RWX Raises $12M Series A for Agentic Software Infrastructure

RWX announced a $12M Series A led by Hyde Park Venture Partners on August 4, 2026. Quiet, The OHIO Fund, R1, and DV participated, alongside angels Kohsuke Kawaguchi, Bryan Johnson, and Theo Browne. The round follows a previously disclosed $7M seed led by Quiet, bringing RWX's announced funding to at least $19M.

The Columbus, Ohio company is building cloud infrastructure where engineers and AI coding agents can build, test, and validate software. RWX combines CI/CD, agent sandboxes, preview apps, container image builds, and test-suite management in a platform designed around faster, more autonomous development loops.

The larger signal is not that AI can write code. Everyone with a demo and a prompt box has been yelling that for a while. The more important shift is that code generation is getting faster than the systems used to verify what the code actually does. RWX is raising capital to own that widening gap.

What Happened

According to the official Series A announcement, Hyde Park Venture Partners led RWX's $12M financing, with Quiet, The OHIO Fund, R1, and DV participating. Quiet's return matters because it led RWX's $7M seed in 2022. A new investor can buy the story in a meeting. A returning investor has already watched the company operate between meetings.

RWX was founded in 2022 by Dan Manges, Founder and CEO, and Tommy Graves, Founder and CTO. Manges previously served as CTO at Braintree and Root. Graves previously worked as a Principal Engineer at Huckleberry and a Senior Engineering Manager at Root. RWX's current leadership page also lists Clara Kridler as COO, bringing operating experience from Hologram, Root, and KoBold Metals.

The company did not disclose its valuation, board changes, revenue, customer count, or a precise use-of-funds plan. Those blanks matter because funding stories get sloppy when the check becomes permission to invent the rest of the spreadsheet. What RWX did disclose is clear: the company is financing a broader push to rebuild software validation for an era when coding agents can generate and modify code at high speed.

Why Agentic Software Changes CI/CD

Traditional CI/CD systems grew up around a human cadence. An engineer wrote code, committed it, pushed it, and waited for a pipeline to run. The process was inefficient, but the delay could hide inside a larger block of human work. Coding agents compress the creation phase, which makes every redundant build step and flaky test suddenly much louder.

RWX's thesis is that the software forge has to change with the worker. Its interface is CLI-first, and code execution is decoupled from version-control webhooks. Agents can test code in sandboxes, run a full CI pipeline, and deploy a preview application without first creating a commit and pushing it to a remote repository. That is not a cosmetic workflow change. It pulls validation closer to the moment code is generated.

RWX supports that loop with an OCI-compatible container runtime, graph-based task execution, and content-based caching. The graph allows independent work to run in parallel. The cache reuses results when relevant files and task definitions have not changed. The platform also provides semantic outputs, granular retries, observability, and tools for detecting and quarantining flaky tests. In plain English, RWX is trying to make verification fast enough that agents can iterate without turning infrastructure waste into a business model.

The Evidence Behind the Pitch

RWX's strongest public proof comes from customer case studies, so the numbers require attribution rather than confetti. In an RWX-published Honeycomb case study, Honeycomb reported a 45% faster CI feedback loop, P95 build times falling from more than 13 minutes to 7 minutes, and a 50% reduction in failures after adopting RWX. Those results are not an independent benchmark for every engineering team, but they show the operational problem investors are underwriting.

The issue is not simply speed. A coding agent that produces more changes also creates more demand for tests, previews, builds, policy checks, and human review. If validation infrastructure fails under that volume, the organization has traded a writing constraint for a trust constraint. The code arrives sooner, but nobody can responsibly ship it sooner.

This is why RWX sits in a useful part of the AI infrastructure stack. Model vendors compete on intelligence. Coding products compete on the quality of generated changes. RWX is focused on the execution and verification layer that determines whether those changes survive contact with a real codebase. That layer may be less glamorous than a cinematic agent demo, but engineering budgets have a long history of finding whatever stops production from breaking.

Why Hyde Park Venture Partners Fits

Hyde Park Venture Partners describes its focus as early-stage AI, software, and technology-enabled services companies connected to the Midwest, Toronto, and the Southeast. RWX is a Columbus software-infrastructure company led by operators with experience at Braintree, Root, Huckleberry, and other high-growth businesses. The geographic and technical fit is unusually direct.

The round also adds to a broader Midwest infrastructure story. In April 2026, RWX became the first Scaleup Suite tenant at Rev1's downtown Columbus innovation hub. That does not guarantee the next developer-tools giant will come from Ohio. It does show that the company is pairing venture capital with a physical operating base in a market that knows how to build software without requiring a zip code approved by a coastal group chat.

What This Signals

RWX's Series A is a bet that the next software-development constraint will be confidence. Teams will not need fewer tests because agents write more code. They will need faster, more observable, and more reproducible ways to prove that a larger volume of code is safe to merge and deploy.

For engineering leaders, the practical question is no longer whether developers will use coding agents. It is whether the surrounding systems can absorb the output without creating longer queues, higher compute bills, and more failures. RWX is selling an answer built around graph execution, caching, sandboxes, previews, and a CLI that agents can use directly.

For investors, the appeal is equally concrete. The application layer of AI changes quickly, but every application still depends on infrastructure that can execute work and verify results. RWX does not need every coding agent to look the same. It needs the volume and speed of software changes to keep rising. The $12M Series A gives the company more room to prove that read, write, and execute is becoming build, test, and trust.