# AI Video Production Budget Guide: Plan and Defend Video Spend

This budget guide gives marketing and procurement teams a defensible way to plan AI video spend: what drives the cost of a finished clip, how the three published price tiers map to deliverables, and how to split a quarter between hero pieces and always-on short-form. CarlosVideo18+ starts at $94 per clip.

## Why video budgets get rejected

Video budgets fail review for the same three reasons almost every time: the unit of cost is undefined, the scope is described in adjectives instead of numbers, and the figure at the bottom has no relationship to anything the business already spends. A request for some brand videos, roughly a minute each, of high quality cannot be approved, because nobody can tell whether the answer is a few hundred dollars or a few thousand.

The fix is to budget in finished deliverables. A finished deliverable is a cut that is approved, exported in the agreed formats and published. It is not a draft, not a render and not a folder of takes. Once every line item on the request maps to one of those, the conversation with finance becomes arithmetic rather than taste.

## What actually drives the cost of a finished clip

Four variables explain nearly all of the variance between quotes. Everything else is detail.

### Duration

Length is the blunt driver. A 10 to 15 second clip and a 60 plus second clip do not differ by four times the effort so much as by category: the longer piece needs more scenes, more transitions, more review surface and more revision risk. Budget line one by length, and be honest about it. A 30-second brief that quietly becomes 75 seconds is a change order, not a rounding error.

### Scene count

Two clips of identical length can sit in different price bands if one is a single continuous shot and the other is nine distinct environments. Each scene is a separate generation problem with its own continuity, lighting and subject consistency. When you write the brief, count the scenes. That number is the second-best predictor of cost after duration.

### Revision scope

Revisions are included in the published tiers, but revision scope still shapes the quote because of when feedback arrives. Notes consolidated into one round per stage are absorbed. Notes that arrive in drips, or that reverse a decision made at script stage, push work back into generation. The cheapest revision is the one you never need, which is why a tight brief is a budget instrument and not merely a creative one. See [how to write an AI video brief](https://carlosvideo18.shop/blog/how-to-write-an-ai-video-brief) for the template that keeps scope fixed.

### Variants and formats

One hero clip exported as a widescreen master, a vertical cut-down and three paid hooks is four deliverables from one concept. Variants that share a locked style are inexpensive. Variants that change the concept, the script or the pacing are new pieces of work. Specify which kind you are buying before the quote, not after delivery.

The [pricing breakdown](https://carlosvideo18.shop/blog/ai-video-pricing-breakdown-2026) covers the same four drivers at greater depth, and the [production workflow](https://carlosvideo18.shop/blog/ai-video-production-workflow) shows where each one enters the pipeline.

## The three published price tiers

CarlosVideo18+ publishes exactly three tiers, so a budget request can be written against them directly. Nothing outside these three tiers should appear on your quote.

**Single clip, from $94.** One 10 to 15 second clip in a single style. This is the test tier: hook variants, a single product point, a short announcement. It is also the natural home for always-on short-form, because the per-clip cost is small enough to approve at team level rather than at finance level.

**Standard, $150.** 30 to 60 second or multi-scene work, custom style, voiceover where the brief calls for it, and revision rounds included. This is where most campaign clips and explainer pieces land, and it is the tier to quote when a marketing lead asks for the proper brand video.

**Premium, $250.** 60 plus second pieces and full packages: complex scenes, multiple environments, alternate formats and the deepest revision coverage. A launch film with a vertical cut-down and social hooks belongs here, as does any project where the deliverable is a set rather than a single file.

Across all three tiers, revisions are included, delivery is seven days from an approved brief, and crypto is accepted. If a figure on a quote does not map to one of the three tiers, query it before approving. The [cost reference page](https://carlosvideo18.shop/seo/ai-video-production-cost) summarises the same numbers.

## Cost per finished second vs cost per finished deliverable

Both metrics are useful, and they answer different questions.

**Cost per finished second** divides the clip price by the runtime you actually publish. It is the right metric for comparing formats against each other and for spotting scope creep: a 15 second clip at the entry tier and a 60 second clip at the top tier may price closely per second, which tells you that length is being priced sensibly rather than arbitrarily. It is a poor metric for judging value, because seconds do not convert. Deliverables do.

**Cost per finished deliverable** divides total project spend by the number of approved, export-ready assets. This is the metric to put in front of finance, because it maps to the publishing calendar: how many cuts do we have to run this quarter, and what does each one cost? It also makes the economics of series production visible, since one approved concept yields several variants from a single style lock and the cost per deliverable falls as output rises.

