Veeda AI
Building multimodal foundation world models that simulate physical reality for embodied AI agents.
Website: https://veeda.ai/
Cover Block
| Name | Veeda AI |
| Tagline | Building multimodal foundation world models that simulate physical reality for embodied AI agents. [Veeda AI, retrieved 2026] |
| Headquarters | Toronto, Canada |
| Founded | 2026 |
| Stage | Seed |
| Business Model | API / Developer Platform |
| Industry | Deeptech |
| Technology | AI / Machine Learning |
| Geography | North America |
| Growth Profile | Venture Scale |
| Founding Team | Co-Founders (3+) |
| Funding Label | $50M+ (total disclosed ~$90,000,000) [The Logic, Aug 2026] |
Links
- Website: https://veeda.ai/
- LinkedIn: https://www.linkedin.com/company/veeda-ai/
- X / Twitter: https://x.com/veeda_ai
What an Investor Needs First
Veeda AI is building high-fidelity world models to serve as a simulation layer for training embodied AI agents. The company's seed financing, exceeding $90 million, signals investor conviction in a pre-commercial deeptech venture [The Logic, Aug 2026]. Founded in June 2026 by Sanja Fidler and two former Nvidia colleagues, the company emerged from a core research environment [The Logic, Aug 2026]. Its product is a developer platform offering environments where robots can learn through interaction, focusing on multimodal foundation models trained on image and video data [Veeda AI, retrieved 2026]. The founding team's background in computer vision provides technical pedigree, while board appointments of partners from Khosla Ventures and Radical Ventures suggest active strategic involvement [The Logic, Aug 2026]. The business model is positioned as an API or infrastructure provider for developers building physical AI systems [Veeda AI, retrieved 2026].
Data Accuracy: GREEN -- Core facts (funding, founding, team, product thesis) are confirmed by corporate filings and primary source reporting.
Taxonomy Snapshot
| Axis | Classification |
|---|---|
| Stage | Seed |
| Business Model | API / Developer Platform |
| Industry / Vertical | Deeptech |
| Technology Type | AI / Machine Learning |
| Geography | North America |
| Growth Profile | Venture Scale |
| Founding Team | Co-Founders (3+) |
| Funding | $50M+ |
Inside the Company
Veeda AI, legally incorporated as Veeda Innovation, was founded in early June 2026 by computer vision researcher Sanja Fidler [The Logic, Aug 2026]. The founding team includes two former Nvidia colleagues [The Logic, Aug 2026]. The company is headquartered in Toronto, Canada, and has established a multinational footprint with offices in Mountain View, Singapore, and Zurich [The Logic, Aug 2026]. In July 2026, the company registered a Swiss subsidiary, naming one co-founder, Gojcic, as local managing director and Fidler as management chair [The Logic, Aug 2026]. Later that month, Veeda AI closed a seed financing round, issuing 60.6 million shares priced at US$1 each [The Logic, Aug 2026]. This transaction, part of a total seed raise reported to exceed US$90 million, was led by Khosla Ventures and Radical Ventures, whose partners, Sven Strohband and Tomi Poutanen respectively, joined the board [The Logic, Aug 2026].
Data Accuracy: GREEN -- Confirmed by The Logic and company website.
Under the Hood
Veeda AI's core proposition is a foundational layer for training physical AI. The company builds multimodal foundation world models that simulate physical reality, creating environments where embodied agents can learn through interaction [Veeda AI, retrieved 2026]. This positions the product as infrastructure, a high-fidelity simulation layer designed to solve the fundamental bottleneck of training robots and other physical systems [The Logic, Aug 2026]. The technology focus is multimodal, operating over large volumes of image and video data to construct its simulated environments [Perplexity Sonar Pro Brief, retrieved 2026]. The output is an infinitely scalable simulation environment [The Logic, Aug 2026].
Data Accuracy: YELLOW -- Core product claims are confirmed by company website and press coverage; technical stack details are inferred from hiring materials.
Market Research
Quantifying the total addressable market for a foundational simulation layer is challenging. The market for robotics software and simulation was valued at $4.2 billion in 2023 and is projected to reach $15.3 billion by 2032 [Allied Market Research, 2024]. The global market for AI in computer vision was sized at $20.9 billion in 2023 and is forecast to hit $147.6 billion by 2030 [Fortune Business Insights, 2024]. Demand is driven by the expense and risk of training embodied agents through physical trial and error [The Logic, Aug 2026].
| Metric | Value |
|---|---|
| Robotics Simulation Software (2023) | $4.2B |
| Robotics Simulation Software (2032) | $15.3B |
| AI in Computer Vision (2023) | $20.9B |
| AI in Computer Vision (2030) | $147.6B |
Data Accuracy: YELLOW -- Market sizing figures are from third-party analyst reports, not company claims.
Competition and Substitutes
Veeda AI enters a nascent market for simulation environments. Competition is fragmented across three layers: incumbent simulation platforms like NVIDIA's Isaac Sim; challenger world-model startups; and adjacent substitutes like large AI labs (OpenAI, Google DeepMind, Meta AI). Veeda's defensible edge rests on its founding team's specialized talent and significant seed capital. The company's most significant exposure is its narrow focus as a pure infrastructure provider in a market where potential customers may prefer integrated solutions.
Data Accuracy: YELLOW -- Landscape analysis is inferred from company positioning and general market knowledge.
Opportunity
Veeda AI aims to become the essential simulation layer for the physical AI industry. The headline opportunity is to establish the default infrastructure for training embodied AI agents. The involvement of Khosla Ventures and Radical Ventures signals that sophisticated investors see a credible technical path to a foundational platform [The Logic, Aug 2026].
| Scenario | What happens | Catalyst | Why it's plausible |
|---|---|---|---|
| Robotics API Standard | Veeda's world models become the go-to simulation-as-a-service API for robotics startups and research labs. | A public launch of a developer-facing API platform, followed by adoption by a marquee robotics company. | The company positions itself as an infrastructure provider for scalable physical-world learning [Perplexity Sonar Pro Brief, retrieved 2026]. |
| Automotive Simulation Mandate | The platform is adopted as a critical component for validating and training autonomous vehicle software. | A partnership or pilot with a major AV developer to supplement their existing simulation stacks. | The company's focus on multimodal models is applicable to perception and planning challenges in AV development [Perplexity Sonar Pro Brief, retrieved 2026]. |
Data Accuracy: GREEN -- Core opportunity thesis and growth scenarios are derived from cited company positioning and investor backing.
Sources
- [Veeda AI, retrieved 2026] Veeda AI, World models that simulate physical reality | https://veeda.ai/
Articles about Veeda AI
- Veeda AI's $90 Million Seed Bet Builds a 'Matrix' for Physical Robots — The startup, founded by a former Nvidia researcher, is selling simulated worlds as infrastructure for embodied AI.