Case Study — SUBMERGE: Beyond the Render
16 artists rendered 18K digital art for a 270-degree space in 2 months
Any artist working in immersive media production frequently encounters a linear hardware bottleneck. ARTECHOUSE NYC is an immersive gallery that features an 18K resolution, 270-degree panoramic canvas with a higher pixel density than even the colossal Sphere in Las Vegas — one of the most advanced 16K LED displays. In this environment, the hardware bottleneck constrains creativity.
Until now, producing high-fidelity content for an environment of this nature required many months of production time. The majority of such a cycle is consumed by the latency of serial rendering on local workstations or centralized farms.
SUBMERGE: Beyond the Render was designed to change this. By using the decentralized GPU power of Render Network, 16 world-class artists simultaneously delivered 18K resolution, cinema-quality immersive works in a fraction of the typical timeline.
The exhibition demonstrates a fundamental shift in the production model, moving from restricted local workstations to a global, parallelized rendering infrastructure. This allows studios to compress months of production into weeks or even days, while substantially reducing the high costs associated with mainstream rendering solutions.
The Challenge in Rendering Multiple Pieces at 18K Resolution
Since its founding in 2015, ARTECHOUSE has focused on creating immersive, technology-driven art experiences that exist between cinema, installation, and spatial computing. ARTECHOUSE’s flagship is located in a 100-year-old boiler room beneath Chelsea Market; the architectural space became a seamless canvas that spans three walls and the floor, reproducing the finest details in full resolution.
The key challenge of staging art on this scale is the computational cost.
When ARTECHOUSE NYC opened in 2019, it set a new benchmark with an 18K panoramic canvas, representing 44% more pixels than the famed 16K Las Vegas Sphere. Back in 2019, an artist producing a single long-form work at this scale would need up to a year using a typical local setup with a single 3090 GPU. Even for experienced studios, the combination of resolution, scene complexity, and long runtimes made immersive production slow, expensive, and operationally risky.
Render Network emerged in 2017 with the goal of unblocking these bottlenecks by providing distributed access to GPU for rendering. Artists and studios tap into on-demand compute at a price that’s affordable for studios and individual artists.
For ARTECHOUSE and the artists participating in SUBMERGE: Beyond the Render, this partnership was an opportunity to align output scale with rendering throughput. By treating extreme resolution as a baseline rather than a technical ambition, the collaboration sought to remove the boundaries of immersive media production.
The Stakes: Bringing Cinematic-Quality Immersive Art to NYC
SUBMERGE: Beyond the Render was conceived as a large-scale exhibition that showcases what becomes possible when creative ambition is untethered from rendering constraints.
The show brings together different creative approaches to 3D and immersive media rather than being curated around a single theme or template. 12 artist teams were invited to contribute works spanning cinematic world-building, generative systems, hyper-real environments, and abstract, atmospheric spaces.
“We envisioned SUBMERGE as a kind of window into the future of immersive mediums, showcasing a broad spectrum of creative, aesthetic, and technical approaches to spatial storytelling. There is something for everyone — from fully immersive SciFi and musical experiences to playful, contemplative, avant-garde, and ethereal artworks.”
— Phillip Gara, VP of Strategy, OTOY (Render Network ecosystem partner that collaborated on SUBMERGE)
However, the technical requirements were uncompromising. Every contribution had to be delivered at a full 18K resolution and for a 270-degree environment.
Content is normally produced at 4K resolution (about 8 million pixels per frame) or 8K resolution (about 33 million pixels per frame). For ARTECHOUSE’s 18K canvas, the pieces had to be rendered at about 95 million pixels per frame. That is about three times the image data of 8K without compromising fine detail, smooth gradients, complex lighting or simulation-heavy effects.
Crucially, all 12 artist submissions had to be created simultaneously on a two-month timeline from initial concept to final delivery — a feat that would have been impossible using traditional cloud or in-house rendering solutions.
This was the point at which Render Network’s role became foundational to production. For the artists, the impact was felt most directly in the creative iteration cycle.