Use seconds to sanity-check the quote and deliverables to defend the budget. Publishing both numbers in your request ends most of the argument before it starts.

## Splitting the quarterly budget: hero pieces and always-on short-form

A workable quarterly split has two lines, not twelve.

**Hero pieces, roughly one third of the quarter.** Launch films, campaign centrepieces and category-defining pieces: few in number, larger in scope, tied to a date that cannot move. They sit in the upper tiers and are briefed weeks ahead. Budget them first, because they are the items that slip when money is committed too late.

**Always-on short-form, roughly two thirds of the quarter.** The publishing engine: weekly cuts, feature explainers, internal comms, seasonal variations. Individually they sit at the entry or standard tier. Collectively they are the larger number, and they are what keeps the channel warm between launches.

The ratio is not sacred, but the structure is. Separate the two lines in the request so that a delay on a hero piece does not freeze short-form production, and so that short-form spend stops being raided to cover hero overruns. Teams that run one pooled video budget find that the loud project consumes the quiet one every quarter.

## Budget checklist to run before approving a quote

Run this list against every quote. If any line fails, the number is not ready for approval.

1. The brief is written and dated, with audience, platform, length, single message, call to action and approval owner named. An approved brief is also what starts the seven-day delivery clock.
2. Scene count is stated as a number, not described as multi-scene.
3. Every line maps to a published tier: from $94, $150, or $250. Nothing invented.
4. Revision terms are written down, including the feedback window and who consolidates the notes.
5. Variants are itemised individually, so master, aspect ratios, hooks and captions cannot drift into scope.
6. Delivery date is explicit: seven days from an approved brief, with the brief date on the purchase order.
7. Payment terms match your process. Crypto is accepted, and invoicing terms should be confirmed before the run starts.
8. A named approver exists at each stage: script, style frames and final set, one name each.

A quote that survives all eight can be defended without a second meeting.

## In-house vs outsourced: the trade-off worth defending

| Factor | In-house stack | Outsourced service |
| --- | --- | --- |
| Cost shape | Software seats plus operator time, ongoing | Per finished clip: from $94, $150, $250 |
| Budgeting | Licence spend forecast, output cost unpredictable | Publishable tiers, quotable before the run |
| Setup cost | Training, pipeline build, prompt discipline | Brief, asset pack, brand book |
| Consistency | Depends on one operator's habits | Style frames locked for every clip in the run |
| Throughput | Scales with people, not with subscriptions | Seven days from an approved brief |
| Hidden cost | Maintenance, model churn, idle seats | None outside the published tiers |

The honest version: an in-house stack wins on volume when you already have a skilled operator and a mature brand book. It loses on predictability, which is precisely what a budget review is judging. Outsourcing converts a variable, hard-to-defend cost into three published tiers and a fixed delivery promise, and the first quarter you run it you will have real numbers instead of projections.

## FAQ

### How much should we budget for AI video in a quarter?

Start from deliverables: count the hero pieces and the weekly short-form cuts you intend to publish, then price each against the three published tiers, from $94, $150, $250.

### What makes one clip more expensive than another?

Duration, scene count, revision scope and the number of variants or formats. Two clips of the same length can differ in price if one runs through nine scenes and several exports.

### Are revisions charged separately?

No. Revisions are included in the published tiers. What affects cost is scope: notes that reverse decisions made at script stage push work back into generation.

### Is cost per second or cost per deliverable the better metric?

Use cost per finished second to sanity-check a quote and cost per finished deliverable to defend the budget, because deliverables map to the publishing calendar.

### How do we split spend between a launch film and weekly posts?

Give hero pieces roughly one third of the quarter and always-on short-form roughly two thirds, on separate budget lines so a delay on one does not freeze the other.

### Is an in-house tool cheaper over a year?

Sometimes, at high volume with a trained operator. For most brand teams the licence, training and maintenance cost is hard to forecast, while a per-clip quote at published tiers can be approved before the work starts.

## Approve the first run

Write the brief, count the scenes, map every line to a published tier, and send it. Seven days from an approved brief, revisions included, crypto accepted: [Order a custom AI video](https://www.CarlosVideo18.shop/order). Published packages and examples of finished work are on the [portfolio](https://carlosvideo18.shop/portfolio) page.