Henrik Bennetsen and the distance between an idea and a business

A great deal of Henrik Bennetsen’s working life can be understood through a question that sounds almost impolite in a room full of inventors: how do you make this into a business? At Otherlab, where engineers worked on new technologies, that was part of his job. It is a question with an excellent talent for spoiling the afternoon. The machine may be fascinating. Someone still has to want it.

Bennetsen went on to help found Canvas, a construction robotics company, and Savor, which makes fats without conventional agriculture. Today he is chief business officer at Formant, working in the world of physical AI. Those industries offer plenty of opportunities for a dazzling demonstration. They also require the less photogenic work of understanding customers, organizing a team and figuring out what happens after the demonstration ends.

His route to that work began somewhere else entirely. Before the startups, there was music. Before the building sites and food chemistry, there were experiments with art and virtual worlds. The career looks unusually restless until you consider how often he has worked on the same practical passage: from something people can imagine to something they can actually use.

A musician enters the virtual world

Born and raised in Denmark, Bennetsen moved to San Francisco in 2005. He had worked as a professional musician before returning to academic life. By 2009, he was associate director of Stanford Humanities Lab, interested in what virtual spaces could do for artistic expression. His work included LifeSquared, an exploration of an immersive archive of artist Lynn Hershman Leeson’s work inside Second Life.

Consider the challenge. An archive normally asks you to look, read and follow an arrangement someone else has prepared. A virtual environment opens up questions about movement, presence and participation. Where do visitors go? What do they encounter? How do they relate to the material and to one another? Putting an archive into three dimensions makes the interface part of the artistic problem.

That period also included Speed Limits, a project involving Denmark’s Bornholms Kunstmuseum. The attraction of a virtual space was tied to expression and collaboration. In a 2010 speaking biography, his continuing interest in creative expression through technology appeared alongside his earlier life as a musician. The arts remained part of the brief even as the tools changed.

The browser had to open the door

In early 2010, Bennetsen co-founded Katalabs as a spinoff from Stanford research. Its OurBricks project connected people looking for 3D content with people who could make it. A digital object needs a route from its creator to whoever wants to use it. That route can be surprisingly tedious. Formats disagree. Software makes demands. A promising collaboration gets lost in the plumbing.

Katalabs worked on that plumbing through open web technologies. KataSpace enabled browser-based virtual worlds; OurBricks let users upload, share and perform simple edits on 3D models. The point of running in a browser was access. An extra plugin might seem like a small request to a developer. To a visitor, it can be an invitation to leave.

“Everything that has made the Web great has been open.”

Henrik Bennetsen, 2011

He also helped gather the people building this corner of the web. In December 2011, WebGL Camp 4 at Mozilla brought together more than 70 technologists and innovators. Earlier that year, he had presented OurBricks at WebGL Camp 3. The programme put him among browser developers, graphics specialists and people exploring how three-dimensional work might become ordinary online activity.

There is a social component to a technical standard. People have to try it, exchange results and find reasons to keep working on it. A conference room can be part of the infrastructure. Bennetsen’s role in organizing one belongs beside the software work because an emerging technology needs a community capable of doing more than admiring it.

Then came an actual wall

At Otherlab, Bennetsen encountered Kevin Albert and Maria Telleria. Their work became Canvas, and he joined them as co-founder and chief business officer in September 2016. The challenge shifted from shared digital space to a construction site. Here the environment would not politely remain where a designer put it. A machine had to fit into an established trade and help people complete a real job.

A drywall finisher gives you a rather concrete definition of usefulness. The wall must be finished. The crew has a process. The equipment has to earn its space in that process. Calling something a robot does little to answer the immediate questions about how it will work, who will operate it and what improvement the customer can expect.

The business role in such a company sits close to the engineering. Customer conversations influence what gets built. A technical choice changes the commercial proposition. The people paying for a machine are buying an outcome inside their own operations, with all the practical constraints that implies. Bennetsen’s work was part of helping the invention reach that setting.

