Netflix says The Eternaut (El Eternauta) contains the first generative-AI material to survive into the final footage of a Netflix original film or series. Co-CEO Ted Sarandos made the disclosure during Netflix’s July 17, 2025, second-quarter earnings interview, describing a Buenos Aires building-collapse visual-effects sequence created with Netflix’s Eyeline Studios and production VFX artists. Netflix said the work was completed about 10 times faster than a traditional VFX approach, but it did not publish an independent benchmark, cost breakdown, model name, or shot-by-shot accounting.
That is a narrower claim than “Netflix made an AI episode.” The available evidence describes AI-generated or AI-assisted imagery incorporated into a conventional production, not a cast, episode, or series made without human artists.
| # | Preview | Product | Price | |
|---|---|---|---|---|
| 1 |
|
The Eternaut | $40.87 | Buy on Amazon |
| 2 |
|
El eternauta 1969 / The Eternaut (Spanish Edition) | $20.30 | Buy on Amazon |
| 3 |
|
The Eternaut 1969 | $9.99 | Buy on Amazon |
| 4 |
|
El eternauta (Cómic) / The Eternaut (Comic Book) (Spanish Edition) | $28.95 | Buy on Amazon |
| 5 |
|
Dungeon Crawler Carl, Vol. 1 (Graphic Novel) | $12.73 | Buy on Amazon |
What Netflix actually announced
During the Netflix Q2 2025 earnings interview on July 17, 2025, Sarandos called the footage Netflix’s first “GenAI final footage” in a Netflix original series or film. The production he identified was the Argentine science-fiction series The Eternaut, which premiered on Netflix in April 2025.
“First” should be read as Netflix’s own public claim: the first use the company identified in which generative-AI material reached final footage in a Netflix original. It is not an independently audited claim to be the first such use anywhere in Hollywood, and it does not mean Netflix had never used machine learning, digital effects, virtual-production systems, or other automated tools in earlier work.
#1 Best Overall
| Phrase | What it means here |
|---|---|
| AI was used in production | Too broad; it could include many unrelated tools. |
| AI helped create a VFX shot | Consistent with the reported workflow. |
| AI-generated final footage appeared | Netflix’s stated claim. |
| AI made the scene | Not established. |
| AI replaced the VFX team | Not supported; reports describe VFX artists’ involvement. |
What viewers are seeing
Public accounts describe a visual-effects sequence showing a building collapsing in Buenos Aires. Some secondary reports place it late in the series, but Netflix’s investor disclosure does not provide a shot list or an independently verified episode location. The defensible description is therefore a building-collapse sequence, not a claim that an entire episode or even every frame of that sequence was generated by AI.
No public production breakdown identifies the model, vendor, prompt-to-video method, number of generated frames, or the division between generated material and conventional simulation, tracking, compositing, matte work, cleanup, and grading. “AI scene” is useful headline shorthand but technically imprecise.
How Netflix says the sequence was made
Eyeline and human VFX work
Netflix’s Eyeline Studios, its in-house VFX and production-technology operation, reportedly worked with the show’s producers and VFX personnel. Eyeline should not be understood as an autonomous AI studio or as a replacement for the production crew. Directors, supervisors, artists, compositors, and other technicians still determine what is usable, integrate it with photographed material, correct errors, and approve the final image.
Rank #2
The 10-times-faster claim
Netflix said the sequence was completed roughly 10 times faster than it would have been through traditional VFX workflows. That is a company statement attributed to Sarandos, not an independently verified benchmark. Netflix did not disclose the comparison’s labor hours, shot count, revision schedule, cleanup time, or total cost. A faster generation step could also create downstream work that is not visible in the headline figure.
Why budget feasibility matters
Reporting said the desired collapse effect may not have been financially feasible for the production without the AI-assisted approach. That could mean the technology enabled a particular shot that otherwise would have been cut or simplified; it does not establish a specific dollar saving or prove that the whole series became cheaper.
