From the Roaring 20s to Conscious Automation: Media at its Inflection Point

Five years ago, we argued that the media industry was entering a new “Roaring 20s”: a decade of accelerated innovation, collapsing assumptions, and uncomfortable trade-offs. In 2025, that prediction stopped being theoretical. AI moved from experiment to infrastructure, post-production economics broke in plain sight, and Big Tech’s gravity reshaped the value chain. As we look toward 2026, the main question to what extent we are willing to let transformation go—and what we insist on keeping under human control.

TL;DR

• The “Roaring 20s” thesis proved directionally right: innovation accelerated, but the casualty list in post-production is now real.
• 2025 marked the transition of AI from early adoption to mainstream operational reality in media.
• Big Tech’s economic gravity (advertising, compute, IP) is reshaping the entire value chain, forcing producers into an efficiency arms race.
• Sustainability temporarily lost mindshare to AI and cost pressure, but the underlying math has not changed.
• 2026 will be defined by interoperability, format-specific technology stacks, and computer-assisted storytelling—under conscious human control.
• Responsible AI and sustainability will re-emerge, not as ideology, but as operational necessity.

Introduction

Back in 2020, at the height of the pandemic, we described the macro-economic shock combined with the rise of AI as the beginning of the “Roaring 20s”: a decade of accelerated innovation, dramatic productivity gains, and limitless opportunity for those willing to rethink how media is made and delivered. We also warned—explicitly—that it would not be a gentle transition. There would be winners, and there would be casualties. That framing was not rhetorical. It was structural.

Five years later, our predictions have aged uncomfortably well. In 2024 and 2025, several long-established and widely respected post-production facilities disappeared from the market. Technicolour, Halo, and Evolutions did not fail because creativity suddenly lost value; they failed because the economic and operational assumptions underpinning traditional post-production no longer hold. Ateliere also went down, though in that case the causes were more idiosyncratic than systemic.

Dave Cadle, CEO of the ENVY Group, when asked to review Halo and Evolutions holding company’s operations and explore whether a strengthened structure or revised strategy could help return the business to long-term financial health, said: “After three years of difficult trading conditions, the UK television and film industry is still facing an exceptionally tough period. A transformed broadcasting ecosystem, combined with wider economic pressures and sector-wide strain, has led to the closure of many respected post and VFX companies, including industry-leading names such as The Mill, MPC, Jellyfish, Directors Cut and, only this week, Wave Studios and potentially more to follow.”

(source: https://www.televisual.com/news/evolutions-and-halo-post-go-into-administration/)

At the same time, a “battle of titans” moved from conference speculation to open combat. Paramount, Netflix, and others circling Warner Bros. Discovery became a live case study in scale, debt, ego, and strategic paralysis—a process which was extensively covered in almost real-time by Doug Shapiro (Ont Battle After Another) and Evan Shapiro (Why Netflix will Fight for Warner Bros. to the Death). These struggles are not side-shows; they are symptoms of an industry re-ordering itself under pressure from technology, capital markets, and changing consumer behaviour.

Overlaying all of this is a tension that remains largely unresolved: AI versus ecological sustainability. Advanced AI systems are constrained not by algorithms but by energy density, compute availability, and cooling. The more aggressively the industry deploys AI, the more sustainability risks being pushed to the background—at least in the short term.

Against that backdrop, Limecraft rolled out eight platform releases in 2025 alone. Not as a marketing stunt, but as a deliberate expression of an agile, customer-driven approach to innovation: shipping value-added and battle-tested AI and advanced media-production capabilities incrementally, in close collaboration with the people who actually use them. That collaboration—rather than abstract disruption—remains our core belief.

2025: the Year AI Became Operational

If 2023 and 2024 were about experimentation, 2025 was the year per excellence AI became operationally accepted across the media industry. Not just by early adopters or innovation teams, but this time by the pragmatic mainstream majority.

