Anti Metal raises $20M Series A to automate cloud infrastructure as AI makes deployment the new bottleneck
Jun 12, 2025 with Shreyas Iyer
Key Points
- Anti Metal raises $20M Series A led by Sound Ventures to automate cloud infrastructure debugging, targeting the shift where AI-assisted coding makes deployment the new bottleneck for engineering teams.
- The startup learns how each company operates and encodes institutional knowledge into reusable systems rather than dictating infrastructure standards, moving teams from maintenance firefighting to feature building.
- AI workloads are creating new failure modes across cloud infrastructure while cheaper model development spawns more internal tooling, compounding sprawl that Anti Metal positions as a systems-engineering problem hitting far more companies simultaneously.
Summary
Anti Metal raised $20 million in Series A funding led by Sound Ventures. Buckley Ventures, Naval Ravikant, Arjun Vohsi, Drew Houston (Dropbox founder), Arvin Srinivas (Perplexity), and other angels also participated.
Where the bottleneck moved
Writing code is no longer the constraint. Tools like Cursor, Lovable, and v0 have largely solved that problem. The hard part is deploying code, maintaining it, and scaling from one user to a million. Anti Metal connects to a company's full infrastructure surface area—source code, cloud environments, ticketing systems—to diagnose failures and surface fixes.
The product stays deliberately non-opinionated. Rather than dictating what good infrastructure looks like, it learns how each company operates and encodes that institutional knowledge into a reusable system. The goal is shifting teams from asking "can we keep this running?" to "what should we build next?"
AI workloads compound the problem
AI is adding new failure modes on top of existing ones. Models become potential single points of failure. If a cloud goes down, products built on top of it fail too. Tooling for monitoring and evaluating AI systems remains immature. Most companies that succeed at this build their own frameworks instead of using off-the-shelf solutions. Cheaper AI-assisted development means companies ship more internal tooling than ever, which compounds infrastructure sprawl.
These are fundamentally familiar systems-engineering problems, not entirely new ones. They are just hitting a much larger pool of companies simultaneously.
Oracle Cloud Infrastructure
When asked which hyperscaler is underrated, Iyer names Oracle Cloud Infrastructure. OCI offers strong foundational compute without the layers of proprietary services sitting on top. It competes well on price and security even though it lacks the brand appeal of AWS, Azure, or GCP.
Growth and tooling
Anti Metal's growth has been unusually viral for an enterprise infrastructure company. Co-founder Mattly's background in growth and marketing drove much of that momentum. Nearly every engineer has been stuck debugging infrastructure instead of shipping features, which explains why the problem resonates. A larger marketing push is planned for 2025.
The team runs Cursor across the board. Some engineers prefer Claude Code for agentic workflows.