FluidCloud’s core product is the Large Infrastructure Model, an AI engine designed to generate, translate and validate Terraform configurations across cloud providers. The system uses a hybrid architecture: standard large language models parse user intent at the front end while custom conditional and foundation models, trained on synthetic Terraform data, perform Terraform generation and cloud-to-cloud conversions. The model scores 0.58 on BLEU benchmarks—approaching a cited human-level score of about 0.60—and supports translation for more than 150 resource types versus the platform’s prior coverage of roughly 25–30 types. Inputs now include existing GitHub repositories with Terraform code (multiple syntax styles and custom mapping overrides are supported), and the product includes features such as compatibility scoring, failure prediction based on provider release cycles and network/upgrade data, and an optimization layer that trades off cost, security and performance. The platform ships with 1,800 compliance policies covering major hyperscalers and emerging providers like Vultr, OVH and Hetzner. Future product plans described include an agent/agent-builder for custom infrastructure workflows via MCP and portable SDKs that abstract cloud provider APIs to simplify switching deployments.

Total raised

$8M

Funding rounds

1

Latest round

Seed

Latest activity

Mar 2026