During the 2025 DPP Leaders’ Briefing this shift was captured loud and clear: AI is no longer “a thing companies do”, but something inseparable from operational effectiveness, differentiation, and scale. Crucially, the centre of gravity moved from novelty to value. Executives spoke less about what AI could do, and more about where it should—and should not—be applied. The consensus was strikingly sober: AI’s true value lies in amplifying human judgement, removing friction, and automating routine work, not in replacing creativity or editorial responsibility. Obviously.

Meanwhile, tectonic shifts across the value chain continued as AI companies sought to “understand video” deeply enough to generate “new” images credibly. The UMG–NVIDIA and Disney-OpenAI partnerships are emblematic: AI infrastructure companies are no longer content to be neutral enablers; they are moving upstream into discovery, creation, and engagement—while carefully signalling commitments to artist protection and responsible use. Not primarily to monetise it directly, but to train models capable of understanding long-form video, narrative continuity, and sequencing.

Looking ahead, all major GenAI players are likely accelerating in the race for intellectual property. Google, ByteDance, Meta, and OpenAI, brilliantly lined up at the Mediatech Hub Conference in Potsdam in September 2025, demonstrated their capability to create premium content, albeit most of it single-shot videos for commercials at that point. To these, the “holy grail” is not a better chatbot; it is AI that can reason across a coherent sequence of shots, scenes, and story arcs.

At Limecraft, we have and will consistently reject the idea of radical automation. Storytelling, cinematography, and craft editing are metier. What can be automated is the grunt work: logging, transcription, shot detection, versioning, compliance tasks. One has to realise that automating these already eliminates up to 75% of post-production time which, in fact, is waste. That reduction is brutal, but it does not automate storytelling an sich. Editorial control remains firmly with the professional.

2025 also reinforced a clear message to supplier ecosystem: interoperability is no longer optional. Limecraft’s partnerships with editingtools.io, Lockit Network, Cuez, Gloocast, and the upgraded Limecraft panels for Adobe Premiere Pro and Avid are expressions of a broader responsibility: in a fragmented tool landscape, vendors must reduce complexity instead of adding to it.

Economically, the pressure intensified. According to The Media Stack, close to 60% of global advertising budgets are now controlled by Alphabet, Amazon, and Meta. This concentration has profound downstream effects. Producers are forced into an arms race to lower production costs, driving automation and insourcing at the expense of specialised post-production facilities. This is not a merely cultural shift, yet another symptom of the collusion between Big Media and Big Tech, as well as real, brutal, and likely irreversible.

Sustainability, regrettably, was largely swamped by these dynamics in 2025. Yet the underlying data did not change. Limecraft remained carbon neutral for the third consecutive year in a row, therefore receiving the DPP Committed to Sustainability mark and the SATIS Coup de Cœur—not only for being best in class, but for advocating structural solutions at industry scale.

💡Environmental Sustainability thrives on Operational Excellence (IABM Quarterly Journal)

Our predictions for 2026

The forces that reshaped the media industry in 2025 will likely not slow down; they are hardening into structure. Cost pressure, AI adoption, and platform concentration are now givens, not variables. The predictions below do not describe a distant future, but describe how organisations will adapt their workflows, technology choices, and operating models in the next couple of months to excel at creativity and remain competitive.

Interoperability Becomes the Norm

In 2026, interoperability will shift from competitive advantage to baseline expectation. Partnerships will increasingly be anchored in standards emerging from EBU initiatives such as the Media eXchange Layer (MXL) and Dynamic Media Facilities (DMF). One-off, bespoke integrations are becoming economically untenable. Standardised, declarative interfaces will form the foundation for scalable collaboration, enabling organisations to interoperate and trade off services to one another efficiently.

Technology Stacks that Align with Formats

One-size-fits-all platforms continue to fracture. Over the past decade, large, company-wide MAM and workflow solutions were assumed the silver bullet: a single system intended to serve every genre, department, and business unit.

In reality, several of these “monster deployments” proved slow to implement, expensive to customise, and extremely hard to adapt. Different formats place fundamentally different demands on technology—news prioritises speed and trust, scripted drama long timelines and complex collaboration, factual scale and versioning, sports real-time performance. Adding to the complexity: formats change over time. Forcing these different requirements into a single monolithic architecture typically results in compromise, operational friction, and the proliferation of parallel shadow workflows that quietly erode the original investment.

