1. Disrupting Japan
  2. What happens to Japanese ani..
What happens to Japanese anime in the age of AI?
2026-11-10 37:38

What happens to Japanese anime in the age of AI?

Can we have AI without the AI slop?

Theoretically, I suppose we can, but the more pressing question is will we?

Today we talk with Takuya Goto, founder of AssetHub, and discuss how AI is now being used by major Japanese animation and game studios. Takuya makes the case that AI slop is just a transitional phase that the industry is already moving beyond.

He also has some very direct advice for Japanese founders thinking about doing business overseas.

It's a great conversation, and I think you'll enjoy it.


Show Notes


They key to maintaining quality with AI cost reductions
Exporting Japanese animation culture
Why Japanese enterprises are the easiest clients for startups to work with
Why a Japanese startup targeting Japanese customers should set up in San Francisco
The core difference between animation professionals and hobbyists
The real cause of the backlash against AI in animation and gaming
How to open new global markets in Japanese IP licensing
Japan's labor mobility problem (and a few possible solutions)


Links from the Founder

Everything you ever wanted to know about AssetHub
Connect with Takuya on LinkedIn



Leave a comment






Transcript
Welcome to Disrupting Japan, Straight talk from Japan's most innovative founders and VCs.
I'm Tim Romero, and thanks for joining me.
Over the past few months, there has been an increasing backlash against the use of generative AI in music, in art, and in video games. Well, today's guest is here to explain that while he understands that backlash, the trend itself is inevitable. But not in a bad way.
Today we sit down with Takuya Goto, the founder of AssetHub, who explains how AI not only can remove much of the drudgery from video game and animation creation and provide the human artists with a greater artistic control, but how generative AI actually is being used that way today. At major animation studios and AAA game companies.
Now, I've got to say, Takuya has a very interesting plan for startup growth. He moved to San Francisco to start his startup there, but from the very beginning, his customer base was always in Japan.
Traditionally, this is not something you ever want to do, but Takuya explains why it's not just a good idea today, but why it might become standard practice for Japanese startups in the future.
Takuya and I talk about where the global gaming and animation industries are heading, the challenges of maintaining quality in a world increasingly infested with AI slop, and the unique global value held by Japanese game and animation studios.
But, you know, Takuya tells that story much better than I can. So, let's get right to the interview.

