Runway has launched Runway Ads, an AI-powered performance marketing platform designed to automate the creation, publishing, measurement and iteration of advertising campaigns. The company says the system has already been used internally since July 2026, increasing weekly ad volume from 77 to roughly 900 while doubling return on ad spend.
Runway Ads connects an advertiser’s ad accounts and brand assets, then generates video and image creatives, publishes approved variants to advertising platforms, reads performance data and creates subsequent variations based on which ads receive spend.
The product is initially being piloted with select enterprise partners, including Lovable, before a wider rollout.
Quick Summary
- Runway Ads is a new autonomous performance marketing product from Runway.
- It can generate video and image ads using brand assets, product imagery and previous winning ads.
- Approved creatives can be published directly to Meta, Google and TikTok.
- Runway says its own ad volume increased from 77 to roughly 900 ads per week since July 2026.
- The company reports 2x ROAS, 34% higher conversion and 41% lower cost per subscriber.
- Runway says more than $100 million in additional annual recurring revenue is directly attributable to its performance marketing.
- Human approval is enabled by default, while automated publishing can be enabled for selected campaigns or variants.
- The product is currently being piloted with select enterprise partners before a wider rollout.
What Runway Ads does?
Runway Ads is built around a continuous creative testing loop rather than treating ad production as a series of separate manual tasks.
Advertisers can connect their ad accounts and provide a brand kit containing information and assets that the system can use when generating advertising creative. Runway Ads can then create video and image variants based on brand guidelines, existing ads and product imagery.
The system also connects the creative process to campaign performance. Approved variants can be published directly to Meta, Google and TikTok, while performance information from those platforms is fed back into the system to inform subsequent creative generation.
That means the workflow covers several stages that traditionally require different tools or teams:
- Generating advertising creative
- Creating multiple creative variations
- Publishing approved ads
- Reading campaign performance
- Producing additional variants based on results
- Localizing creative for different markets
- Resizing assets for different placements
Runway says the system also performs an automated brand check before variants reach an approval queue. Human approval remains enabled by default, although teams can move to automated publishing for selected campaigns or variant types.
Runway reports significant results from its own advertising
The performance figures behind the launch come from Runway’s own use of the technology.
According to the company, weekly ad production increased from 77 ads to roughly 900 since July 2026. That represents an increase of approximately 1,068%, which Runway describes as a 1,000% increase in its launch announcement.
The company also reports that its return on ad spend doubled during the same period. Conversion increased by roughly 34%, while click-through rate remained steady.
Runway says its cost per subscriber fell 41% despite the larger advertising operation.
The company additionally reports more than $100 million in annual recurring revenue directly attributable to its performance marketing. This figure refers to the revenue impact Runway attributes to the broader performance marketing program rather than a claim that customers have generated $100 million in revenue from Runway Ads.
| Metric | Runway’s reported result |
|---|---|
| Weekly ad volume | 77 to roughly 900 |
| Ad-volume increase | Approximately 1,000% |
| Return on ad spend | 2x |
| Conversion | +34% |
| Cost per subscriber | -41% |
| Additional ARR attributed to performance marketing | $100M+ |
From AI creative generation to automated advertising
Runway Ads extends Runway’s existing work with AI agents beyond generating individual pieces of creative.
The company says the product builds on Runway Agent and automates key parts of both ad generation and buying. The system can identify which combinations of creative attributes are receiving spend and use those signals to generate the next batch of advertising variants.
This approach is particularly relevant to performance advertising, where advertisers often need to test numerous combinations of creative hooks, formats, audiences and messaging.
Runway’s system can also localize advertising assets according to rules defined by a marketing team. The company says this can include on-screen text, product screenshots and voiceover rather than limiting localization to spoken audio.
Creative can also be reformatted for different aspect ratios and channel specifications.
Human approval remains part of the workflow
Despite describing Runway Ads as an autonomous system, Runway has not removed human oversight from the default workflow.
Human approval is enabled by default, and generated variants pass through an automated brand check before reaching the approval queue. Teams can subsequently enable automated publishing for specific campaign or variant types.
The platform also provides controls over how advertising budgets can shift between variants. Runway says advertisers can establish limits including maximum daily changes, minimum shares of new audiences and retargeting caps.
These controls are intended to keep automated campaign adjustments within parameters established by the marketing team.
Who can use Runway Ads?
Runway says the system is designed for a range of performance marketing operations, from smaller in-house teams to large enterprises.
Potential use cases identified by the company include direct-to-consumer brands running ongoing campaigns, ecommerce companies managing large product catalogs, mobile applications entering new markets and agencies operating advertising campaigns for multiple clients.
For ecommerce, Runway says advertising creative can be generated from existing product feeds rather than requiring teams to brief individual creatives for every listing.
For mobile applications, the system can localize a winning advertisement across multiple languages and adapt on-screen text and interface screenshots alongside the voiceover.
Agencies can use the same testing and regeneration workflow across client accounts while retaining approval and brand-check controls.
Runway Ads enters enterprise testing
Runway is currently piloting Runway Ads with select enterprise partners, including Lovable. The company says it plans to expand availability more broadly in the coming weeks.
The product therefore represents a shift in Runway’s positioning from AI-assisted creative production toward a more integrated advertising workflow. Rather than stopping after generating an image or video, Runway Ads connects creative generation with distribution, performance measurement and subsequent iterations.
For advertisers, the central change is the integration of those stages into one system. Runway’s own reported results provide the initial evidence for the approach, although those figures are company-reported performance from its internal advertising operation rather than an independent benchmark across customers.
Also Read –
Runway Gen-4.5 Video Generation on NVIDIA Vera Rubin
Runway Workflows: AI-Powered Audio & Video Editing Guide