“Some of these creators would not have taken this job if they couldn’t iterate at this speed. What would normally be a multi-week rendering marathon locally was condensed into days or even hours, changing the stakes in terms of creative freedom.”
– Riki Kim, Executive Creative Director at ARTECHOUSE
Render Network enabled the parallel rendering of 12 productions, comparable in quality to those shown on the Las Vegas Sphere.
The result was a production environment where 12 teams, using different tools and pipelines, could simultaneously work toward a common delivery target without flattening their ideas to fit narrow technical constraints.
Why Existing Rendering Solutions Fail
For a project of SUBMERGE’s scale, centralized rendering workflows would quickly stall due to technical limitations and cost-prohibitive infrastructure. The challenges with working at this scale are manifold.
Establishing a Quality Baseline. Large-format immersive visuals can be produced quickly using composited layers, stylized effects, or procedural graphics. But the 270-degree immersive space used for SUBMERGE demanded more than simply staging large-format production. When viewers are surrounded by an image, even small inconsistencies in lighting or depth can break the sense of realism. Maintaining immersion requires frames that behave like coherent physical scenes rather than flat visual compositions.
This required SUBMERGE to deliver fully rendered 3D environments where every frame preserves realistic lighting, reflections, shadows, and depth at full 18K resolution. From the get-go, the creative ambition for this exhibit was to be of cinematic-level quality.
Creating Hollywood-grade content is computationally intensive. Traditional rendering infrastructure struggles to deliver the required quality at scale.
Local GPU workstations cannot render scenes of this complexity fast enough, and centralized render farms become prohibitively expensive when multiple teams need to iterate simultaneously. Without access to massively parallel GPU compute, maintaining this visual standard across every project would make the production timeline unworkable.
The Burden of Infrastructure Ownership. For high-fidelity 3D scenes involving dense geometry, advanced shading, volumetrics, and simulation, rendering at 18K resolution requires access to hundreds — often thousands — of high-performance GPUs operating in parallel. For high-impact, one-time events, it doesn’t make financial sense for a studio or production organization to invest in the massive hardware required to render such complex scenes.
While hyperscalers such as AWS allow renting, they essentially require users to “buy” infrastructure for a set period, paying for the GPU regardless of whether they’re using 10% or 100%.
Prohibitive Enterprise Pricing. Cloud rendering services often rely on enterprise-grade GPUs such as NVIDIA’s A-series to comply with data center commercial terms. These cards are priced at significant multiples of consumer-grade hardware — a cost that is passed directly to the user, creating a massive barrier to entry for boutique teams.
VRAM and Complexity Ceiling. Standard local workstations and even many mid-tier render farms would hit a hard limit on Video RAM (VRAM) when handling SUBMERGE’s massive 18K assets. When a scene’s complexity exceeds the memory capacity of a single machine, the render either fails or slows to a crawl, creating a barrier that prevents artists from realizing the full detail of their vision.
Ingestion and Licensing Bottlenecks. Centralized cloud solutions require a setup that adds significant friction. Artists must often set up Docker containers, manage complex remote licensing for each Digital Content Creation (DCC) tool, and provide credentials for every third-party plugin. This technical learning curve diverts time and energy away from actual creative production.
The Iteration Bottleneck. With a total project timeline of only two months, the established “upload-render-download” cycle of farms becomes a fatal bottleneck. Because existing solutions require a new upload of the entire rendering file for each change, however minor, the constant back-and-forth iterations required for a 95-megapixel project would have consumed the entire production schedule.
In a typical “staggered” pipeline, rendering is a final, separate step that can take days or weeks. This limits the rapid iteration required for a world-class immersive experience. Without the ability to render in parallel across hundreds of nodes, even minor adjustments to lighting or shading could set the project back by days.
How Render Network Delivered for SUBMERGE
For the SUBMERGE project, Render Network didn’t just provide more rendering power. It fundamentally changed the mechanics of the production pipeline in several ways, delivering benefits that turned the project from vision to reality. Specifically:
- 12 artists worked in parallel over 8 weeks
- 18K resolution across a 270-degree panoramic canvas
- 70x pipeline acceleration (compared with traditional cloud rendering)
- 26 months → 1 week rendering time
1. Limitless Scalability and Scene Complexity
Render Network uses a decentralized pool of high-VRAM GPUs, orchestrated via parallel processing. This gives even solo artists the same computational horsepower as a Tier-1 VFX house.
