Businesses
Space Raises $2.4M to Build AI-Native Filesystem for Humans and Agents
19 Aug 2026
The a16z Speedrun-backed startup is developing a distributed filesystem that lets applications, teams and AI agents work across cloud-hosted data without relying on local disk storage.
San Francisco, CA — 18 August 2026 — AI infrastructure startup Space has raised $2.4 million in pre-seed funding, led by a16z Speedrun, to build what it describes as an AI-native distributed filesystem for humans and software agents.
The company is targeting a long-standing problem in cloud computing: storage capacity may be effectively unlimited, but access is still often constrained by the device in front of the user.
Space says its system allows cloud-hosted files to behave more like local files. Instead of downloading entire files, its technology streams only the data required by an application or agent at that moment.
Moving Beyond Traditional Cloud Storage
Conventional cloud platforms solved the problem of where files are stored.
Space is focused on what happens next: how people and AI systems actually access those files.
Today, teams often need to download, sync, duplicate or re-upload data before work can begin. AI agents can face an even larger bottleneck because information may need to be copied into a separate tool or indexed before it becomes usable.
Space wants to remove those steps.
Its distributed filesystem sits above the operating system and is designed to make the same live data available to applications, teammates and AI agents through one shared layer.
Built for Data-Heavy Workflows
The technology is aimed at teams working with large files and datasets.
A video editor, for example, could work with a project larger than the computer’s available storage. An architect could access a large building model without downloading all its assets first. Developers could work across large repositories while AI agents access only the parts of files relevant to a task.
The company says this approach could be useful across media, design, software development and AI-heavy workflows.
Space is currently in private beta, with around 100 users and teams onboarded.
a16z Speedrun Leads the Funding Round
The $2.4 million round was led by a16z Speedrun.
Other participants include Golden Ventures, Northside Ventures and several angel investors with experience across software companies including Parsec, Sentry, Stan, Superwhisper and Modem.
Jonathan Lai, General Partner at a16z, said Space is challenging the assumption that a file must exist on a local device before someone can work with it.
The company was founded by Matthew Ao, Arihant Bapna and Jason Zhao.
Why AI Changes the Storage Problem
The timing is closely linked to the rise of AI agents.
Traditional cloud tools were designed primarily around human users and file syncing.
Agentic systems work differently. They may need to inspect small parts of thousands of files, search across shared company data or act on information without moving entire datasets into another environment.
Space believes the next generation of infrastructure needs to make data immediately accessible at the filesystem level rather than through one integration at a time.
That could reduce duplication while also improving how humans and AI systems work from the same source data.
What Comes Next
Space is beginning with sectors where local storage limitations are already a major problem.
Its early focus includes video, marketing and architecture, engineering and construction.
The company plans to expand into areas such as:
- media and entertainment;
- AI training-data infrastructure;
- computer vision;
- enterprise data systems; and
- world-model pipelines.
- Its longer-term ambition goes beyond storage.
Space describes its ultimate vision as the Space Computer: a model where the physical device becomes a window into effectively unlimited storage and compute.
About Space
Space is building a distributed filesystem designed for human creators and AI agents.
Its technology makes cloud-hosted files accessible inside existing applications without requiring full-file downloads, duplicated copies or application-specific integrations.
The company was founded by Matthew Ao, Arihant Bapna and Jason Zhao.