What “AI-assisted” can mean in a VFX pipeline
A generative system may contribute reference imagery, previs, a background element, a destruction plate, or material later composited with conventional effects. Artists can then track the shot, match perspective and lighting, remove artifacts, enforce continuity, and combine the result with practical photography or computer graphics. The public record does not establish which of those steps occurred on The Eternaut.
Rank #3
That distinction matters because final-shot responsibility remains human even when a model supplies pixels. A production may use AI to explore an idea quickly while relying on conventional tools for repeatability, revisions, physics, and delivery specifications.
Why the disclosure has become controversial
Labor and bargaining power
VFX artists, animators, concept artists, editors, illustrators, writers, and performers worry that an “assistive” workflow can become a reason to reduce headcount, rates, or entry-level opportunities. A shot that becomes feasible with fewer labor hours may expand a production’s ambition, but it may also shift bargaining power from specialized workers to studios and technology vendors.
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallCrashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteCredit and authorship
If an artist directs, selects, edits, composites, repairs, and approves generated material, “AI-generated” can hide substantial human work. Productions still need decisions about screen credits, compensation, residuals, ownership, and responsibility for mistakes.
Consent, likenesses, and union rules
The 2023 SAG-AFTRA television and theatrical agreement included informed-consent and compensation protections for digital replicas of performers. The union’s announcement describes those protections. They do not automatically answer every question about non-performer VFX imagery, but they show why a production’s use of generated likenesses or new performances requires contractual scrutiny.
Copyright and training data
The undisclosed model and data terms leave important questions: Was the system enterprise-secured? Were production assets excluded from vendor training? What rights attach to the output? Can the studio defend the production if training data or generated material is challenged? Netflix has not publicly answered those questions for this sequence.
Quality and continuity
Generative video can produce unstable textures, changing architecture, incorrect motion, flicker, and physics that do not hold across frames or revisions. Those are known workflow risks, not claims that such defects appeared in The Eternaut. A frame-by-frame assessment or production testimony would be needed to make that judgment.
Quick wins for a faster PC:
Scan for outdated or missing drivers - takes under a minuteDriver Scan →Clear out junk files and repair common Windows errorsFree Scan →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Best Value
Netflix’s later production guardrails
Netflix subsequently published guidance for production partners using generative AI. The guidance asks partners to disclose intended use to their Netflix contact and emphasizes:
- human input and oversight;
- enterprise-secured tools where possible, with production inputs and outputs not stored, reused, or used for training;
- no replication or substantial recreation of identifiable characteristics from unowned or copyrighted material;
- consent before replacing or generating talent performances or other union-covered work;
- written approval when generated material appears in final deliverables or involves talent likenesses, personal data, or third-party intellectual property.
This guidance was published after the July 2025 disclosure. It is a later operational framework, not proof that The Eternaut was produced under every provision or that the rules apply retroactively.
What this does—and does not—prove
What it demonstrates
- Netflix is willing to put generative-AI imagery into final VFX footage in a scripted original.
- The company sees speed and production feasibility as meaningful benefits.
- Human-supervised AI work can be integrated into a conventional VFX pipeline.
What it does not demonstrate
- that Netflix made an AI episode or an AI-made series;
- that the whole collapse sequence was generated by a model;
- that VFX artists were eliminated;
- that generative AI is cheaper after cleanup, revisions, supervision, and rights clearance;
- that AI can replace conventional VFX across Netflix productions.
Netflix has not published evidence, as of August 18, 2026, showing that conventional VFX has been replaced across its slate.
The unanswered questions that matter
A fuller production account would identify the model or models, the exact shots and pixels affected, the artists and vendors involved, the labor-hour comparison behind the 10× claim, the total-cost calculation, the rights and data safeguards, and how many revisions the workflow supported. It would also clarify whether the approach was a one-off solution for an otherwise unaffordable shot, a pilot for wider adoption, or a signal in negotiations with VFX suppliers and workers.
Free tools Windows power users keep installed
One-click scans. No signup required.
For now, The Eternaut is best understood as a significant production precedent, not proof of an AI-native television future. Netflix has shown that it will use generative systems for final visual effects; the unresolved question is whether that expands what artists can make, reduces the number of artists hired, or does both at once.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