For SUBMERGE, this meant:
Breaking the VRAM Ceiling: SUBMERGE’s 95-megapixel assets were too heavy for standard local hardware to hold in memory. Render’s ability to tap into nodes with 24GB+ of VRAM ensured that 18K frames could be processed without crashing or downscaling. Scenes included high-density particle systems, volumetric lighting, and physics-based simulations that would crash local workstations.
A Simultaneous Pipeline: Render supported artists working simultaneously, rather than a staggered, serial workflow. While 12 different artist groups worked concurrently from across the globe over the 8-week window, they could all render high-resolution tests and final frames within a compressed timeline, often overlapping with one another, but without queuing.
Time Compression: By distributing these massive files across thousands of nodes, Render turned a process that would have taken weeks on a local workstation into one that took days or even hours, ensuring the team met an unmovable launch deadline.
2. Pipeline Efficiency through Differential Uploads
In a high-resolution 18K environment, file sizes are enormous. Typical cloud ingestion, which requires re-uploading an entire scene for every small change, would have been prohibitive given the two-month timeline. Render leverages differential uploading, a system that hashes scene data to identify exactly what has changed.
For SUBMERGE, differential uploading had a significant impact on iteration, since only the new or modified data needed to be synced with the network.
The 70x Acceleration Factor: This streamlined workflow accelerated scene preparation and ingestion by up to 70x compared to typical centralized cloud workflows. The SUBMERGE artists could iterate at their own speed, moving from a “submit and wait” culture to a “tweak and render” workflow that supports the constant process of creative refinement required for world-class quality.
3. Seamless Interoperability (Octane & ORBX)
To understand Render’s impact, it is essential to clarify the relationship between OTOY’s OctaneRender engine and the ORBX file format.
OctaneRender is the industry’s first and fastest unbiased, physically correct GPU render engine. OTOY does not license Octane for use on cloud render farms, and OctaneRender is available for distributed GPU rendering on Render Network. This means Render Network is an optimal solution for studios that require Octane’s photorealistic output at scale.
ORBX is an interchangeable “container” format designed to encapsulate every element of a 3D scene, including geometry, high-resolution textures, complex lighting, and materials, into a single, atomic file natively understood by Render Network. The ORBX file format facilitates hashing and thus enables differential uploading.
For SUBMERGE, this meant:
- Zero Setup Overhead: Because ORBX acts as a self-contained exported output, it removes the dependency on native Digital Content Creation (DCC) files. Artists could work in their preferred tools — such as Cinema4D or Houdini — and export directly to the network without configuring Docker containers or virtual machines. This allowed the 12 artist groups who worked on SUBMERGE to focus entirely on the immersive creative experience rather than technical infrastructure management.
- Eliminating Licensing Friction: Cloud farms require studios to set up their own licensing and provide credentials for every remote node. In contrast, Render’s ORBX-based workflow includes licensing entitlements, meaning the SUBMERGE artists could render distributed jobs without re-entering proof-of-purchase for every individual node.
Render Network vs. Centralized Solutions
Pipeline Overview: 12 Artist Groups, 5 Distinct Technical Bottlenecks
The exhibition features twelve unique pieces produced by a diverse group of individual artists and artistic pairings, ranging from Emmy winners to emerging 3D luminaries. They faced various technical challenges when scaling their work to an 18K environment.
- AI-native pipeline scaling
- Heavy physics + light simulations
- Compressing “impossible” production timelines
- Legacy IP upscaling
- High-fidelity texture and color precision
1. AI-native pipeline scaling: Bridging the resolution gap between generative AI outputs and cinematic exhibition standards
William Selviz + Manuel Sainsily, a creative duo working at the intersection of immersive media and generative tools, created Protopica: Sin Agua No Hay Tiempo. Their core creative tool was OpenAI’s Sora, which introduced a major production constraint, as Sora outputs video at standard resolutions and cannot natively produce ultra-high-resolution frames for final exhibition.
