--- title: "Q&A: Filesystem Startup Space's Founders on Building a Cloud Workspace for AI Agents" description: Space co-founders Matthew Ao, Arihant Bapna and Jason Zhao on streaming cloud files to AI agents, enterprise pilots and lessons from running Eden. author: Darie Nani (Editor-in-Chief) date: 2026-09-15T09:03:12.143Z updated: 2026-09-15T09:03:12.166Z canonical: https://www.sovereignmagazine.com/article/space-founders-qa-ai-agent-filesystem image: https://cdn.nanimediahouse.com/space-cofounders-qa-hero-v1.webp categories: Artificial Intelligence, Startups content_type: Spotlight region: San Francisco publication: Sovereign Magazine schema_type: Article about: - type: Organization name: Space description: Space (Space Computer, Inc.) is a San Francisco startup building an AI-native distributed filesystem that makes cloud-hosted files behave like local ones. Files appear in Finder and open in native applications while Space streams only the byte ranges an application or AI agent requests, using no local disk space. Founded by Matthew Ao, Arihant Bapna and Jason Zhao, the company raised a $2.4 million pre-seed round led by a16z Speedrun in August 2026. url: https://spacefs.com foundingDate: 2025-11-01T00:00:00.000Z industry: Cloud Storage sameAs: - https://x.com/getspaceso - https://github.com/Space-Computer-Inc - https://www.linkedin.com/company/spacecomputers --- Space recently announced [a $2.4 million pre-seed round led by a16z Speedrun](https://www.sovereignmagazine.com/article/space-a16z-speedrun-streaming-filesystem), with participation from Golden Ventures, Northside Ventures, and angel investors. Now live and generating revenue, the company is building a distributed filesystem for humans and AI agents. Co-founders Matthew Ao, Arihant Bapna, and Jason Zhao discuss the product and their roles in building it. **Space started with a problem your own team faced. How did that become the company you are building today?** **Jason Zhao:** I had accumulated years of footage across hard drives and cloud services, making finding and moving files a recurring problem. At Eden, our previous company, that became a team-wide constraint. Producing content together meant moving large files between people and losing time to transfers and duplicate copies. We wanted files available inside the applications where people work. Cloud storage gave us somewhere else to put our files, but it never made them feel truly local. Space grew out of closing that gap. **Your focus now includes infrastructure for AI agents. What connects that to the original problem?** **Matthew Ao:** People and agents face a similar challenge: their files need to stay accessible as work moves between environments. For people, that might mean switching computers or applications. For agents, it might mean moving between cloud environments or handing work to another agent or person. Agents need a workspace to read files, create new ones, and save progress. That work should remain available when the machine running the agent changes or shuts down. We are giving agents persistent cloud filesystems they can share with collaborators. An agent can then continue a task on another machine, and a person can open and review its work in the applications they already use. **What difference does it make for an agent to have its own filesystem in the cloud?** **Arihant Bapna:** Many enterprise cloud products are built around uploading, downloading, and syncing copies. In that model, agents download files into their own environment before their tools can analyze or modify them. The data is already in the cloud, but it still has to move onto the machine doing the work. Across more agents and larger datasets, that means repeated transfers, waiting, and duplicate storage. Space mounts that data as a filesystem inside the agent's cloud environment. Files appear immediately, and Space streams the portions its tools request from object storage. Work can begin without downloading complete files or syncing the dataset first. Agents access ordinary files and directories, receive data as needed, and save changes back to the shared filesystem. **How does Space enable humans and AI agents to collaborate on real work?** **Jason Zhao:** Humans and agents can work on Space with the tools and applications they already use because it is architected as a filesystem rather than a sync layer. Any task involving reading or writing a file can be completed in the cloud and on Space. One of our current enterprise pilots uses Space to store and back up email data, then runs agent workflows against it to assist with sales. Other teams are evaluating Space to power internal agent tools and build persistent company knowledge. **What does each founder lead, and how does that shape the product?** **Matthew Ao:** I lead product, translating the technology into something people and companies can use and deciding which problems we address first. **Arihant Bapna:** I lead engineering, focusing on the filesystem infrastructure: reads, writes, and the operations applications depend on. **Jason Zhao:** I lead go-to-market: communicating the technology's value, understanding customer workflows, and developing adoption. We established these responsibilities at Eden and brought that working experience into Space. **What is technically distinctive about Space?** **Arihant Bapna:** Space turns object storage into a streaming-based filesystem existing applications can use, streaming requested portions of files without downloading the entire dataset first. Applications also need renaming, directory management, locking, and random access. Our architecture tracks filesystem metadata and transactions directly in object storage. The engineering challenge is supporting those operations while minimizing remote requests and coordinating changes across a shared filesystem. Agents' tools can use the filesystem interface they already expect. Build tools, scripts, and desktop applications can access files without a separate object-storage integration for each tool. **How did building Eden prepare you to build Space?** **Matthew Ao:** We had already bootstrapped a revenue-generating business together. That gave us experience making product decisions with real constraints and working through the consequences as a team. **Arihant Bapna:** It also gave us an established working relationship. We understood our responsibilities and how to collaborate before starting a technically demanding infrastructure product. **Jason Zhao:** And we had experienced the storage problem as operators. Moving files was part of growing the business, so we understood firsthand how storage friction could become a bottleneck at scale. **What are you building toward next?** **Matthew Ao:** With Space addressing the file-access problem, we are exploring interfaces that let people orchestrate agents to complete more complex tasks with those files: editing videos, generating CAD models, and carrying out entire workflows within Space. The goal is to move from accessing files to directing work. People can give agents a task, coordinate their tools and compute, and review the results in one place. **About Jason Zhao** Co-Founder, Space Jason Zhao is a co-founder of Space, the San Francisco startup building a streaming filesystem. A filmmaker and YouTuber who has documented his life cinematically for a decade, he ran into the limits of local storage managing terabytes of footage, the problem Space was built to solve. **About Matthew Ao** Co-Founder, Space Matthew Ao is a co-founder of Space, the San Francisco startup building a streaming filesystem. He previously co-founded Kortex, a note-taking app built with the writer Dan Koe, which was later rebuilt as Eden, an AI workspace for creatives. **About Arihant Bapna** Co-Founder, Space Arihant Bapna is a co-founder of Space, the San Francisco startup building a streaming filesystem. He was previously CTO of Kortex and a co-founder of Eden, and studied computer science at the University of Toronto. **About Space** Space (Space Computer, Inc.) is a San Francisco startup building an AI-native distributed filesystem that makes cloud-hosted files behave like local ones. Files appear in Finder and open in native applications while Space streams only the byte ranges an application or AI agent requests, using no local disk space. Founded by Matthew Ao, Arihant Bapna and Jason Zhao, the company raised a $2.4 million pre-seed round led by a16z Speedrun in August 2026. [Website](https://spacefs.com)