Waldek Technologies Convinces the Smartphone to Render the Digital Twin

A Techstars-backed bet on browser-native Gaussian splatting is quietly testing with Tesla and Volkswagen.

About Waldek Technologies

Published

You upload a video, shot on a phone, of a half-finished construction site. In a browser tab, a cloud of shimmering, colored points resolves into a three-dimensional room. You click on a pipe, highlight it, and drag it a few virtual feet to the left. This is the promise of Gauzilla Pro: that the most complex spatial data can feel as malleable as a Google Doc, and that it can all happen without leaving Chrome [LinkedIn / Techstars community, Jan 2025].

For Yoshiharu “Yoshi” Sato, the solo founder behind Tallinn-based Waldek Technologies, this moment is the wedge. His company is building an AI-powered, no-code web editor for 3D and 4D digital twins. The core technical bet is on 3D Gaussian Splatting (3DGS), a rendering technique that turns video into interactive, photorealistic point clouds. Waldek’s twist is running it all client-side, in the browser, using WebGL and Rust-based rendering so that sensitive data never needs to leave a user’s machine [radiancefields.com].

The Wedge of the Web Browser

The ambition is to collapse a multi-step, expert-driven process into a single, accessible action. Waldek is aiming for the space between consumer photogrammetry apps and industrial platforms like Matterport, targeting AEC, e-commerce, agriculture, gaming, and AI/robotics [LinkedIn / Techstars community, Jan 2025].

By keeping computation on the edge, it addresses latency and privacy concerns for enterprise clients. By building a no-code editing layer on top of the 3DGS renderer, it promises to let a project manager, not just a CAD technician, segment objects, add annotations, or manipulate scene geometry.

Traction and the Tallinn Proof Point

Validation has come in two forms. The first is programmatic: Waldek was part of the Techstars Berlin accelerator class of 2024, from which it raised a $120,000 seed round [CB Insights, September 2024]. The second is more tantalizing: the company reports having conducted proofs of concept for Tesla, Volkswagen, and CoStar Group [Techstars Job Board].

Competitor Primary Focus Capture Method Key Differentiator
Luma AI 3D asset creation Smartphone video AI-generated neural radiance fields
Polycam 3D scanning LiDAR / smartphone High-quality scans
Scaniverse 3D scanning LiDAR / smartphone Precision measurement tools
Matterport Professional digital twins 360° cameras Industry-standard platform
Gauzilla Pro No-code 4D editing Smartphone / drone video Browser-native 3DGS

The Single-Founder Counterfactual

The risks are as sharp as the ambition. The company is betting its future on a rapidly evolving technical frontier. The most visible question mark is organizational: Waldek is a solo-founder venture at a seed stage where building both cutting-edge R&D and a sales motion for complex enterprise PoCs is a monumental dual track.

Yet, the bet makes a distinct kind of sense. The world is generating more spatial video data than ever. If Waldek can prove that its browser-based editor is a reliable tool that saves time and secures data, it could carve out a defensible niche.

The Next Twelve Months

The immediate roadmap is clear: move from closed beta to a launched product, convert proof-of-concept interest into first paid pilots, and raise a larger round to build out the team. Success won’t be measured in waitlist numbers, but in whether a construction foreman or a factory planner starts their day not in a specialized program, but in a browser tab, editing reality as easily as they edit a spreadsheet.

Read on Startuply.vc