We have been running ZER0ONE Lab Cinema Render in beta for four weeks and the support channel has become a very direct feedback instrument. Most questions come from solo freelancers and 3D studios with two to five artists — people who have never signed a cloud render farm contract and need honest answers before they upload a first .blend.

These six questions come almost verbatim at least once a week. The answers below are our standard answers, no marketing.

1. "What does a 4K render actually cost with you?"

Depends on which frame and which scene. A 30-second 4K loop (720 frames), 512 samples, Cycles OptiX, medium complexity costs between US$8 and US$18 on our A6000 tier. The wide range comes from scene complexity (sub-D meshes, volumetrics, motion blur) and denoising setting (OptiX AI-Denoise lets you cut samples for identical visual quality — saves 30–50 %).

What we won't do: promise a fixed $/frame number before we have seen the scene. That would be dishonest. Instead we offer a free calculation — send the .blend, we tell you what the job will cost before you click render.

2. "How long does uploading a 12 GB .blend take?"

On a 100 Mbps link, about 20 minutes for 12 GB. The upload chunks internally — if the connection drops, the resume starts at the last chunk, not from zero. For very large assets (environment renders, character rigs with 50 GB texture packs) we offer direct rsync to a landing pod — the cloud node then pulls the file with internal bandwidth. About 6× the speed, drop us an email to enable.

3. "What happens if the job crashes on frame 187 of 240?"

The orchestrator marks the frame as failed, spawns a new pod (or reuses a running one), renders only the missing frame, and you pay only the 187 successful frames plus the one retry. You never pay twice for the same frame.

If the crash comes from the scene itself (Cycles bug, out of memory, broken texture reference), the job auto-stops after 3 failed retries on the same frame and pages us. We look manually — usually find the problem in under 10 minutes and re-trigger.

4. "Why do you accept clients whose scene you have never seen?"

Because we do a very sharp technical check before the pod spawns: Blender-version match, Cycles engine check, texture reference graph validation, and an estimate based on sample count × tile count × frame count. 95 % of incoming jobs pass this check cleanly. The 5 % that fail, fail not because of us, but because of broken scenes. In those cases you don't pay — we refund automatically.

5. "Can I store my assets in your cloud?"

Yes, but we don't recommend it. Storage is $0.015/GB/month in R2 EU-Central, downloads into your renders are free. What we recommend instead: HDRIs and CC0 assets from our integrated PolyHaven library (128 HDRIs pre-installed) — you can reference them for free in any render without uploading.

6. "Why should I take you over RunPod or Vast.ai?"

Honest answer: if you already have a working RunPod pipeline, there's no compelling reason. We recommend ourselves for three groups: solo 3D freelancers who don't want to manage SSH keys (our CLI is zer0one-cli, plus a Blender add-on), DACH clients with EU-privacy requirements (storage EU-Central, origin in Falkenstein/DE), and clients with tight budgets who fit on the A4000 — see our cost-per-frame deep dive.

For ML fine-tuning or non-Blender workloads, take RunPod or Lambda Labs. We have no advantage there.

What we wish clients would ask

"What happens to my job if your company shuts down?" — you'd have a 30-day grace period during which all your started jobs and R2 assets are retrievable. Daily database backup, off-host replicated, documented shutdown playbook.

"Where does support run?" — direct to Michael, usually within 2 hours, at most 24. No ticket system. Email [email protected], phone +49 1573 1093501.

Ask your own question — same channel, honest answer.

Six questions 3D studios ask us about cloud rendering — and the honest answers