LIVE
+++ US agencies warn of AI-built exploits against Siemens S7 PLCs +++ Amazon expands Prime Air drone delivery to nearly 500 US cities +++ Munich Re to buy cyber insurer At-Bay for $575 million +++ Veeda AI raises $90M+ seed for world models +++ Queen One raises $25M for AI-native commerce CRM +++ Reuters: US military research paved way for Unitree's robot dogs ++++++ US agencies warn of AI-built exploits against Siemens S7 PLCs +++ Amazon expands Prime Air drone delivery to nearly 500 US cities +++ Munich Re to buy cyber insurer At-Bay for $575 million +++ Veeda AI raises $90M+ seed for world models +++ Queen One raises $25M for AI-native commerce CRM +++ Reuters: US military research paved way for Unitree's robot dogs +++
All news ›
AI IN LIFE AI IN LIFENEWS
DAILY
Research

Veeda AI Raises $90M Seed to Build World Models for Robots

Former Nvidia researcher Sanja Fidler has raised over $90 million for world-model startup Veeda AI — one of Canada's largest seed rounds ever.

Veeda AI Raises $90M Seed to Build World Models for Robots

Illustration · AI-generated (AI IN LIFE)

At a glance

  • Seed round: over $90 million, led by Khosla Ventures and Radical Ventures
  • Founded roughly three months before the 19 August 2026 announcement
  • Team: Sanja Fidler (CEO), Huan Ling, Zan Gojcic — all formerly at Nvidia
  • Based in Toronto; one of Canada's largest-ever seed rounds
  • Focus: world models as training simulation for robotics ("Physical AI")

What happened? Veeda AI, a Toronto startup founded only about three months ago, has closed a seed round of more than $90 million, led by Khosla Ventures and Radical Ventures. It ranks among the largest seed financings ever raised by a Canadian startup.

Who is behind it? Founder and CEO Sanja Fidler previously led Nvidia's Spatial Intelligence Lab in Toronto, where her team built world models for autonomous vehicles and humanoid agents. Her co-founders are former Nvidia colleagues Huan Ling and Zan Gojcic.

What is Veeda building? Multimodal "world models" — AI systems that simulate physical environments so robots can learn inside them. "To scale interactive learning, robots will need to learn in a simulated reality," Fidler writes, adding that Veeda's sole mission is "to build simulated reality for Physical AI".

Why are world models so hot right now? Real-world training data for robotics is expensive and slow to collect. Simulated environments promise millions of parallel training runs — an approach also pursued by Google DeepMind with Genie and by Nvidia itself. Just a day earlier, AI startup Queen One announced fresh capital too; the funding market for applied AI remains hot.

What comes next? Veeda plans to use the capital to grow its team and compute infrastructure. No product is on the market yet, and the company gave no timeline.

◈ AI-GENERATED REPORT · SOURCES LINKED

FAQ

What are world models?

AI systems that realistically simulate physical environments so robots and agents can learn inside them — faster and cheaper than in the real world.

Who is funding Veeda AI?

The $90 million seed round is led by Khosla Ventures and Radical Ventures.

Who is Sanja Fidler?

A leading AI researcher who previously headed Nvidia's Spatial Intelligence Lab in Toronto and teaches at the University of Toronto.