Met Ai
passed deal
← All Deals passed
MetAI - Deal Memo
Date Reviewed: January 8, 2026 Source: Brandon Chiang (Addin Ventures) via email forward Stage: Pre-A (raising $25M) Website: www.met-ai.com
Overview
AI-powered digital twins simulation platform for industrial/physical AI training. Converts CAD/blueprints into simulation-ready environments in minutes (vs. 100,000+ man-hours traditionally).
- Founded: Q1 2023
- Location: Taiwan & US
- Contact: Daniel Yu (CEO) - daniel@met-ai.com / +1-650-880-0025
Problem
| Current State | Challenge |
|---|---|
| Training AI in real world | Too risky, slow, expensive |
| Real-world data collection | Time-consuming, difficult |
| Existing simulators | Not built for scalable, AI-native learning |
Solution
MetGen — Generative platform that instantly converts blueprints/CAD into SimReady digital twins on NVIDIA Omniverse.
| Feature | MetAI | Traditional |
|---|---|---|
| Time to digital twin | Minutes | 100,000+ man-hours |
| Approach | AI-native, generative | Manual, consulting-heavy |
| Output | Reusable, cross-industry | One-off, fragmented |
Twin Framework:
- Design Twin — CAD/CSV/BIM → SimReady in minutes
- Simulation Twin — High-fidelity virtual validation
- Operation Twin — Real-time control system integration (PLC, SCADA)
Verticals: Warehouse, Semiconductor, Data Center
Team
| Name | Role | Background |
|---|---|---|
| Renton Hsu | Founder/CTO | NVIDIA Jetson AI Ambassador (1 of 2 in Taiwan), NVIDIA DLI Instructor, 5+ years 3D tech, Vancouver Film School |
| Daniel Yu | Co-founder/CEO | Web3 CFO, AI/XR product lead, led family manufacturing digital transformation, NTU Finance |
| Dave Liu | Co-founder/President/COO | Serial entrepreneur, Founding Partner Newton Capital, ex-Wall Street trader, AppWorks VC, NTU Semiconductor B.A. & M.S. |
Market
| Metric | Value |
|---|---|
| Physical AI opportunity | $50T (NVIDIA) |
| Industrial facilities | 10M factories, 200K warehouses, 1.5B commercial cameras |
Thesis: Digital twins are to physical AI what training data was to LLMs. Essential infrastructure.
Business Model
- Co-creation revenue — Enterprise deals with manufacturing/semiconductor leaders
- Scalable SaaS — Modular environments + AI tools (future)
Traction
| Metric | Value |
|---|---|
| 2025 Q2 Revenue | $4M secured |
| 2025 Target | $8-10M |
| Key Customer | TSMC (deployed) |
| Partners | NVIDIA product teams, Wistron, automation leaders |
| Seed Round | $4M from NVIDIA (first Taiwanese startup backed by NVIDIA) |
Fundraising
| Detail | Amount |
|---|---|
| Round | Pre-A |
| Raising | $25M |
| Use of Funds | US expansion, scale MetGen |
| Previous | $4M seed (NVIDIA) |
Investment Thesis
1000x Opportunity?
Potentially YES.
- Physical AI is NVIDIA's stated next frontier ($50T)
- Infrastructure layer for industrial AI training
- If this becomes "AWS for physical AI simulation" — massive
- NVIDIA backing validates the thesis
- Contrarian timing — physical AI emerging but not consensus
Kingmaker Fit?
WEAK.
| My Value Add | Relevance |
|---|---|
| 800+ CTO intros | Low — customers are TSMC, semiconductor giants, not B2B SaaS |
| Technical scaling | Medium — but they already have NVIDIA backing |
| GTM expertise | Low — enterprise sales to industrial companies ≠ my domain |
| DeepMind network | Low — not directly applicable |
Verdict: Their customers are industrial giants in Taiwan/Asia. My network doesn't unlock their next 10 customers.
Green Flags
| Signal | Note |
|---|---|
| Taiwanese founders | Immigrant signal — building US company from Taiwan |
| NVIDIA backing | Strong technical validation |
| Real revenue | $4M secured, not vaporware |
| Elite customer | TSMC deployed |
| Technical moat | Deep NVIDIA Omniverse/OpenUSD integration |
| Contrarian | Physical AI not yet consensus |
Red Flags
| Signal | Note |
|---|---|
| Round size mismatch | $25M round vs my $10-50k check = tiny, no leverage |
| Platform risk | Heavily dependent on NVIDIA Omniverse |
| NVIDIA could build | Why won't they just do this themselves? |
| Taiwan-centric | US expansion unproven |
| Enterprise sales | Long cycles, complex deals |
| Team exits | No major exits, unproven at scale |
Key Questions for Due Diligence
- Why won't NVIDIA just build this themselves?
- What's the defensible moat beyond NVIDIA relationship?
- Unit economics on enterprise deals?
- How sticky are customers once onboarded?
- Path from services revenue to scalable SaaS?
- Founder conviction — would they turn down a $100M acquisition?
Decision
PASS — January 8, 2026
Reason
Not a bad company — interesting market, strong validation (NVIDIA, TSMC). But failed kingmaker criteria:
- Round size mismatch — $25M round, my $10-50k check is irrelevant
- No kingmaker fit — my value-add (CTO network, B2B GTM) doesn't unlock their next customers (TSMC, semiconductor giants)
- Just another check — if I can't materially change trajectory, pass
Appendix
Contacts
| Name | Role | Phone | |
|---|---|---|---|
| Daniel Yu | CEO | daniel@met-ai.com | +1-650-880-0025 |
| Dave Liu | President/COO | dave@met-ai.com | +886-956-155-882 |
| Renton Hsu | CTO | renton.hsu@met-ai.com | — |
Source
- Investor deck (14 slides) via Brandon Chiang (Addin Ventures)
- Email forward dated Oct 2, 2025
Contacts — MetAI
Founders
| Name | Role | Phone | ||
|---|---|---|---|---|
| — | — | — | — | — |
Relationship Timeline
No interactions logged yet.
Last Touchpoint
- Date: —
- Status: Passed
- Next Action: —
Notes - MetAI
2025-10-02 | Email | Deal Introduction
From: Brandon Chiang (Addin Ventures) Type: Forwarded intro from Dave Liu (MetAI COO)
Summary:
- Brandon forwarded MetAI deck and blurb
- Original email from Daniel Yu (CEO) to Sydney Sykes at NVIDIA requesting VC intros
- MetAI positioning: "scalable infrastructure for industrial simulation and physical AI training"
- Key claim: reduces 100,000+ man-hours to minutes for digital twin creation
- Raising $25M Pre-A for US expansion
Impressions:
- TBD — no founder interaction yet
- Deck is polished, NVIDIA validation is real
- Need founder call to assess George/Alex quality
Follow-ups:
- Decide if worth taking founder call given kingmaker fit concerns
- If proceed: schedule call with Daniel Yu or Dave Liu