Illoca Raises $13M to Reinvent Architectural Design

According to VentureBurn and Bessemer Venture Partners, AI-native startup Illoca raised $13 million in seed funding led by Bessemer Venture Partners, with participation from AIX Ventures, Root Ventures, and Alt Ventures. The company, founded in 2024 by Chin-Yi Cheng and Chiaowei Yu, unveiled a generative design product called Tracing Paper™ that the founders and investors describe as an interface for converting sketches and bubble diagrams into editable 2D and 3D models and exporting construction-ready data into tools like Revit. VentureBurn reports the platform adds natural language controls and aims to reduce the time architects spend transcribing ideas into precise models. Bessemer characterizes the product as a step toward more generative schematic and detailed design workflows. Reporting also highlights the precision and code-compliance challenges that differentiate architectural use cases from purely visual generative tools.
What happened
According to VentureBurn and a Bessemer Venture Partners announcement, Illoca raised $13 million in a seed round led by Bessemer Venture Partners, with participation from AIX Ventures, Root Ventures, and Alt Ventures. The company was founded in 2024 by Chin-Yi Cheng and Chiaowei Yu, per VentureBurn, and announced a first product called Tracing Paper™. Bessemer's post describes Tracing Paper as a generative interface that accepts hand sketches and bubble diagrams and produces editable 2D or 3D models which can be exported into industry tools such as Revit.
Technical details
Per VentureBurn and Bessemer, Tracing Paper supports three headline capabilities: sketch-to-3D conversion, natural language interaction to modify layouts, and export of constructible, code-ready model data for downstream CAD/BIM workflows. VentureBurn reports the founders combine prior experience in generative AI research and BIM practice, noting Chin-Yi Cheng's background at Google DeepMind and the Autodesk AI Lab and Chiaowei Yu's role as a BIM lead at Tesla. Bessemer frames the product as addressing both generative creativity and the precision requirements of building modeling.
Industry context
Editorial analysis: Architects and engineers commonly use legacy CAD and BIM tools that prioritize millimeter accuracy and regulatory compliance. Industry reporting from VentureBurn and Bessemer highlights a recurring tension: visual generative tools produce compelling imagery but lack the groundedness required for structural calculations, code compliance, and constructible deliverables. Companies attempting to bring generative AI into design workflows therefore face the dual challenge of usability for ideation and strict data fidelity for construction.
Why this matters
Editorial analysis: For practitioners, a reliable sketch-to-BIM workflow could materially change early-stage iteration velocity. If the product capabilities reported by VentureBurn and Bessemer are realized in production quality, architects and BIM teams could shift time spent on transcription and manual modeling toward exploration and coordination. That said, the core engineering and regulatory interfaces for architecture impose higher validation costs than creative imaging applications.
What to watch
What to watch
Observers should follow three indicators in the coming months:
- •evidence of robust Revit and BIM interoperability in customer pilots
- •case studies showing compliance and constructability checks integrated into generated models
- •enterprise adoption signals from architecture firms or construction engineering teams. Reporting to date is promotional and investor-facing; independent trials or third-party validations will clarify practical limits in complex projects
Source attributions and notable claims
VentureBurn provides the funding details, founder biographies, and feature claims about Tracing Paper. Bessemer's blog post frames the product narrative and user experience. Yahoo Finance and other syndications report similar funding and product descriptions. VentureBurn includes the investor-quoted line, "Any architect can generate code-compliant designs in just the time it takes to open a file," attributed to Bessemer Venture Partners.
Limitations of current reporting
Editorial analysis: Existing coverage is largely company and investor material. Independent performance benchmarks, published interoperability specifications, and published customer pilots are not present in the scraped sources. That gap leaves open questions about model robustness on regulatory and structural validation tasks.
Scoring Rationale
This is a notable seed round that highlights venture interest in applying generative AI to professional CAD/BIM workflows. The story is relevant to practitioners because product claims target time-consuming parts of architectural workflows, but independent validation and enterprise adoption remain unproven.
Practice with real Ad Tech data
90 SQL & Python problems · 15 industry datasets
250 free problems · No credit card
See all Ad Tech problems


