Participated · Seed · Aug 2026

Space develops an AI-native filesystem that sits above the operating system to let applications, people, and agents access cloud-hosted files as if they were local without full-file downloads or duplicated copies. Files appear in native file browsers and open in existing applications while Space streams only the exact byte ranges required in real time. The company launched a prototype in November 2025 and is currently in private beta with roughly 100 users and teams and an organic audience of over 80,000. Founders Matthew Ao, Arihant Bapna, and Jason Zhao built the product from operational pain points encountered while working with terabytes of media. After raising a $2.4M pre-seed led by a16z Speedrun, Space plans to focus initial go-to-market efforts on video, marketing, and AEC industries and then expand into AI training-data infrastructure, media and entertainment, and other data-intensive workflows.

Participated · Seed · Jan 2026

Overworld is a Providence, Rhode Island-based startup building AI-native world models that function locally on devices such as gaming PCs and Chromebooks. Its core technology pairs diffusion models with real-time, responsive systems so open-ended, human-directed digital environments can operate without data-center latency or cloud dependence. By running on local GPUs, the company aims to reduce environmental overhead while delivering seamless interactive experiences. A recently released research preview demonstrated that its diffusion models can serve as continuous, interactive worlds on consumer hardware, validating the feasibility of its approach. Led by CEO Louis Castricato and Chief Science Officer Shahbuland Matiana, Overworld positions itself at the intersection of generative AI and immersive computing. The company plans to use fresh capital to deepen R&D, build a foundation for scalable AI-native world models, and advance its on-device technology stack. No revenue, user, or other operating metrics have been publicly disclosed.

Participated · Seed · Jan 2026

Founded in November 2025, Canada Rocket Company is a launch-focused startup staffed by a 24-person team of mostly 7+ year launch veterans. The company came out of stealth in January 2026 and was selected for Canada's Launch the North challenge. It has reported raising $22.5 million (around $21 million Canadian). Operational progress includes bringing an engine test facility in Etobicoke online and conducting a first component test this month. The team is targeting a gas-generator hot-fire test by year-end. No revenue, user metrics, or detailed product specifications were disclosed in the provided articles.

Participated · Seed · Jun 2025

Unbound’s core product is an AI Gateway that plugs into commonly used tools (e.g., Cursor, Roo, Cline or internal copilots) to provide real-time protection, model routing, and usage analytics for enterprise AI. Instead of simply blocking prompts, the platform redacts sensitive content in real time and reroutes high-risk requests to internal open-source models hosted in the organization’s cloud, preserving workflows while preventing data exposure. The company reports it has prevented the leakage of hundreds of secret credentials and more than 500 instances of personally identifiable information; customers have also cited prevention of over 7,000 potential data leaks. Unbound enables fine-grained control over model access and cost, letting teams allocate premium models to high-stakes workflows and route lighter tasks to smaller open-source models—mid-market customers have saved more than $10,000 annually and, per the company, cut AI tooling costs by nearly 70%. The product is in use with a growing base of mid-market and enterprise customers across sectors including tech and healthcare, including a deployment that supported 100+ engineers integrating Gemini 2.5 into production. The founding team—CEO Rajaram Srinivasan (former data security lead at Palo Alto Networks and Imperva) and CTO Vignesh Subbiah (former founding team member at Tophatter and Shogun)—draws on enterprise security and infrastructure experience to scale the product. Future plans for the new funding include expanding integrations across the AI ecosystem, deepening model routing capabilities, and supporting internal model orchestration for enterprises adopting open-source LLMs.

Participated · Series A · May 2025

veritree provides a platform that captures verified, site-level ecological data and transforms it into reporting to support nature restoration. Its proprietary Smart Forest Design™ includes AI-Backed Bioacoustic Monitors, a Social Impact Dashboard, and the Collect App for field data capture. The company serves nearly 300 corporate customers and, since 2023, has reached 100 million trees pledged. Led by CEO Derrick Emsley, veritree plans to roll out AI monitoring and analysis features and test and deploy the platform across ecosystems in South America and Asia. It will also accelerate go-to-market efforts and build integrations to enhance auditing, strengthen monitoring, and improve platform accessibility. veritree raised new capital to fund these initiatives.