You click a link, and a person is standing in your browser. Not a video, pinned to a rectangle, but a three-dimensional figure you can orbit with a mouse drag. The light on their skin shifts as you move, catching the texture of fabric. The audio pans spatially. The whole thing is running at 60 frames per second, and you’re doing it on a phone. This is the demo experience Gracia AI uses to show what its WebSDK can do, and it feels less like a tech preview and more like a sudden, frictionless window into another room [Andrey Volodin - Gracia AI | LinkedIn, 2026]. It’s the kind of moment that makes the abstract promise of ‘volumetric video’,the ability to capture and view a real-world scene from any angle,feel immediately, disruptively tangible.
For co-founders Andrey Volodin and Georgii Vysotskii, that moment of tangible magic is the starting pistol. Their London-based startup, Gracia AI, isn’t trying to be the Netflix of this new medium. Instead, they are building the production studio, the film lab, and the distribution network,the full-stack infrastructure required to make professional-grade volumetric content as producible as a traditional video shoot [gracia.ai/careers/cto]. Their wedge is a suite of tools based on 4D Gaussian Splatting (4DGS), a nascent AI technique that converts synchronized multi-camera footage into dynamic, photorealistic 3D scenes. The bet is that by giving filmmakers, VFX studios, and brands the pipelines to create this content easily, they will define the standards for a medium that is still figuring out what it wants to be.
The Infrastructure Wedge
The initial instinct, according to company lore, was to build a ‘YouTube for volumetric video’ [StartupHub.ai]. They quickly pivoted. The problem wasn’t a lack of audience curiosity; it was a crippling lack of professional-grade tooling to create the content that would fill such a platform. Earlier volumetric methods often relied on polygonal meshes, which could struggle with fine details like hair, fabric motion, and realistic skin, creating an uncanny, video-game-like result [Forbes, December 2025]. Gracia’s approach uses 4DGS, which represents a scene as a cloud of millions of tiny, semi-transparent ‘splats’ that can be rendered in real time. The output is notably more photorealistic and handles motion more gracefully.
Gracia’s product suite aims to be end-to-end. It covers the capture process, the AI processing to create the 4D asset, and crucially, the distribution and rendering through dedicated SDKs for major creative engines.
- Creator Toolkits. The company provides plugins and SDKs for Unity, Unreal Engine, Blender, and Adobe After Effects, aiming to slot directly into existing professional workflows [EU Tech Future].
- The Web Frontier. Their WebSDK, deployed on AWS and built with Spark.js/Three.js, is a strategic push to make high-fidelity volumetric experiences instantly accessible without a headset or specialized app [Andrey Volodin - Gracia AI | LinkedIn, 2026].
- Validation Vectors. The technology is also being pitched for applications like Physical AI and robotics simulation, where its sub-millimeter accuracy and lack of a ‘real-to-sim gap’ are key advantages [Polina Sycheva - Gracia AI | LinkedIn, 2026].
The goal is to be the plumbing. If a fashion brand wants a navigable virtual runway show, or a film studio needs to capture a stunt performer for smooth VFX integration, Gracia wants to be the underlying platform they license to make it happen.
A Team Built on Scalable Magic
The founders bring a specific kind of pedigree to this ambitious infrastructure play. Andrey Volodin, the CTO, was an early employee at Prisma, the photo-editing app that transformed selfies into art styles using neural networks. Prisma achieved a viral scale that is the stuff of startup legend: over 1 million active users within a month of launch, and more than 150 million downloads in its heyday [AI Journal, December 2025] [Prisma (app) - Wikipedia]. That experience,shipping an AI-powered visual product that captivated a global consumer audience,informs Gracia’s focus on both technical robustness and user delight.
Co-founder Georgii Vysotskii, the CEO, also worked at Prisma and brings a complementary commercial track record. He previously co-founded Netmonet, a QR-code-based tipping business that was later acquired by Alfa-Bank, Russia’s largest private bank [Founder Lodge, 2026]. This blend of deep technical AI experience and a founder who has navigated a business to an exit has resonated with investors.
| Round Date | Amount | Lead Investor(s) | Key Notes |
|---|---|---|---|
| June 2024 | $1.2M | The Venture Reality Fund, Triptyq Capital, LVL1 Group, Future Fund | Triptyq’s founder previously founded Vrvana, acquired by Apple [TechFundingNews, June 2024]. |
| December 2025 | $1.7M | EWOR | Included an angel investor described as a pioneer of Neural Radiance Fields (NeRF) [TechFundingNews, December 2025]. |
The two rounds bring Gracia’s total disclosed funding to $2.9 million [Perplexity Sonar Pro Brief]. The investor mix is telling: a blend of funds focused on virtual reality (The Venture Reality Fund), deep tech (EWOR), and angels with direct expertise in the field’s precursor technologies.
