3D AI

Building Meshy, a 3D AI Platform with 5M+ Users and 18× Annual Revenue Growth | Founder Interview

Honghao Wang

— 3 min read
Building Meshy, a 3D AI Platform with 5M+ Users and 18× Annual Revenue Growth | Founder Interview

Meshy — Generate 3D Models from Text or Images

Cover image generated by AI

Generative AI is transforming 3D content production — shifting from the traditional “DCC tools + outsourcing” model to a scalable “large-scale asset generation + pipeline-ready output” paradigm.

In the past five years, technologies have rapidly evolved from real-time volumetric rendering → NeRF → Score Distillation → 3D diffusion.

Demand has expanded beyond gaming and film into 3D printing, e-commerce prototyping, digital humans, education/training, AR/VR, and more.

---

The Changing Cost Curve

While challenges remain — geometric precision, UV/topology, material detail, controllability — the speed × cost curve has been fundamentally reshaped:

  • Old: 2 weeks / $1000 per model
  • Now: Minutes / $1 per model

Lower entry barriers are creating new business opportunities.

Future winners will scale asset production, integrate into standardized pipelines, and iterate rapidly.

---

Hu Yuanming: Founder’s Cross-Domain Edge

Meshy founder Hu Yuanming combines:

  • Computer graphics
  • Physics simulation
  • Systems engineering
  • Product empathy

Background:

  • Tsinghua Yao Program
  • MIT PhD
  • Creator of Taichi programming language

Key shift: From “tool licensing” to market-aligned “asset generation services” — evolving from tech-driven to user value–driven.

---

Product Positioning & Breakthroughs

Meshy’s mission: Generate 3D models from text or images.

Latest version highlights:

  • Geometry: Spatial resolution doubled on all three axes (~8× detail boost)
  • Quality: Top-tier character and fine-object generation
  • Speed: 20–30 seconds for 3D model, ~1 minute for texturing

Workflows:

  • Text-to-3D
  • Image-to-3D
  • Retopology
  • Animation
  • API integration

---

Growth & Market Performance

  • Users: 5+ million
  • Monthly site visits: 2.5–3 million
  • Revenue growth: Double-digit monthly, 18× over past year

Moat:

  • First-mover focus
  • High-efficiency organizational execution
  • Active user ecosystem

---

Entrepreneurial Lessons

Key Insights:

  • From curiosity (“how the world works”) to capability (“create worlds”)
  • Always enter growth markets, not shrinking ones
  • Shift from selling tools to selling assets
  • Cost slash → 2 weeks/$1000 → 2 min/$1
  • Speed gets you in, quality keeps you in
  • CEO mission: make the whole team win
  • Bottleneck in 3D: algorithmic paradigm for expressive, controllable generation

---

Open-Source Ecosystems & Distribution Platforms

Platforms like AiToEarn官网 integrate:

  • AI content generation
  • Cross-platform publishing (Douyin, Kwai, WeChat, Bilibili, Rednote, Facebook, Instagram, LinkedIn, Threads, YouTube, Pinterest, X/Twitter)
  • Analytics & model ranking

Such tools turn high-quality 3D creations into sustainable revenue streams.

---

From Academia to Entrepreneurship

Growth Journey:

  • Early programming sparked by love of video games
  • Gold medal in NOI → Tsinghua Yao Class
  • MIT PhD in graphics, physics simulation, computing systems
  • Built Taichi language — acted as product manager + engineer + marketer
  • Transitioned from academic research to practical products

Pivot Lesson:

Initial attempts (Taichi commercial model, renderer) faltered → recognized shrinking markets.

User feedback revealed assets > tools → inspired Meshy.

---

Product Evolution: Meshy 6 Preview

Upgrades:

  • Geometry precision: doubled spatial resolution → ~8× detail
  • Eliminated major structural errors (e.g., incorrect finger counts)
  • Speed maintained: 20–30 sec (model), ~1 min (texture)
  • Strongest improvements in characters & fine objects
  • Expanded suitability for 3D printing use cases

Remaining challenges:

  • Topology not yet artist-level
  • Limited controllability for direct AI edits

---

Core Functions

Meshy offers:

  • Text-to-3D
  • Image-to-3D
  • Texture generation
  • Animation generation
  • API integration

Design philosophy:

  • Ease of use
  • Accessible features for beginners
  • Extensible APIs for professionals

---

Real-World Impact: User Stories

  • Stratton Studio: Tank model workflow reduced from 4 hrs → 4 mins using Meshy + Blender
  • Injured 3D modeler regained creative ability using Meshy after losing capacity for traditional tools

---

Technical Roadmap

Focus areas:

  • Data quality
  • Algorithm innovation
  • Compute scaling

Challenges:

  • Imprecise geometry
  • UV/topology optimization
  • Material/texture detail
  • Enhanced controllability

Approach:

Find top talent → structured problem-solving → fast iteration.

---

Pricing & Business Model

  • Freemium core product
  • Paid tiers: Pro, Studio, Enterprise
  • Educational discounts
  • Future: deeper industry partnerships (gaming, film, AR/VR)

---

Future Directions (3–5 Years)

  • Complete 3D scenes generation
  • Improved pipeline integration
  • Features like “3D-to-Video”
  • Higher quality & controllability
  • Exploration into interactive 3D, VR/AR

---

International Strategy

  • Internationalized UI from the start
  • Localization for prompts & UX
  • Current focus: Europe/North America
  • Strategy: adapt to local ecosystems

---

Competitive Landscape

Meshy’s moat:

  • First-mover advantage in 3D generation
  • Fast, flat-team execution
  • Deep user relationships

Open-source: valuable for ideas & R&D → Meshy Scholarship supports students.

Big tech: possible partners/clients; Meshy benefits from startup agility.

---

Pivot from Taichi Graphics to Meshy.AI

Path:

  • Explored NeRF → chose generation over reconstruction
  • Built first version in 8 hours → 1K users day one
  • Iterated despite initial low quality → improved steadily

---

Leadership Evolution

CEO shift:

  • From personal execution (coding) → team orchestration
  • Focus on:
  • Strategic decisions
  • Team standards
  • Role alignment

---

Closing Thoughts & Advice for Founders

Lessons:

  • Face reality — validate actual user need & retention
  • Stay brave — accept failure, iterate quickly

Inspiration:

  • OpenAI — persistence & first-principles approach
  • Epic Games — synthesis of tech & commercial success
  • Lovable, Cursor — AI in real user workflows

---

(Links preserved)

---

Contact Directory

(WeChat IDs preserved as in original)

---

Final Note

In the evolving AI 3D content space, tools like Meshy and AiToEarn官网 illustrate how combining creation + distribution + monetization can empower global creators.

By integrating generation, cross-platform publishing, analytics, and model ranking, they enable creators to focus on delivering value — and scaling success.

Read more