the brakes, clutches and engines, but rather built around the software. This may sound trivial today, but it didn't in 2015.
Also having known about his previous success with his GPU monitoring company and the continuous interest showed by industry players like Bosch and Nvidia, we knew we are on the right track in terms of founder-market-fit.
DayOne's thesis changed very little since its founding in 2013. We want to back young entrepreneurs early on who can leverage the relatively cheap but skillful technical talent pool of the CEE region to scale a company for a third or even quarter of the cost of a Western European or US competitor.
The success story of AImotive reassured us that our thesis is right. Even though there are many companies valued above AImotives exit price, only a handful of startups were able to realize the valuation that their late-stage investors put on them. As the largest ever startup exit in Hungary it is an even bigger boost for the Hungarian ecosystem.
Building the greatest Hungarian startup
The industry's focus is shifting back from full self-driving (level 5) to lower levels, where the AI could be put to use for safety and comfort features. AImotive has built a very strong foundation to be a leading player in this field.
In the past seven years they've built a portfolio of four products: Self Driving, Simulation, Data and Chip development. Though they could be used as standalone solutions,most seem to be almost inseparable from one another.
Chip design might be an exception,which started out as László's hobby, but it might make the most sense know, having become part of the world’s third largest manufacturer.
László guided us through his reasoning, which is no doubt shared by Stellantis.
"The automotive industry, as always, is aiming for the cheapest and most efficient solutions on the market. The GPUs made by Nvidia were designed for the gaming industry, thus they will hardly be sufficient for self-driving cars."
To illustrate the issue, he told us that the currently available solutions could cost up to 700 000 USD, while being less efficient than the 700 USD smartphones in our pocket.
Simulation technology has been a game-changer in the automotive industry, especially for autonomous vehicles. It allows us to test and perfect driving algorithms without the cost, time and potential risk associated with real-world testing. AImotive has taken this concept and elevated it to a new level of affordability and efficiency.
Again, László shared the magnitude of this development with us:
"Traditionally, using human drivers for testing costs around $2 per kilometer. However, with AImotive's advanced simulation technology, this cost plummets to a mere $0.02 per kilometer. The true beauty of this technology lies in its ability to focus on significant events, considering that the majority of driving consists of uneventful periods like waiting in traffic or driving straight without disturbances. Thus, the cost-effectiveness of AImotive's simulation isn't merely a hundredfold, but probably ten thousand times more economical."
Data is the cornerstone of any technological innovation. In AImotive's case, the role of data has been pivotal, yet it only recently transitioned into a standalone product.
Large corporations like Nvidia and Bosch have been investing hundreds of millions of dollars each year intohuman resources to train their models. AImotive, however, has taken a different approach.
Being able to utilize just a third or a fifth of the funding compared to their US competitors, AImotive always had to be mindful of the effectiveness of their approaches.They have developed a data product that doesn't necessitate real-time processing. Instead, it facilitates the automation of processing two million frames in two hours, all without the need for human workforce. This innovation results in substantial cost and time savings, as this would cost significantly more even in developing countries, not to mention that it wouldn’t be done in two hours, but two or three weeks.
How much is AImotive worth for Stellantis? Well, as the transaction closed Carlos Tavares,Stellantis's CEO told one thing to every executive he met in Budapest:
"Don't let us to kill you!"
Surely, everything makes sense in hindsight, but it’s never easy to scale a team from 15 to 200 people, nor tomanage investors in your board who might have different incentives or time horizons.
As a parting advice from the most successful Hungarian startup founder:
"...although every decision was consensus based, there are a lot of stomach punches under water in scenarios like this and the investors you choose for your journey might be a lot more important than how much money you raised or at what valuation."