Interview
Tim: So, we're sitting here with Takuya Goto, the founder and CEO of AssetHub, which is a generative AI-powered 3D modeling for game asset creation. So, thanks for sitting down with us.
Takuya: Ah, yeah. So thanks for inviting me.
Tim: Oh, it's absolutely my pleasure. So, as I understand it, the simplest use of AssetHub, a customer can upload just a simple image, and AssetHub will break this into discrete parts and create 3D meshes of each part, and then lets the user edit each part individually. Is that pretty accurate?
Takuya: Yes, yeah, it's right. So, we're basically given the AI network for 3D production, especially for making the high-grade contents. So, AI agents to handle the AI native production workflow.
Tim: So, can you walk me through a typical use case? How does an average user use your system?
Takuya: The general clients are game developers or animation. Actually, the 3D asset is the most expensive part of the production. So, in gaming, around 50% of the whole budget is spent on 3D, like character, environment, or something. And then getting positive is also quite expensive. At least it requires 10 million dollars, so it's 50 million, 100 million, or something, right? So, it's getting higher and higher. This is the biggest bottleneck of creating content. So, we support them to efficiently create a 3D asset.
Tim: So, is a customer uploading existing models? Are they uploading just 2D images and sketches? You know, where are they when they start?
Takuya: We are giving two services. The one is a self-serve style product. And then another service is FDE. I mean, so we generate 3D model for clients. So, we are working as an AI-based BPO in this case, and they customize the agent for them. First, let me explain the more typical tool usage. At first, users need to give some image to our agents. Sometimes this is a concept art. Giving those images to our agents, and then our agents analyze, okay, this is a human character with anime style. Then this pre-process is required, and they automatically apply those techniques. Separate a single character image into parts, but it requires several efforts. So, we automate that part with agents. So, give image, agent analyze and think about what is the best way to reproduce 3D model, and then just apply those techniques, and then general 3D model. So, user experience is basically, yeah, upload image, see 3D model. And if there is any issues, inspect what AI agent does. 80% of 3D assets can be done by agents. And of course, the rest of 20% must be done by human, but still a major part of the creation is automated by AI.
Tim: So, tell me about your customers. You mentioned that it's game studios and animation. How many customers do you have now? Where are they located?
Takuya: First, self-serve tools. Major users are more indie creators or personal creators. Around 98% of users globally. I mean, only 2% is Japanese users. This is called a self-serve style toolkits.
Tim: With the self-service tier, 98% is global. Can you break that down? Is it mostly the US or Korea or…?
Takuya: Yeah, so 40% is from US. Rest of the countries quite vary. European countries, Korea, China, quite broad.
Tim: It makes sense. And you were talking about your other service.
Takuya: Yeah, FDE style case, 90% clients from Japan. Our business strategy is that we'd like to export the Japanese IP creation capability. It means, I think that in physical AI case, just convert the syokunin, some expertise knowledge into software. So, at first, we heavily work with top Japanese IPs. And then we are customizing our agent for them. So, if we can meet those kinds of demands, it means our agent is quite usable, even for top contents. Once we could construct agents while working with top contents creators, just export it as software. And then the quality is very high. So, that's why the self-serve style users from other countries really love it. So, this is a basic private structure of our gold market.
Tim: So, the fact that 90% of your FDE users are coming from Japan is just because that's who you've targeted for the launch. You've decided you want to work with these studios rather than trying to work with studios in the US or some other country.
Takuya: Yeah, Japan is the easiest place to work with enterprise clients.
Tim: Why is that? I know a lot of B2B companies who would disagree with that.
Takuya: Actually, this is not my first company. I previously co-founded Gaudiy. Dollar is in total $17 million from enterprises. Basically, a major part of the Japanese startup is working with enterprise companies. But compared to the US, for example, the YC company giving the suits for SMB or even sometimes same cohorts and then show more exponential growth. It works in the US because if the service is loved by the US and the rest of regions' users also love it. But in Japan, Japanese SMB, number of them is quite limited. So basically, I need to work with enterprise clients. And then enterprise also want to work with startups. So yeah, that's why.
Tim: So, are you saying like the Japanese enterprises are easier for Japanese startups just because of the language or that they prefer to work with Japanese startups? What makes Japanese enterprises easier to work with for startups compared to, say, US enterprises?
Takuya: So Japanese enterprise people cannot speak English sometimes. So, they may prefer Japanese. But it doesn't mean that they want to work with Japanese. Just a language barrier. Just a reason why Japanese enterprise work with startup is that in Japan, less number of M&A. And also, same person keep working for same company. So they don't do any layoff and they don't acquire another company. And then just the same person keep working for same company. So, it's fine. But every enterprises sometimes have difficulties doing new things.
Tim: Well, it would seem like that would make it harder to work with the bigger companies, right?
Takuya: No, no, no. Yeah, in US, if they want to have a new business, they would acquire startup, right?
Tim: Oh, I see, I see. So, it's easier to make like long-term partnerships and things.
Takuya: Yeah, they won't fire…
Tim: Yeah, the same person is in the role for many years so you can develop this long-term trust.
Takuya: In Japan, changing their employees is really tough. To do something new, they require some support. Changing the workplace by themselves is really tough. So, they need to have a quite longer term relationship. And so just to bring some outside workforce is quite helpful. And in this case, startup is an easy way to do it because they are cheaper and sometimes do new things. This is why Japanese startup is easy to work with enterprises. I think the US, they require many things compared to Japan.
Tim: That's a good point. That's a good way of thinking about it. So, we've been talking a lot about Japan versus the US here. But you actually founded AssetHub in San Francisco, not Japan. So, why San Francisco?
Takuya: Because as a Japanese founder, I'd like to get accustomed to more of the global business. Many Japanese startups, even unicorn companies, fail to enter into the US market. They launch a branch in the US but it doesn't work well and then just fin

感想

まだ感想はありません。最初の1件を書きましょう!

37:38

コメント

スクロール