A Small Round With a Repeat Investor Base
Visoid, an Oslo-based startup that turns existing 3D architectural models into photorealistic renders, said on August 10 that it had closed a EUR2.16 million ($2.5 million) seed round led by Skyfall Ventures. Existing backers StartupLab, OBOS and Antler returned for the round, joined by two new investors, Farvatn and byFounders Angel Collective.
The raise follows a EUR700,000 seed round the company closed in March 2025 from StartupLab, Antler and OBOS, meaning three of the five investors now backing Visoid have already watched it grow through a full product cycle before writing a second check. Founded in 2023 by Mark Szoke, a former practicing architect, and serial entrepreneur Joachim Holwech, the company says the new capital will expand its engineering, product and commercial teams, accelerate new AI capabilities and deepen partnerships with architecture and design firms.
The Job This Software Is Actually Replacing
What makes this round worth a second look is not the amount, which is modest against the funding numbers usually attached to AI raises this year, but what the product quietly displaces. For decades, architecture firms across Europe that needed polished visualisations for client pitches or planning submissions have routinely outsourced that work to specialist rendering studios, many based in lower-cost markets in Asia and Eastern Europe, because producing a single photorealistic render in-house required dedicated GPU workstations, specialist software skills and enough turnaround time that most firms found it cheaper to send the job out and wait.
That outsourcing relationship is built on a cost-and-speed gap: an external studio could turn a render around faster and cheaper than a firm could justify staffing for the skill in-house. Visoid's pitch is that its AI closes that gap by taking a firm's own 3D model and producing a comparable render in minutes rather than days, without the firm needing to hire a renderer or brief an outside studio at all.
Why In-House Now Beats Outsourced
What separates Visoid from a general AI image generator, according to the company, is that it works directly from a firm's existing 3D model rather than reconstructing a scene from a text prompt, so the render preserves the actual design intent instead of approximating it. That distinction matters because it means the tool can be used earlier in the design process, to test massing, materials or lighting choices while a project is still being iterated on, rather than only at the end to produce a single polished image for a client presentation, which is the point at which outsourcing to an external studio traditionally made sense.
The company says its platform is now used by architects and designers in 120 countries, including established practices such as OMA and Sweco, and has produced more than 5 million renders to date. Chief executive Mark Szoke said the new funding gives the company "a strong foundation to develop Visoid for bigger, design-driven workflows," with the capital going toward hiring and product investment rather than into the render pipeline itself.
Re-Price the Rendering Line Before the Next Renewal
Visoid's own team of nine plans to double by the end of the year, but the hiring is going into engineering, product and commercial roles rather than rendering capacity, which is the detail an outsourcing studio's client should sit with. The category of work Visoid is chasing has always been driven by turnaround time and per-render cost, and both of those are exactly what AI-native, in-house tools are built to compress.
For a mid-size architecture or design practice, the immediate move is not to sign up for Visoid specifically, but to look at what is still being budgeted for external rendering studios and ask whether that line item reflects a genuine capability gap or a workflow habit nobody has revisited since GPUs and AI tools got fast enough to close it. Firms that wait for the next renewal cycle to ask that question will spend another year paying outsourced rates for work a subscription can now do in-house, the same day.
Read next: 9,437 Checkouts Just Learned to Spot Theft Together | Omilia Raises $67M as Banks Trust AI With Live Calls