To bridge this, the duo built a hybrid pipeline using Render Network to upscale, denoise, and re-render AI-generated assets at full production resolution. Distributed rendering enabled an AI-native workflow to move from creation to final large-format delivery without rebuilding the project in a separate pipeline.
“Render Network allows me to compete as an independent creator. In the future, this kind of IP will come from individuals, not studios.”
— William Selviz, SUBMERGE artist, quoted in PCMag
2. Heavy physics + light simulations: Overcoming VRAM ceilings and computationally heavy light-path calculations for complex, reflective 3D environments
This group dealt with immense computational loads, where high-density particle systems and physics-based lighting exceeded the memory of standard workstations.
FVCKRENDER is a Montreal-born self-taught artist whose work is defined by sharp, crystalline geometry and futuristic architecture. In his piece, Infinite Fields, the hurdle was the complexity of light-path calculations across countless sharp, reflective crystalline surfaces, which typically creates a bottleneck for centralized hardware.
David Ariew and Chelsea Evenstar faced a similar challenge. Known for their mastery of light and “cosmic” architectural environments, their piece — 1685: Immersive — was originally created as a “cosmic cathedral” collaboration with Grammy-winning artist Zedd for his 2025 Coachella set. For displaying at SUBMERGE, they utilized complex reflections and physics-based lighting that required massive VRAM to calculate.
Likewise, Josh Pierce’s digital art and motion designs focus on the intersection of nature, presence, and spiritual stillness. In his piece, Satori: The Infinite Now, the lush-looking landscapes utilized high-density scatter systems for foliage and volumetric light-beam simulations that exceeded the capacity of high-end workstations.
By using Render Network, this group had the computational headroom required to process heavy-duty simulations and industry-leading light calculations, without simplifying the creative vision or suffering system crashes.
“It feels like more of a partnership with the 3D software because it’s really based on how physics work. I throw some light in there, and then I get to see what comes out of that kind of mystical box.”
— David Ariew, SUBMERGE artist
“Render Network is an incredibly powerful tool for 3D artists to be able to generate these worlds on a really rich level that can’t be done any other way.”
— Josh Pierce, SUBMERGE artist
Compressing “impossible” production timelines: Reducing multi-year render forecasts for procedural and algorithmic work
For these artists, the main barrier was less about complexity than the timeline required to render their work locally.
MHXis a self-taught visual artist specializing in generative art and algorithmic growth. His contribution, LATTICE, is a complex display of procedural simulations. He estimates it would have taken 26 months to render at 18K on his local setup — a timeline that he compressed to one week by harnessing Render Network.
Amrit Pal Singh’s work has a very different aesthetic, but he faced a similarly impossible timeline. Singh is a Delhi-based 3D illustrator celebrated for blending digital craft with the tactile nostalgia of physical collectibles. His piece, Adventures of the Toy Maker, required rendering vast, playful worlds with soft lighting and high-density textures at a scale that was computationally prohibitive on his local workstation.
In both of these cases, Render Network acted as a “force multiplier,” compressing a 26-month forecast (MHX) into less than a week, and a multi-month estimate (Singh) into a three-day turnaround.
“It’s not just about speed; it’s about making the impossible actually possible for a solo artist.”
— MHX, SUBMERGE artist
“Seeing your work on this scale is mind-blowing for any artist. It’s almost like a superpower. On my computer, it would still take a few months to render the whole 3-minute animation. With Render Network, it took me three days.”
— Amrit Pal Singh, SUBMERGE artist
Legacy IP upscaling: Reimagining and expanding standard assets for an 18K, 270-degree panoramic canvas
These artists transformed existing high-quality work into immersive spatial experiences, bypassing the “render penalty” that usually prevents such scaling.
Maciej Kuciara and Emily Yang are Emmy award-winning creators who adapted their celebrated White Rabbit universe into a physical, spatial experience for SUBMERGE. The original version of White Rabbit, which won the 2025 Emmy for Outstanding Innovation in Emerging Media Programming, was authored for 4K on standard screens, but the ARTECHOUSE canvas required a leap to 18K for a 270-degree environment.
