Survey: 59% of organisations run no AI video in production

Fifty-nine per cent of organisations with video at the core of their business have no AI video features in production, despite AI dominating industry debate, according to research published on 3 November 2026 by bunny.net, a European provider of cloud infrastructure and content delivery networks (CDNs).

The Running Video in 2026 report surveyed more than 500 organisations across streaming, education, enterprise communications and creator platforms. The headline finding cuts against the conference-floor narrative of scaling, monetisation and next-generation delivery. Teams remain fixed on a more basic problem: building infrastructure that scales without adding operational burden.

What does the survey actually show?

Adoption of emerging video technology is thin across the board. Among organisations that do use AI, almost 40 per cent deploy just one AI feature. Transcription and translation rank among the most common applications. AV1, the next-generation codec with clear efficiency benefits, is in production at only 14 per cent of respondents; another 38 per cent are considering it but have no immediate implementation plans.

The report frames this as sequencing, not stagnation. Many organisations are building operational foundations first, with simplicity and scalability cited as the key drivers of video infrastructure decisions.

The survey's infrastructure numbers:

• 77 per cent use a managed video platform to avoid running encoding and delivery themselves
• 41 per cent identified predictable cost as a key consideration
• Just 10 per cent cite cost as a challenge without also citing operational complexity

Costs: bandwidth and storage dominate

Cost concerns centre on usage that grows as businesses scale and audience appetite rises. Nearly two-thirds of organisations, 63 per cent, named CDN bandwidth as the cost factor most difficult to control. Half, 50 per cent, cited storage. Premium encoding troubled 26 per cent, and enterprise DRM just 15 per cent.

The pattern is consistent. The variable, usage-linked costs of delivery weigh far more heavily on operators than the licensing and processing line items that vendors often highlight.

What does this change for video providers?

Dejan Grofelnik Pelzel, CEO at bunny.net, said the research contradicts the industry's feature-first orientation: "The video delivery industry is often focused on new features and what's next, but our research shows something different. More capability does not necessarily equate to more value."

He added: "There's clearly a demand from organisations for fewer moving parts and they're carefully weighing up the benefits of adopting new technology against the operational complexity it could introduce. The providers that succeed will be those that help organisations scale efficiently, control costs and innovate without making video delivery harder to manage."

That last sentence doubles as the company's commercial argument, given bunny.net sells exactly the managed delivery its survey says the market wants. The data on bandwidth and storage cost anxiety, however, stands on its own as a signal of where buyer budgets are strained.

Selective adoption, not rejection

The findings indicate demand for innovation remains, but organisations are becoming more selective in how they adopt it. As video grows more critical to how businesses operate, the report identifies the next challenge as building simplified, scalable foundations and greater cost visibility, so that new video technologies can be adopted at scale effectively. Providers that read the 59 per cent AI laggard figure as an education problem, rather than an operations problem, are likely reading it wrong.