SgtrAiArt

AI: The Battle for Video – The Future Is Here, or Is It?

Lately, we’ve been witnessing a real battle in the field of generative art, particularly in video content production through AI models.

Every independent artist has likely realized that, from this point forward, things will change dramatically. Some concepts have already been touched on in my previous articles, but it’s worth highlighting what I believe are the key points for gaining full control over the creation of an artifact.

As previously discussed, subscription-based models, whether accessed via Web Portal or API, come with pros and cons:

Pros

  • New features, upgrades, and ongoing evolution of models released over time.
  • Some platforms offer “customization” of existing models (a sort of checkpoint or fine-tuning), which is a key feature for generating specific styles.
  • Models released in recent months boast advanced filmmaking capabilities, such as controlling camera movements, synchronizing lip movements with audio for subjects, and interlacing multiple static scenes into videos of varying durations (start frame, evolution, end frame, etc.).
  • A framework that accounts for essential aspects of creating films (not just short clips) longer than 8-10 seconds.
  • An emerging feature is the ability to define in detail the character of subjects and scenes.

Cons

  • The cost per generation and the available features, from the simplest to the most complex, depend on the pricing plan; this is already clear, as companies need to recoup their investments.
  • Partial or no control over the generation of the artifact.
  • Possible withdrawal of one or more models from the market.
  • Rising subscription or plan costs tied to new feature releases or exponentially increasing demand.
  • Currently, advanced features seem limited to text-to-video generation, not image-to-video (highly likely).
  • It’s often impossible to access all model features at a “low level” via Web or API, meaning the variable input parameters users could tweak to customize the artifact (see previous articles). This is perhaps the biggest issue, but it’s a deliberate choice by companies targeting retail users.

Last week, new applications were released. For industry insiders, there’s no need to name them; the only certainty is that the average price for a complete suite with advanced features, as described, is around $300/month for a consumer subscription.

At this point, some might rightly start asking questions…

Here’s my take.

Do you really think a company that invests millions of dollars in training models, paying engineers, and covering general business costs would offer a paid service with unlimited features for consumer users? Think about it: it’s not economically feasible. Look at the history of other startups in the field—many failed years ago.

Economically, only two types of entities will likely invest in AI to produce advanced models and features:

  • Large multinationals with vast capital.
  • Eccentric philanthropists willing to throw money at it.

In other words, the stakes in the video market are too high to allow individuals, associations, or communities to take market share from major companies that rely on video media for revenue.

The real multimillion-dollar business in video artifact creation will focus on B2B agreements between large companies or studios, with customized models capable of producing, for example, feature films or movies. I don’t believe this technical opportunity will be “gifted” to the general public.

After all, the real economic value lies in the trained model, which I’ve always called the “AI-Brain.” To fully exploit its potential, you’ll need to pay a hefty price.

Those wanting to use professional services will have to shell out big bucks to get noteworthy video outputs or generate significant added value quickly. However, there are some possibilities for independence!

A certain company with a yellow and blue logo, which officially launched its suite in January 2025, if I’m not mistaken, takes a bold and noteworthy open-source approach. It gives everyone the chance to create consistent, high-quality artifacts useful for professional businesses. This node-based framework is, in my view, the true revolution in the industry. There have been similar systems before, but none as advanced in terms of both abstract and concrete system logic.

Let’s hope some “good Samaritan” supports this initiative and those who continue to embrace the passion for open source as a social model. Only then can we break free from the constraints that many will try to impose on creativity for mere economic gain.