Woosung Kang (aka VJ WOOO) is an Emmy-nominated motion designer who transformed a decade of his legacy work into a piece for SUBMERGE titled Nebula Junkyard. It blends high-intensity scenes from multiple past projects, re-rendered for ARTECHOUSE’s 18K, 270-degree canvas.
Gavin Shapiro, known for his viral loops with billions of views, also reached into his archive, expanding his 2021 loop called Wingin It into a high-fidelity experience. The transition from a 4K screen to an 18K environment while preserving the crisp, satirical detail presented a severe hardware bottleneck.
By using Render Network, this group bypassed hardware limitations, allowing them to treat massive data sets as fluid creative playgrounds and hit deadlines that were otherwise unreachable.
“For most of my career, the scale of my imagination was limited by access to resources. SUBMERGE changes that equation… this show proves we can produce work that rivals, and even surpasses, what comes out of the biggest studios.”
— Maciej Kuciara, SUBMERGE artist
“It’s amazing that in Artechouse you can see the audience react to an immersive space with three side walls, even the floor. It’s a truly special experience.”
— Woosung Kang, SUBMERGE artist
“I wanted to take advantage of the emotional response from seeing a physical piece of performance art, and combine that with the strengths of digital tools, which let you do things you’d never be able to do otherwise.”
— Gavin Shapiro, SUBMERGE artist
High-Fidelity Texture & Color Precision: Maintaining visual integrity and seamless color transitions at a 95-megapixel scale
When work is projected at museum scale, subtle technical flaws like “banding” become glaring errors, presenting a specific technical challenge for artists relying on visual integrity for impact.
Blake Kathryn and Maalavidaa are a collaborative duo whose work explores “femme surrealism” and abstract color therapy. The River Remembers is a vibrant, pastel-hued dreamscape that relies on soft gradients and subtle transitions between hues, meaning it’s highly sensitive to compression and resolution limits. On a standard render farm, these smooth gradients often break, creating visible lines that shatter the immersive illusion.
Jess Wiseman is a 3D artist and creative technologist known for her ability to find the surreal within mundane environments. Her SUBMERGE piece, Cycle Reset, includes hyper-realistic experiences that rely on texture-first fidelity to recreate chrome sheens, fabric textures, and even the fluff of popcorn, which all demand ultra-high resolution to remain effective in a 270-degree immersive space.
For this group, Render Network provided “big studio” infrastructure to output high-bitrate files that maintained seamless color transitions and fine detail, ensuring the emotional “flow” and tactile realism were not lost.
“For the first SUBMERGE show, myself and a handful of artists rendered at half the full scale due to limitations at that time, which shows how far the technology has come in a year.”
— Blake Kathryn
“Normally, you have an idea, but then you also have the logistics of ‘Can my computer actually do it in time?’ Not having to worry about that opened pretty much any idea to be created.”
— Jess Wiseman, SUBMERGE artist
While these creative visions vary in aesthetic, they all share a common reliance on a specific set of technical breakthroughs enabled by Render Network.
What This Means for the Future of Immersive Art
The success of SUBMERGE signals a shift in the economics of digital creation. As the industry moves toward increasingly sophisticated spatial environments and higher resolutions, the accessibility of high-performance compute has become the most significant barrier.
By decentralizing the rendering process, Render Network has effectively democratized the tools for large-scale production, allowing independent artists and boutique studios to compete with Tier-1 studios. The future of immersive art lies in removing technical friction, where the leap from a 4K concept to an 18K museum installation is a seamless operational step rather than a prohibitive financial or time-based risk.
Looking toward the next iteration of digital experiences, the ability to scale compute power on demand will be a defining factor in which stories get told. Render Network is committed to providing that scalable infrastructure, ensuring that as display technology evolves, creators have the raw power needed to inhabit it.
To see how you or your studio’s production pipeline can be accelerated affordably, check out the Render Network at rendernetwork.com.
