Taking deeptech from innovation to industrialisation requires a healthy appetite for risk — meet Tristan Rousselle, CEO of Aryballe Technologies.
How Aryballe translated its cutting edge research into a scalable solution.
Full disclosure: Aryballe is a Hardware Club portfolio company.
Over a century ago Alexander Graham Bell declared “If you are ambitious to find a new science, measure a smell”. Molecular scientist-turned Aryballe Technologies cofounder Tristan Rousselle has taken the challenge further by developing a universal electronic nose. The idea had remained in the realm of science fiction - until now - since the scale of the problem is vast: humans can detect up to 10,000 distinct odors but the possible molecular combinations that form smells number in the billions.
A previous wave of startups in the 90’s had succeeded in producing limited, clunky devices with a price tag of upwards of $100,000. But no single solution offered a comprehensive catalogue of smells. Rather, they were trained to only identify key compounds, giving a partial nasal snapshot at best. Development of gas sensors led the way in terms of miniaturisation — yet remained specialised resulting in a one-sensor-per-gas model.
Aryballe’s eclectic group of scientists and investors aimed to leverage the latest breakthroughs from the intersection of nano, bio, cognitive sciences and software to create a portable electronic nose. Designed as a portable stand-alone device, the first application of the technology is planned for the fields of flavour and fragrance, food and industrial safety. The database of smells their device can recognise is exploding exponentially.
“True olfaction requires a sensor to be able to record a broad range of different molecules in a family — they may have some homogeneity but represent hundreds of thousands of different molecules. The possibilities are endless” Rousselle says.
Success in deeptech depends on how you view risk
“It’s our job as early developers to be able to evaluate the risks” he says. After some initial market research, it was the positive reaction from the flavour and fragrance industry that convinced him. This was encouraging as they had already trialled most existing odour analysis solutions.
When speaking to industry experts Rousselle met both detractors and those who never thought they would see the solution offered in their lifetimes. “As an entrepreneur, you always try to take good lessons from the naysayers,” he says “but we were encouraged by others also”. Aryballe even secured some early partnerships which granted them some vital early traction. “Plan A worked so we haven’t had to resort to Plan B,” he says wryly.
“It’s our job as early developers to be able to evaluate the risks,”
Deeptech startups like Aryballe are extremely challenged from three perspectives:
- Business model: they must define new benchmarks.
- Priorities: they are unable to tackle every vertical at once.
- Marketing: they have no pre-existing customers.
Many technologies fail to successfully make the leap from a research project to commercialisation. The time to market is both long and uncertain and cutting edge research requires years of testing and clinical trials before producing a viable technology. Deeptech is often geographically constrained: it requires access to talent with training, expensive instruments and advanced infrastructure. The scaling process then requires deeptech startups to have the ability to link innovative technology with existing manufacturing processes and components. This produces a general haze of risk in the eyes of investors and other stakeholders.
Having skin in the game
Rousselle co-founded Aryballe Technologies two years after the acquisition of his previous biotech company PX’Therapeutics. For him, the appeal of founding Aryballe lay in the challenge: “I wanted to work on a breakthrough technology” he says “where the innovation came from the combination of different disciplines”. Headquartered in the CEA-MINATEC campus in the French city of Grenoble, the key to Aryballe’s technology lies in Surface Plasmon Resonance: a combination of optics and biochemistry. This biosensor had been very well developed in liquid analysis but had never been tested for gas. “After selling my first company I met [Aryballe Scientific Partner] Thierry Livache, who told me ‘You’ve come just in time. We just wrote a patent covering gas’”. After some initial tests, they felt confident in commercialising the technology.
The initial obstacle was one familiar to many deeptech startups: financing. Rousselle threw everything he had into the idea. “I spent all my money” he reveals “roughly one million euros”. In terms of mindset, Rousselle is clear that the reward outweighs the risks. “Investing in R&D with your personal money is not easy” he laughs “you have to negotiate without knowing what the final result will be”.
“I put myself in a position where I need to succeed. That’s why I invest my own money”.
It’s interesting to examine the role of the entrepreneur in connecting the wide range of stakeholders that enable deeptech innovation. Rousselle describes his job explicitly as an ‘early developer’ of technology. Although they lack important things that bigger companies possess, such as capital and market access, entrepreneurs have advantages in identifying innovation: “When you have a small structure and team you have the liberty to meet the right people” he says. There is a cultural and organisational drive towards innovation within startups that bigger companies lack. “There is focus because we are only working on one technology”.
A different mindset is at play individually: “It’s easier for an entrepreneur than an employee to be hungry,” he continues, “I put myself in a position where I need to succeed. That’s why I invest my own money”.
Making the leap from early stage to mid-stage startup
Aryballe has identified 50+ business potential applications beyond food and flavour, such as industrial safety and disease identification. However, it took almost two years to progress this far. Investors were well aware of the failure of previous generations of electronic noses. “Deeptech is not easy to explain” Rousselle says, “it’s more a question of trust than a question of understanding”. A lot of their investors’ confidence stemmed from the interest expressed by potential customers.
Aryballe successfully navigated its initial proof of concept and prototyping and raised a $3.5M series A in 2016. “It wasn’t a classic VC round,” Rousselle says. The synergy between theirs and investors’ aims was the key concern. “They are either well connected to the industry or the business angel community”. It pays to differentiate between partners who fund R&D and classic investors. As Rousselle notes “We need to maintain the core technology and make sure that nothing should prevent us from industrialising it”.
Turning innovation into industrialisation
“When you’re inventing a technology there is no reference. Our competition is the human nose!” he laughs. “Industrialising without specification is a nightmare”. In order to convince potential partners Aryballe had to develop an entirely new way to specify what their product did — almost like creating a new language for partners to learn.
“When you’re inventing a technology there is no reference. Our competition is the human nose!”
Caution is important when speaking to investors about ‘sci-fi’ applications:
- It’s all about avoiding overpromising and underdelivering. Establishing clear lines of communication between your investors and team can give you leeway to adapt to new circumstances. Aryballe pivoted away from launching a medical application towards gearing their first device for industrial safety. This cost them a six-month delay in their product road roadmap — but the return on investment promised to be worth 20 times as much as the initial application.
- Lead with the vertical that will deliver the highest return on investment. This requires you to be ready to adapt your strategy and be open-minded about the possible use-cases that present themselves. “We were cautious in spending and good at securing complementary financing,” he concludes. “It’s not done yet but it’s quite encouraging. The first phase will only be a success when we sell”.
- Hiring is a process of convergence and learning. “It’s a matter of recruiting people with different skills to adapt to our tech. A scientist will progress towards industrialisation” he says, “and an engineer will begin to grasp the fundamentals of the technology”.
