InsForge raises $8M seed
for InstaCloud —
the agent-native cloud taking on AWS.

On September 29 2026, Hang Huang — co-founder and CEO of InsForge — announced on X that the company had raised an $8 million seed round and launched InstaCloud, their agent-native serverless cloud platform. The framing was direct: they are building the alternative to AWS, Google Cloud, and Azure. Not a feature on top of them — an alternative.

InsForge is a Y Combinator Spring 2026 company. Six employees. Headquartered in San Francisco. Founded in 2025 by Hang Huang and Tony Chang — Huang on the product and business side, Chang as CTO coming out of Databricks network infrastructure engineering.

The numbers behind the pitch

The $8M seed announcement came with public metrics that give the claim some weight. These are not projections — they are published numbers from the InsForge team.

11,000 GitHub stars in three months — five times growth — is developer interest that cannot be manufactured. 33,000 projects created on a platform that has existed for roughly a year means people are building with it, not just starring it. The weekly active project count doubling is the more interesting number because it suggests retention, not just acquisition.

Where the money came from and where it went before

InsForge did not appear at the $8M seed stage fully formed. There is a clear progression from the initial open-source launch through the cloud pivot to the current raise.

The pivot that led here

The InsForge founders have been unusually transparent about the process. In a post titled "Focus or Die" published on their own blog, Hang Huang described the moment the team lost direction: they had launched the open-source version, hit 500 stars quickly, and then tried to build everything at once. A planned August 31 launch of their cloud version became September 14. They were 30% as efficient as before and slowing down.

The response was to strip back. Pick one thing. Build it. The cloud version was the thing. InsForge Cloud followed, and the progression from that early hosted product to the current InstaCloud agent-native platform is the arc that the $8M seed is now funding.

The original problem they were solving for themselves is what makes the product credible. Coding agents could write application code. The backend services those agents needed — databases, storage, compute — were still built for humans to manage. Dashboards. Manual steps. UIs that agents cannot navigate efficiently. InsForge built the CLI-first interface that lets agents do the infrastructure work without a human in the loop for every provisioning decision.

The competitive claim

Calling your six-person startup the thing that kills AWS is the kind of statement that invites scepticism. It is worth separating the ambition from the current reality.

AWS, Google Cloud, and Microsoft Azure were built for humans — human operators, human DevOps teams, human administrators. That was the right design for the last twenty years of software development. The InsForge argument is that as AI coding agents write more and more of the application code, the infrastructure layer needs to be redesigned from the ground up for agents as the primary operators, not humans.

That argument is not obviously wrong. Supabase — their most direct comparable in terms of developer-facing backend tooling — has raised $544 million and is considered a mature product. InsForge is positioned adjacent to that category but with the agent-first design as the differentiator. The 11,000 GitHub stars and 33,000 projects suggest there is a real audience for the bet they are making.

What is missing from the public record: customer names, production metrics beyond the GitHub and project counts, regional availability beyond US East, and the specific investors in the $8M seed round beyond Pioneer Fund. Those are the things to watch for in subsequent announcements.