Archil provides a cloud‑native, POSIX‑compatible filesystem that unifies data storage and compute for AI workloads, mounting directly on servers and offering serverless execution containers that run commands against the disk without separate sandboxes. Its NVMe‑backed distributed cache delivers sub‑millisecond read latency and read‑after‑write consistency, while writes are replicated and asynchronously flushed to object stores such as S3, GCS, or Azure Blob. The platform supports GPU clusters, dev notebooks, CI/CD pipelines, and agent sandboxes, enabling AI developers to use existing code and tools without modification.
Funding
$6.7M raised to dateRaised to date based on public sources. This may differ from the amount the company actually raised and is based only on what is publicly available on the internet.
8OFounders
Product
Problem
AI agents and machine‑learning workloads often require fast, persistent storage that can be accessed directly from code, but existing cloud storage services are separate from compute and involve latency, complex mounting, and limited POSIX compatibility. This fragmentation makes it difficult to build scalable, serverless pipelines where agents can read and write data as if it were a local filesystem.
Solution
Archil offers a cloud‑native, POSIX‑compatible filesystem that unifies data storage and compute for AI workloads. The system mounts directly on servers and provides serverless execution containers that can run arbitrary commands against the mounted disk without provisioning separate sandboxes. Data is cached on NVMe‑backed nodes, delivering sub‑millisecond read latency and read‑after‑write consistency, while writes are replicated and asynchronously flushed to the underlying object store (S3, GCS, Azure Blob). Because the filesystem presents a standard file interface, existing code and tools run unchanged, eliminating the need for SDKs or code rewrites. The platform also supports GPU clusters, dev notebooks, CI/CD pipelines, and agent sandboxes, enabling developers to focus on model logic rather than infrastructure.
Target Audience
Primary customers are AI developers, machine‑learning engineers, and data‑science teams building serverless or containerized pipelines, as well as cloud infrastructure teams that need a high‑performance, POSIX‑compatible storage layer for compute‑intensive workloads.
Features
- POSIX‑compatible mount that works with any application without code changes or client libraries
- Serverless execution model where each command runs in an isolated container directly on the mounted disk
- NVMe‑backed distributed cache providing sub‑millisecond read latency and read‑after‑write consistency
- Automatic replication of writes across cache nodes with asynchronous flush to Amazon S3, Google Cloud Storage, or Azure Blob
- Seamless integration with GPU clusters, dev notebooks, and CI/CD pipelines for AI and data‑intensive workloads
- Unified view of any cloud object store as a single filesystem, simplifying data access across multiple providers
- Built‑in support for agent sandboxes, allowing AI agents to maintain persistent context and state