EUVC

The episode delivers solid technical and market insights - the shift from horizontal to verticalized AI, the bottleneck of quality 3D data vs. model parameters, and the 5-minute scanning vs. 2-day manual asset creation - but much runtime is consumed by founder narrative (South Tyrol origins, international expansion philosophy), culture debates, and soft positioning. The core insight density is strong when focused on the problem and solution, but diluted by extended sections on culture, hiring, and strategy that lack novelty for experienced operators.

“The core bottleneck today is not more parameters, it's better reality grade data.”

“You use some of the larger models for coding, you need other ones more for the physical world. So that's a huge trend where one size fits all doesn't work.”

The framing of 3D data as the native language for physical AI and the bits-to-atoms convergence thesis is well-articulated but not particularly novel - it's a logical extension of existing AI infrastructure thinking. The hardware-as-defensibility angle and the 'shovels in a gold rush' analogy are familiar venture framings. The specific execution (All Sides' scanner approach) is novel, but the conceptual landscape relies heavily on established AI infrastructure tropes.

“converging from the bits level to the atom level, which is basically enabling AI to really do stuff in the physical world”

“if there's a gold rush, you should invest into the shovels”

Franz Tschimben is a founder building at scale with real customer traction (Meta, Amazon, Nike, Adidas, Nvidia partnerships) and deep technical expertise in 3D scanning and AI. Michael Brehm and Ben Scheidt are practicing venture investors with demonstrated pattern recognition in AI infrastructure (prior investments in Xayn, AMI). All three are operators with skin in the game, though the investors speak more to thesis than day-to-day execution challenges.

“we work with customers such as Meta, Amazon, Nike, Adidas for various applications”

“Michael, you mentioned world models which falls under this category... Michael and Ben were introducing at the beginning”

Good specificity on the technical solution (2x2x2m scanner, 5 minutes per object vs. 2 days, sub-millimeter geometry, PBR + mesh output) and real customer names (Meta, Amazon, Nike, Adidas, Zara, Nvidia). However, the episode lacks concrete data on TAM sizing (only 'Decacorn' hype), customer deal sizes, production volumes, retention metrics, or hard numbers on quality differentials. The market sizing is asserted rather than evidenced.

“various products of that. And you can imagine that to be a machine of 2 by 2 by 2 meters, lots of lights, lots of cameras, lots of other sensors that in about five minutes capture all the data you need”

“we have a team of um, close to 45 people right now from 16 different nationalities”

The host (Speaker D) asks solid setup questions that elicit founder framing, but rarely pushes back or forces specificity. The 'temporary problem' question about whether 3D mapping will eventually saturate is the most incisive moment, but Franz's answer about 'billions of new items each year' sidesteps the deeper concern. The investors add perspective but aren't challenged on valuation claims or TAM assumptions. Missing are follow-ups on customer churn claims, pricing power, or competitive risk from open-source 3D reconstruction.

“Is this not a temporary problem? As in we have LLMs when we needed to build LLMs... isn't it like that, that once you've taken your big machines, you've kind of defined what a shoe look like, what a hammer looks like, and then in two years you're done”

“I would love to ask both of you from investing in France on all sides, what has this taught you?”