AI Video Awareness Campaigns 2026: When a View Costs $0.0022 and Reach Becomes the Product

What the new awareness economics actually look like

AI video awareness campaigns have quietly become the cheapest way to buy attention in 2026: a Veo 3-made Lay's spot delivered views at $0.0022 each with a 65% completion rate, and an AI-led homeware campaign bought Meta reach at a $1.90 CPM before returning 1.9x ROAS. This article breaks down both numbers and the playbook that produced them.

Start with the Lay's case. The snack brand produced a 30-second spot entirely with Google Veo 3 alongside the agency VIVID, then launched it on Display & Video 360 in Ukraine with a split buying structure: CPM line items for maximum visibility, CPV line items to pay only for actual user engagement. The view-based line items cleared at an effective $0.0022 per view, which the agency describes as a record low, while standard targeting and brand-safety settings stayed in place.

The attention metrics held up on the other side of the trade. The AI-made spot posted a 64.7% TrueView rate and a 65% completion rate, and viewers watched an average of 24 seconds of the possible 30. On view-based buying, watched seconds are the product: an advertiser paying per view is buying completed attention, and the generated creative delivered it at snack-brand scale. Numbers like these also need {{link}} before anyone celebrates a record low.

Numbers like these also need video benchmark discipline before anyone celebrates a record low.

Why AI-made creative unlocked the view-price floor

The price per view fell for a structural reason, not a promotional one. When a 30-second spot no longer requires a shoot, a crew, or a physical location, the marginal cost of producing one more candidate version collapses toward the cost of generation and review. That surplus flows directly into the media plan as creative volume: more hooks, more openings, more variants to place against the same CPV bid.

Agencies can then treat the creative layer as iterable inventory. Generate variants, push them into view-based line items, let completion data decide which versions keep running, and retire the rest within days. That logic mirrors {{link}}.

Cheap views only matter if enough of them are real watches, and this is where completion rate becomes the economic hinge. A $0.0022 view that skips after two seconds is wasted spend; a view that watches 24 of 30 seconds is genuinely cheap attention. The 65% completion figure is what converts a low sticker price into a low effective price per usable impression, and it is the number awareness planners should audit before celebrating any CPV quote.

That logic mirrors the shift where generative creative becomes real-time inventory.

The House playbook: buy reach cheap, convert in the promo window

The second half of the evidence comes from Australian homeware retail. House, working with the agency Adcore, ran a top-of-funnel Meta Ads campaign ahead of an Afterpay seasonal promotion with one defined job: build a large retargetable audience as cheaply as possible before the promo window opened.

The creative mix tested three approaches against each other: AI-generated video as the hero asset, UGC-style clips, and CapCut-edited assets. The AI video assets won the bake-off. The campaign reached 285,150 unique users at a $1.90 CPM, roughly 80% below the team's efficiency expectations, and Adcore reports that AI video drove most of the purchases and revenue, with the overall campaign returning 1.9x ROAS.

The architecture matters as much as the headline numbers. Targeting was deliberately broad, avoiding restrictive segments so the retargeting pool would grow fast. Measurement shifted to upper-funnel KPIs such as reach, impressions, CTR, video completions, and assisted conversions, because the actual conversion event was scheduled for the promotion window, not the awareness flight.

That sequencing is the real lesson. The awareness phase was priced and judged as audience acquisition, not as direct response. Judged on last-click conversions during its own flight, a $1.90 CPM reach play would look wasteful; judged on the size and quality of the pool it hands to the conversion phase, it was the cheapest asset in the campaign.

The delivery quality that justifies view-based buying

Cheap views would be worthless if AI-made creative under-delivered once bought, and 2026 platform data says it does not. VDO.AI's data, reported in September 2026, put AI-generated ads at a 92% video completion rate on connected TV against 90% for non-AI ads, and 71% versus 70% on online video. The gap is narrow, and narrowness is the point: the quality penalty that used to offset AI production savings has shrunk to rounding error on delivery metrics.

The caveat deserves equal weight. Completion can improve while trust or purchase intent falls if the creative feels inauthentic, and VDO.AI's own co-founder notes that consumers respond to relevance and timing rather than authorship. Cheap attention that annoys the audience it pools will hand a polluted retargeting list to the conversion phase, which is why engagement-quality signals belong in the stage-two filter, not just view counts.

For awareness planners the implication is a changed binding constraint. The question is no longer whether audiences will watch an AI-made spot, because at current delivery rates they will. The question is whether the campaign design turns watched seconds into a usable, well-segmented audience, which is an architecture problem rather than a model problem.

AI video awareness campaigns: the three-stage pre-promo architecture

Stage one is buying broad, AI-made reach in the two to four weeks before a promotion opens. The objective is unique reach at minimum effective CPM, the creative is generated in volume with hooks varied systematically, and the buy leans on reach and CPV objectives so payment tracks watched attention rather than impressions nobody saw.

Stage two is building the retargeting pool from engagement signals instead of clicks. Video viewers who watched past a completion threshold, profile visitors, and page engagers form the pool, while one-time clickers get weighted lower. House's campaign explicitly grew what its agency called a large, retargetable audience base, and that pool, not the reach number, is the deliverable the awareness flight produces.

Stage three is switching objectives when the window opens. Conversion campaigns run against the pooled audience with promo-window offers, and the earlier spend is accounted for as acquisition cost per pooled user. Teams that automate this buying layer should track {{link}} before they scale the spend.

Run this way, AI video awareness campaigns stop being an experiment bolted onto the media plan and become the pricing mechanism for the top of it. The generated creative lowers the cost of the reach, the view-based objectives lower the cost of the attention, and the staged architecture makes sure the cheap attention compounds into something the conversion phase can actually spend.

Teams that automate this buying layer should track how agentic systems reached the media-buy side in 2026 before they scale the spend.

When cheap views stop being cheap

The first failure mode is frequency. Cheap creative tempts teams to hammer the same broad audience until the effective CPM stops mattering, because the marginal view stops adding pool quality. Awareness flights need frequency caps and rotation schedules just as conversion campaigns do, and the pool should be judged on its engagement distribution, not its gross size.

The second failure mode is incrementality. Attributed engagement at $0.0022 a view can flatter a campaign that would have pooled much of that audience anyway, so holdout regions or geo-splits remain the honest test of what the cheap reach actually added. The single-campaign ROAS figures in the cases above come from vendor reporting and should be read as direction, not as a guaranteed multiple.

The third constraint sits underneath the whole pool. {{link}} explains why the signals underneath retargeting pools are shrinking, and it pushes relevance back into the creative itself. A retargeting pool built from watched seconds is more durable than one built from third-party signals, but its addressability still depends on consented data, and teams should size that risk before they scale the spend.

the 2026 consent reset for video targeting explains why the signals underneath retargeting pools are shrinking, and it pushes relevance back into the creative itself.