His name also appears among the inventors on an automated wall finishing system and method patent granted in September 2022. It is a useful detail in a career often described through executive titles. The formal record of the work includes a technical invention, shared with his Canvas colleagues.

In January 2026, JLG announced its acquisition of Canvas’s core technology. The relationship had begun six years earlier, when Canvas selected a JLG platform for its robot. That history makes the transaction especially relevant to the original problem: a new technology finding a home among equipment already used on building sites. The announcement also welcomed Canvas team members into JLG.

The climate ambition meets the dinner plate

In 2021, Bennetsen was considering his next chapter with climate impact in mind. Conversations with Ian McKay and Kathleen Alexander introduced him to a way of making fats through chemistry. The three became Savor’s co-founders. The proposition took him into another essential part of everyday life: the ingredients people eat.

Fats sit inside recipes and manufacturing processes, where their behaviour matters as much as the story told about them. A new ingredient has to perform. A baker’s interest in a lower environmental footprint still leaves questions about the dough. For a business founder, this is a familiar sort of negotiation between an ambitious idea and a user’s existing expectations.

Bennetsen’s account of joining Savor gave taste a place in the ambition. Eating food prepared with the company’s product was a meaningful milestone for him. The climate argument had brought him to the project; the experience of the food helped make the proposition tangible. There is a considerable difference between understanding why an ingredient ought to exist and wanting another bite.

He left Savor in May 2024. His description of the professional handover emphasized preparing the company, giving the leadership group time to take responsibility and talking individually with team members. A founder’s departure changes the organization as well as the person’s calendar. The transition deserves a place in the career story alongside the founding.

It also prevents the Savor title from swallowing the rest of his life. He remains a co-founder of the company, while his current executive work is at Formant. There are several chapters here, and the business-card version has trouble holding them all.

The machines are talking. What happens next?

At Formant, the problem is again grounded in use. The company works on software for physical automation, connecting devices, data and people. Bennetsen’s recent writing focuses on what hardware teams should do with the information their fleets produce. Collecting information can leave a team with plenty to inspect and too little capacity to act.

That distinction matters in ordinary work. An alert creates an obligation. Someone must interpret it, decide whether it matters and follow the problem through. As fleets grow, a few experienced people can become the point through which everything passes. A dashboard may make the queue visible without making the queue shorter.

In his public discussion of product development, Bennetsen has also argued that clearer intent becomes more valuable as AI tools make building software more accessible. Knowing what a customer means is part of the work. A conversation can supply the detail a specification misses. The attraction is recognizable from the earlier ventures: shorten the distance between the people experiencing a problem and the people making its solution.

He has acknowledged uncertainty about how companies should operate as AI changes their working practices. Experience has supplied questions without giving him a finished rulebook. That is a more useful stance for this career than pretending the browser, the bakery and the building site all follow the same instructions.

A company is also a group of people

Outside the commercial ventures, Bennetsen served on the board of the Scandinavian School in San Francisco and chaired it from 2012 to 2015. In 2013, he promoted its spring fundraiser through a community mailing list. The appeal was straightforward: help spread the word, suggest channels and bring people together around a shared cultural institution.

His irritation with context-free LinkedIn connection requests belongs to the same smaller scale of human interaction. He has asked people to say why they are getting in touch. A couple of relevant sentences can make an introduction easier for the recipient. It is a modest request, and apparently one the professional internet finds surprisingly strenuous.

The guitar has survived in his public biography, too, as something he hopes to make more time for. After so many changes of industry, it is a pleasingly familiar item to find still on the list. A career can acquire new machines without requiring the disposal of every earlier interest.

Follow Bennetsen through these projects and the recurring work becomes visible: bring people into a room, understand what they need, and help a technical possibility find its practical form. The materials have included digital models, walls and fats. Each has supplied its own resistance. The business begins when someone else can put the idea to work.