The End of the Era of “800,000 Yen per Video”

I want to ask small and medium-sized business owners: When was the last time you created a recruitment video or a product promotional video for your company?

“We made one three years ago for 800,000 yen,” or “We can’t afford the budget, so we just uploaded something we shot on our smartphones.” I’ve heard these sentiments countless times. When outsourcing to a video production company, the costs for planning, shooting, editing, and color correction range from 600,000 to 1,200,000 yen. For small and medium-sized enterprises, video content has been the top example of something they want to create but can’t.

This structure is beginning to fundamentally change as we enter 2025. The open-sourcing of Netflix’s AI video editing model “VOID,” the emergence of the AI color grading system “LumiVideo,” and the evolution of automatic video generation technology utilizing reward feedback are all contributing to the collapse of the cost structure in video production.

800,000 yen is becoming 80,000 yen. This is not an exaggeration; from a technical standpoint, that level is within reach.

How Netflix’s “VOID” Shattered Conventional Wisdom—AI Editing That Understands Physical Laws

Netflix’s “VOID (Video Object and Interaction Deletion)” is a technology that removes objects from video footage. However, it operates on a completely different level compared to traditional “eraser tools.”

In conventional video editing, the task of removing people or objects was done manually by professional VFX artists, frame by frame. Even for a 10-second video, it was not uncommon for the process to take several hours. Outsourcing this task alone could cost anywhere from 100,000 to 300,000 yen.

What makes VOID revolutionary is its understanding of “physical causality.” For example, if you want to remove a person holding a cup on a table, traditional methods would leave the cup floating in mid-air. VOID understands the physical relationship that the person is supporting the cup and can naturally process both the person and the cup simultaneously. It automatically adjusts for changes in shadows, light reflections, and the impact on surrounding objects.

The significance of this technology being made publicly available is immense. Video processing that was previously impossible without Hollywood-level budgets is becoming feasible with just a single PC.

The Democratization of Color Grading—The Shock of LumiVideo

One often overlooked aspect of video production costs is color grading.

The biggest difference between amateur footage and professional footage lies in the “color.” Even when shot with the same camera, the impression of the video can change dramatically depending on the color grading. Many corporate PR videos feel “cheap” primarily due to inadequate color grading.

Hiring a professional colorist can cost between 50,000 and 200,000 yen per video. Even for short videos, this expense can be significant. It’s understandable that small and medium-sized enterprises cannot allocate budget for this.

LumiVideo is a system that replicates the workflow of a professional colorist using AI. It analyzes physical lighting conditions, scene content, and the overall tone of the video to automatically generate cinematic color grades. Moreover, it can respond to natural language instructions like “make it a bit warmer” or “deepen the shadows,” allowing for creative feedback to be incorporated.

What used to cost 50,000 to 200,000 yen for color grading could potentially be achieved for just a few hundred to a few thousand yen in AI processing costs. The impact of making the process that determines the “professional look” of a video nearly free is immeasurable.

Reward Feedback-Based Automatic Video Generation—From “Days” to “Hours”

The third technological innovation is automatic video generation utilizing reward feedback.

Traditional video generation AI operated by inputting a prompt and producing a video in one go. Naturally, the chances of getting the desired footage were low, requiring multiple attempts.

The new approach provides “reward signals” at each stage of the generation process, allowing the AI to autonomously improve quality. By giving feedback like “this part is good” or “this part is not,” the AI learns and produces better results in the next generation.

The practical impact is that the production of short videos (30 seconds to 1 minute), which used to take 3 to 5 days, could now be completed in just a few hours. The dramatic reduction in lead time from planning to completion will enable small and medium-sized enterprises to meet their unique speed demands, such as “I need this ready for next week’s exhibition” or “I need to respond to sudden recruitment needs.”

How the Cost Structure Will Change

Let’s compare the traditional video production costs with the costs after utilizing AI.

Typical Costs for Traditional Recruitment Video Production (about 2 minutes):

• Planning and Structuring: 100,000 to 150,000 yen
• Shooting (Cameraman + Equipment): 150,000 to 250,000 yen
• Editing: 150,000 to 250,000 yen
• Color Grading: 50,000 to 150,000 yen
• Narration and BGM: 50,000 to 100,000 yen
• Total: 500,000 to 900,000 yen

Estimated Costs After Utilizing AI:

• Planning and Structuring (Human): 50,000 to 100,000 yen
• Shooting (Smartphone + Simple Equipment): 10,000 to 30,000 yen
• AI Editing (VOID, etc.): A few thousand to 10,000 yen
• AI Color Grading (LumiVideo, etc.): A few hundred to a few thousand yen
• AI-Generated Narration and BGM: A few thousand yen
• Total: 70,000 to 150,000 yen

This roughly translates to one-fifth to one-tenth of the original costs. Of course, the quality of the shooting material and the planning skills still rely on human effort. However, the drastic reduction in costs for the “post-editing” phase significantly lowers the barrier for “let’s just try making one.”

From “Can’t Make It” to “No Reason Not to Make It”

At this point, I want to ask: Is your company unable to make videos, or is it simply that the costs are too high?

If it’s the latter, that reason is fading away.

For small and medium-sized enterprises without recruitment videos, job seekers may see them as “companies with nothing to show.” Data shows that without product PR videos, conversion rates on e-commerce sites can be up to 80% lower compared to competitors with videos (Wyzowl, 2024). The cost of not having videos is greater than the production costs.

Of course, AI is not a panacea. Neither VOID nor LumiVideo guarantees perfect output at this stage. Human direction is still necessary, and creative judgment is essential for expressing a brand’s worldview.

However, the world that offered only the choice of “spending 800,000 yen to hire a professional or doing nothing” now includes the option of “let’s try making one for 80,000 yen.” This change will fundamentally alter how small and medium-sized enterprises communicate their information.

The day when outsourcing costs for video production disappear has not yet arrived. However, the day when those costs become “one-tenth” is already here. If you’re going to act, now is the time.