Early Signals and Deployments
For a seed-stage infrastructure company, Gracia is already logging deployments that serve as proof-of-concept vignettes. It powered what it calls the first 4DGS runway experience for streetwear brand Karl Kani [FoundersToday]. It enabled “Expedition ULUM: The Jurassic Awakening,” a mixed-reality family experience at PortAventura World [Menlo Times]. Perhaps more strategically, the company reports its system is being tested in next-generation VFX pipelines and for creative R&D at unnamed global technology companies [FoundersToday]. These are early beachheads, not yet Fortune 500 enterprise contracts, but they point to a product that is moving out of the lab and into real creative and commercial projects.
The competitive landscape includes established players like 8i and Imverse, which have been in the volumetric capture space for years. Gracia’s argument is that its 4DGS-based pipeline represents a technical leap in fidelity and motion handling, and that its focus on being a full-stack infrastructure provider,rather than a service bureau or a consumer app,creates a different kind of moat. They are selling picks and shovels, not mining the gold themselves.
The Fidelity Gambit
The most credible risk for Gracia is the same as its opportunity: it is betting on a specific, emerging technical standard. 4D Gaussian Splatting is academically exciting but still nascent in production environments. The entire field of volumetric video, while decades old in concept, has repeatedly struggled to find a sustainable, scalable business model beyond niche applications in high-end visual effects and location-based entertainment. Gracia must not only prove its tools are the best, but also that a large enough market of professional creators is ready to adopt a new, complex pipeline for a medium whose commercial outlets are still evolving.
The company’s answer appears to be a pragmatic two-pronged approach. First, by embedding directly into the tools creators already use (Unreal, Adobe), they lower the adoption barrier. Second, by pushing the WebSDK, they are expanding the potential distribution channels far beyond expensive VR headsets to any device with a browser, potentially unlocking marketing, e-commerce, and social applications that have broader budgets. The question they are implicitly answering is not ‘Can we build the most photorealistic volumetric video?’ It’s a more cultural one: what happens when the distance between being a spectator and being a virtual camera operator collapses to a mouse click? Gracia is building the infrastructure on the assumption that once that distance is gone, we’ll want to fill the space it left behind.
Sources
- [AI Journal, December 2025] Gracia AI Raises $1.7M to Build the First Full-Stack Infrastructure for 4D Gaussian Splatting | https://aijourn.com/gracia-ai-raises-1-7m-to-build-the-first-full-stack-infrastructure-for-4d-gaussian-splatting/
- [Andrey Volodin - Gracia AI | LinkedIn, 2026] WebSDK post | https://www.linkedin.com/company/gracia-ai
- [EU Tech Future] Gracia AI raises $1.7M to build 4D Gaussian Splatting | https://eutechfuture.com/artificial-intelligence/gracia-ai-builds-the-first-production-infrastructure-for-4d-gaussian-splatting/
- [Forbes, December 2025] Article on Gracia AI | https://www.forbes.com
- [Founder Lodge, 2026] Gracia AI raises $1,200,000 at Seed | https://founderlodge.com
- [FoundersToday] Article on Gracia AI deployments | https://founderstoday.com
- [gracia.ai/careers/cto] Gracia AI is a volumetric-video infrastructure startup | https://gracia.ai/careers/cto
- [Menlo Times] Article on Expedition ULUM | https://menlotimes.com
- [Polina Sycheva - Gracia AI | LinkedIn, 2026] LinkedIn post on Physical AI | https://www.linkedin.com/company/gracia-ai
- [Prisma (app) - Wikipedia] Prisma app Wikipedia entry | https://en.wikipedia.org/wiki/Prisma_(app)
- [StartupHub.ai] 4D Gaussian Splatting tools just got $1.7M for Hollywood-scale | https://startuphub.ai/ai-news/funding-round/2025/4d-gaussian-splatting-tools-just-got-17m-for-hollywood-scale
- [TechFundingNews, June 2024] Gracia AI raises $1.2M | https://techfundingnews.com/gracia-ai-snaps-1-2m-for-ai-powered-ultra-photorealistic-volumetric-videos-for-vr-and-ar/
- [TechFundingNews, December 2025] Gracia AI lands $1.7M for AI-powered ultra-photorealistic volumetric videos | https://techfundingnews.com/gracia-ai-lands-1-7m-for-ai-powered-ultra-photorealistic-volumetric-videos-for-vr-and-ar/