Good software engineers can easily find lucrative employment in Zurich. Why did you decide to found a startup instead?

I come from an entrepreneurial family. My grandparents and my parents were small business owners themselves. Funnily enough, when I told them about my decision to found Sphere, they asked: why would you do this to yourself? As a kid, I was always fascinated by technology and knew I wanted to build something. I devoured Sci-Fi literature. Bill Gates was a role model to me. When Microsoft first presented the concept of the HoloLens AR goggles, that was a defining moment for me. I knew that spatial computing would become a big topic one day. So, a bit later, in 2017, I started Sphere. We were lucky to be accepted into Airbus’ accelerator which helped us shape the company and get our software product on the market in 2020.

Sphere develops spatial computing software for companies. What exactly does that mean?

It simply means that our software allows you to place virtual content in the space surrounding you. We didn’t come up with this concept ourselves, of course. But we built an application that gives enterprise users the tools they need to make the most of spatial computing, no matter what kind of hardware they use. Sphere’s unique selling proposition is its multi-user functionality that allows several people to share the same experience and work together, no matter where they are. It turns spatial computing to spatial collaboration.

Can you give an example to illustrate this?

Of course. The carmaker Renault, for example, uses Sphere for collaborative car design. In the past, car design meant building a prototype, 3D-printing different parts and discussing results and improvements over the physical model. Sphere lets people work together in the same room regardless of the continent on which they live. They can put on a headset and see the exact 3D data from the company’s CAD software projected on the prototype. It is easier, faster and more affordable to change something digitally and discuss variants. This type of collaboration is relevant for many companies that build things.

A second good example is remote assistance. Here we could cite Airbus as a client. Remote assistance means that experts can guide someone else in a different city and see the same thing as the person they’re helping. Spatial computing opens up many new forms of working together, for example, experts can remotely interact within technicians’ physical environments and place virtual content into these environments. And wherever people work with machines, they’re better off having their hands free anyways.

Since February 2024, the Apple Vision Pro mixed reality headset has been available for purchase. What impact does this have on Sphere?

The impact has been huge because of the signaling effect. Since the announcement, the number of inbound leads from potential customers, partners and investors has risen substantially. Once again, Apple managed to create hype around a topic that makes everyone else take it seriously. Just look at the volumes. According to public estimates, Microsoft sold around 300’000 units of its HoloLens and HoloLens 2 MR headsets in 8 years. Apple sold 200’000 devices in the first 10 days alone.

For now, it still seems like people are trying to figure out what to do with that new pricey gadget.

It depends a lot on what your expectations are. If you think you can throw away your monitor and work with Excel sheets the whole day on your Vision Pro, this might not be the ideal use case. For us, this platform is an epiphany because it is so well done and polished. But since we’re not addressing consumers with our software but rather businesses, one thing hasn’t changed: our clients want to see demonstrable added value.

There are a lot of different hardware approaches. Could you elaborate on which are most relevant for enterprise customers?

Sphere is hardware-agnostic software which means it can be used with any kind of hardware. Some of our customers start with smartphones and tablets, which are ubiquitous. They are not hands-free, but useful nonetheless and important in this transition period. One step up is the robust low-tech assisted reality glasses used in the oil and gas industry. The next level is virtual reality headsets from different vendors such as Oculus, HTC and PlayStation. These are mainly used for simulations. When it comes to augmented reality, we must distinguish between see-through and pass-through. True AR with glasses you can see through, such as the Magic Leap 2 and Hololens 2, are very complex and need billions of dollars in investment. This is probably the reason why Microsoft doesn’t plan a Hololens 3. Pass-through, the new category, is much easier to construct. You just have cameras that film the environment, and this image gets displayed in front of your eyes. I think this category, including the Vision Pro and the Meta Quest 3, which is much cheaper, will become the standard over the next few years. Which model you choose as a company depends on several factors, if you’re employing knowledge workers or are an industrial company, for example. For us, it’s a big plus to be able to support any platform, since the hardware is developing very quickly.

How will AI contribute to the extended reality experience?

This will be a big topic going forward. We have watched the development of the space with a lot of interest and will be announcing an initiative with a US-based university and a large tech company in the coming months. We want to give AI eyes and ears.

Why and how?

If you think about it, our interactions with this ground-breaking AI technology are very similar to how people interacted with MS-DOS in the 1990s. People talk a lot about prompt engineering. For me, that sounds funny. XR [extended reality] and AI are a match made in heaven because XR can provide the spatial context to leverage the full potential of AI. Imagine looking at a machine your company is building through a headset and simply asking your AI what is wrong with it. This sort of anomaly analysis, of course, necessitates training the AI model on the data and manuals that are proprietary to your company – which is a completely different thing than training an AI model on internet data. You need to make sure this data doesn’t leave your company, and this is what our partnership will be about. Speaking to AI is much more natural than inputting data through your mouse and keyboard, which doesn’t really work in the XR space. Whereas being in XR allows AI to see, therefore solving the context problem that AI has when it’s used on a desktop.

That sounds a bit like AI could become a tutor thanks to XR.

We already have use cases in education today! Sphere’s software helps teach technicians who have to learn highly complex manual workflows. An instructor can guide several students through the relevant steps. If you would add AI to this, it could help supervise the class and correct wrong movements, and it would help a lot with remote training as well. People often speculate that AI will make this or that profession obsolete. The opposite is true. We have a pronounced shortage of skilled workers and skilled instructors in many professions, and AI married with XR could help train more of them to alleviate this problem. I live in the US, where distances between towns are huge. Remote training has incredible potential to upskill people no matter where they live. When I look at the developments in the hardware space and in AI, our timing to launch Sphere was quite lucky. We were early but not too early.