· Updated · Venture Capital Tracker · investment-strategies · 3 min read
Veeda’s $90M+ Seed: Sanja Fidler’s World Models for Physical AI
Ex-Nvidia Spatial Intelligence Lab lead Sanja Fidler’s Veeda AI raised more than $90M seed backed by Khosla Ventures and Radical Ventures — multimodal foundation world models to train embodied agents in simulation. Among Canada’s largest seeds. No product customers disclosed.
VCT data record
Funding event facts
Source-backed financing and transaction details. Unknown terms remain undisclosed rather than estimated.
Veeda AI raises more than $90M seed
Khosla and Radical back Fidler’s world-model lab. Filings show 60.6M shares at $1. No valuation disclosed.
- Event type
- Funding Round
- Event date
- Aug 19, 2026
- Stage / label
- Seed
- Amount
- $90M+
- Confidence
- Reported
Company / target: Veeda AI
Lead: Khosla Ventures , Radical Ventures
Sources: thelogic.co unite.ai
Veeda AI raised more than $90 million in seed financing reported August 19, 2026, backed by Khosla Ventures and Radical Ventures. The Logic’s exclusive — people familiar plus filings — is the source spine; Khosla confirmed. Radical has no /fund/ page here.
Unexpected truth: three months after incorporation, a Toronto world-model lab from Nvidia’s Spatial Intelligence alumni prints one of Canada’s largest seeds — before naming a single customer.
This page is for scouts comparing world-model labs (AMI, World Labs, Waabi) on capital and thesis, not ARR. Last verified August 31, 2026.
Five-minute decision
| If you need… | Verdict |
|---|---|
| What happened | Funded (reported). >$90M seed. Khosla + Radical. No valuation. |
| What Veeda is | Multimodal foundation world models for embodied-agent training. |
| Whether it is shipping | Hiring / compute buildout. No named product customers. |
| Whether to diligence | Yes, if you underwrite simulation-first Physical AI. No, if you need logos. |
Investigate further when: Fidler’s Nvidia SIL pedigree is your diligence shortcut.
Wait or pass when: you need revenue proof or a disclosed post-money.
What happened
| Field | Detail |
|---|---|
| Company | Veeda Innovation d/b/a Veeda AI (Toronto HQ; offices MV / Singapore / Zurich) |
| Founders | Sanja Fidler (CEO), Huan Ling (chief scientist), Zan Gojcic (CTO) — ex-Nvidia |
| Round | >$90M seed · reported Aug 19, 2026 |
| Paper trail | Late July: 60.6M seed shares @ $1; Strohband + Poutanen join board |
| Leads | Khosla Ventures, Radical Ventures |
Incorporated early June 2026 per The Logic’s filing review. Fidler spent eight years at Nvidia building Toronto research into the Spatial Intelligence Lab; she remains a University of Toronto CS faculty member and Vector Institute founding member.
Who uses it — and why this fund fit
Users today: none named. Target buyers are teams training embodied agents / humanoids / AV stacks that cannot parallelize unsafe real-world trial-and-error. Veeda’s own framing: high-fidelity simulated environments — “the Matrix” for physical AI.
Why Khosla + Radical write this check:
- Same syndicate co-led Discovery Loop (Jeff Dean) earlier in August 2026 — pattern: research-heavy AI founding teams.
- Direct category overlap: Radical → World Labs, Decart; Khosla/Radical → Waabi ($750M January 2026). Veeda is another ticket on world models, not a random SaaS seed.
- Capital job is people + GPU clusters, not classic go-to-market. A $90M+ seed matches that burn profile.
Likely reason Fidler raises from them: they already understand the category’s capital intensity and will not force premature enterprise packaging.
Why now
World-model financings set the bar in 2026: AMI Labs ~$1.03B (March), World Labs large rounds, Waabi’s $750M. Against that cohort, Veeda’s $90M+ is modest absolutely but enormous for a Canadian seed — and arrives weeks after Fidler left Nvidia with a public “world models is where the next breakthrough lies” line.
What is not proven
- No valuation, ARR, or named design partners.
- Share count × $1 is a paper trail, not a full cap-table reconstruction.
- Simulation fidelity claims are thesis until third-party benchmarks appear.
Competitive map
| Lab | Focus | Veeda contrast |
|---|---|---|
| World Labs (Fei-Fei Li) | Spatial intelligence | Radical sits on both — compare, don’t conflate |
| Waabi (Urtasun) | AV world models | Same Toronto / Vector network; different product surface |
| AMI Labs (LeCun) | Foundation world models at mega-scale | Veeda is smaller capital, Physical-AI training angle |
Implication
The unresolved issue is whether Veeda becomes the training substrate others build on — or another research lab that burns seed compute before a durable API. Watch hiring velocity and first named simulation customers before treating the $90M as proof of product-market fit.
Follow Venture Capital Tracker in Google
Add VCT as a preferred source to make our venture-capital coverage easier to find in Google Search.
Last updated:
Editorial note: AI tools assisted with research, structure, or drafting. Venture Capital Tracker retains human editorial responsibility for factual accuracy, relevance, and source quality before publication.
