Summary of Key Points
Hu Yuanming, a technical genius with a background from Tsinghua University's Yao Class and a PhD from MIT, failed in his first two startups by developing underlying tools and cloud rendering services (to the point where he considered returning money to investors). In his third attempt, he shifted focus to "selling 3D models" after hearing a customer complaint. He launched Meshy, an AI-powered 3D tool that recently raised nearly $400 million in Series B financing, with a valuation exceeding $10 billion, setting a new record for the AI 3D industry. Meanwhile, there are already several players in the AI 3D space, including VAST, Yingmou, Tencent, and ByteDance. Investors continued to support him despite his previous failures, raising questions about whether it's better to invest in the individual talent or the direction of the technology.
I. Even Geniuses Make Mistakes: Why Did His First Two Startups Fail?
Hu Yuanming's first startup aimed to create an underlying tool that could replace CUDA, NVIDIA's core technology for programming GPUs. However, such foundational technologies are either too complex for giants to handle or only generate small profits through outsourcing, making it difficult to sustain a company.
His second attempt was a cloud rendering subscription service, but established software like Blender and Maya had dominated the market for decades, and users were reluctant to switch. Additionally, the soaring cost of GPUs and high global latency led to the closure of the service in 2024.
In short, his first two attempts focused on selling tools, but either the competition was too strong or users were not interested.
II. A Turning Point: From Selling Tools to Selling Results
The breakthrough came when a customer complained, "I won't pay for the software, but I'll buy the 3D models inside it for $5 each." This realization prompted Hu Yuanming to change his approach: users didn't want to learn how to create models; they wanted directly usable models.
Thus, Meshy was born. By inputting text or a 2D image, users could get a 3D model with textures ready for use in game engines in just one minute—something that would have taken professional teams two weeks to complete. Although the initial models were rough (for example, faces might appear distorted), they met the needs of horror game developers, who welcomed this "unusual effect," leading to a successful launch.
III. The AI 3D Space: A Competition Among Powerhouses
The AI 3D industry currently has several major players with different strategies:
- Meshy: Targets creators in gaming and 3D printing, offering a one-stop platform for model generation, texturing, and animation, generating annual revenue of $60 million.
- VAST: A young team founded in 1997 that focuses on producing high-quality 3D models in seconds and is exploring the possibility of creating interactive 3D worlds.
- Yingmou Technology: Established with a focus on high-precision services for businesses (80% of revenue comes from overseas), and its products were showcased at CES by Huang Renxun.
There are also open-source tools like Hunyuan3D developed by Tencent and ByteDance, but giants tend to complement these efforts rather than compete head-on. It's still unclear whether one company will dominate the market.
IV. Investors' Surprising Decision: Why Invest $400 Million Despite Previous Failures?
Existing investors in Meshy, including Source Code, BAI, and Sequoia Capital, continued to invest heavily in Series B despite the previous failures. There are two reasons for this:
1. Trust in the Talent: Hu Yuanming possesses top-notch graphics technology (the Taiji framework is used by over 300 institutions globally), and he has shown the ability to learn from mistakes and quickly adjust his approach.
2. Right Industry Trend: AI 3D represents the next big opportunity after generative art, so the direction is solid, even if the initial focus was on the wrong part of the technology stack (underlying tools rather than applications).
There's debate in the investment community about whether to support repeat entrepreneurs who may overpromise. Hu Yuanming is an example of someone who can correct course when things go wrong.
V. Lessons for Startups
Hu Yuanming's story highlights the importance of:
- Technological Talent with Practical Applications: Even exceptional skills are useless if they don't meet user needs. His failures showed that he failed to understand users' desire for direct results, not just tools.
- Failure is Not the Endgame: The ability to quickly pivot after setbacks is more valuable than initial success.
- Choosing the Right Industry Trend: While AI 3D is a promising field, finding the right focus (what users are willing to pay for) is crucial.
The AI 3D industry is still in its early stages, and no one knows who will emerge as the leader. However, Hu Yuanming's story demonstrates that combining technical talent, user-centric thinking, and resilience can lead to success in a booming market.
In summary, Hu Yuanming's journey shows that even with a brilliant background, it's essential to connect technology with real user needs and be willing to adapt.