As a result, technology strategies are shifting toward smaller and format-specific tech stacks delivered as reusable, composable templates rather than bespoke, organisation-wide builds. These templates combine proven components—media intelligence, collaboration, editing, delivery, and compliance—in configurations aligned with real production patterns. This model lowers deployment friction, reduces integration risk, and accelerates time-to-value. For smaller producers in particular, it avoids the need to bet scarce resources on multi-year transformation programmes that promise coherence but rarely deliver it, while still allowing systems to evolve as formats, budgets, and business models continue to change.

Post-production Budgets Keep Shrinking

Accessible automation and AI-assisted workflows will drive a further, dramatic reduction in post-production budgets. This shift is not cyclical; it is structural. As software becomes more intelligent, tasks that were historically outsourced—logging, transcription, subtitling, versioning, compliance preparation—are increasingly being pulled back into production teams. Not because producers aspire to become post facilities, but because the marginal cost, speed, and control advantages have become too compelling to ignore. What used to require specialised infrastructure and manual labour, becomes emancipated and executed reliably as part of the production workflow, provided the tooling is well integrated and editorially safe.

This evolution is already visible in practice. At Limecraft, producers such as Warner Bros., for continuing drama, and Arrow Media demonstrate how tasks formerly associated with post-production have progressively being insourced. In both cases, AI-assisted transcription, dailies, review, and versioning were moved closer to the production team—not to replace editorial expertise, but to improve visibility and avoid external handovers. By redesigning workflows around fewer copies, fewer file transfers, and fewer manual interventions, these productions primarily achieved faster turnaround times, lower operational cost, and greater predictability, while preserving creative intent as a non-negotiable constant.

Computer-Assisted Storytelling Matures

AI will take a meaningful leap in computer-assisted storytelling and likely automate editing at the same time. Shorter cycle times and the ability to output multiple versions are the driving forces. However, especially in news and journalism, control remains non-negotiable. As Tony Haile argues, trust, context, and accountability cannot be automated away. What will change is how early AI enters the creative process.

Together with Sanoma, Limecraft has been working on tooling that transforms raw interview material directly into a first draft article. Not designed as a publishing shortcut, but as an accelerator for journalistic work. The journalist remains fully responsible for framing, verification, and editorial judgement; AI simply collapses the distance between source material and a usable starting point. Auto-editing and AI-assisted storytelling will mature along similar lines: powerful, increasingly reliable, and deeply embedded—but always under explicit human supervision.

💡 More info: https://wan-ifra.org/2025/09/media-companies-and-start-ups-join-wan-ifra-finnish-innovation-programme/

From Smarter Models to Smarter Workflows

Progress will not come from making AI models statistically “smarter” in isolation. While foundation models will certainly continue to improve, their impact in professional media environments is inherently limited without context.

The real breakthrough will come from embedding AI into well-designed workflows and enriching it with hard, structured production data: metadata, rights information, schedules, compliance rules, and explicit editorial intent. Without these constraints, AI remains impressive but unreliable; with them, it becomes predictable, auditable, and operationally useful.

Limecraft’s thesis is therefore that AI should not be implemented as a black box, but serve as AI-powered media intelligence that consciously automates decisions within clearly defined boundaries. “Conscious” referring to the fact that any pre-existing production data are taken into account. Further to this, automated actions should be transparent, explainable, and reversible, so professionals understand why something was suggested and can override it as appropriate. This distinction matters. In an industry where trust, accountability, and provenance are non-negotiable, smarter workflows will ultimately deliver more value than ever-smarter models operating in isolation.

Sustainability and Responsible Use AI – Now for Real

Finally, we are convinced that sustainability and the responsible use of AI will make a come-back in 2026. Not as a moral statement, but out of pragmatism and operational imperatives. Energy costs, intellectual property protection, and reputational risk will drive deliberate choices. In the light thereof, we must be mindful that optimising for the cheapest model makes little or no sense when marginal cost is already near zero. Therefore Limecraft will always advocate for for quality, efficiency, and credibility as a key enabler to deliver better results, with lower long-term impact.