AI agents are moving from an ambitious idea toward a crowded technology race. Startups and major technology companies are trying to build programs that can complete multi-step tasks without constant user direction, but /dev/agents says the missing layer is not just another agent. It is the operating system those agents can run on.
The new company, founded by former Google executives with deep Android experience, has raised a $56 million seed round and was valued at $500 million, according to a person with knowledge of the investment. Its central claim is simple: AI agents may need the kind of shared platform that helped mobile software become easier to build.
Why /dev/agents sees a platform gap
The company’s thesis starts with a practical developer problem. Many teams are working on AI agents, but /dev/agents argues that builders do not yet have standard tools and systems for creating reliable agent software. Without that common foundation, the promise of AI may remain harder to turn into useful products.
David Singleton, co-founder and CEO of /dev/agents, compares the situation to the early smartphone era. Before Android, mobile computing looked powerful, but building for it was difficult. His view is that AI is at a similar stage: the opportunity is visible, while the development environment remains fragmented and immature.
That is why /dev/agents is working on what it describes as a unified platform for the AI world. The company says its cloud-based operating system will work across devices and use generative AI to present personalized user interfaces.
The platform is aimed at a future in which AI agents do not simply answer requests, but coordinate with one another. The source example is a complex travel task, where agents might need to collaborate to book and pay for flights, hotels, and excursions.
The funding reflects confidence in the team
The seed round was led by Index Ventures and co-led by Alphabet’s independent growth fund, CapitalG. Conviction Capital, the venture firm founded by Sarah Guo, also participated.
Several prominent technology leaders joined the round as well, including OpenAI co-founder Andrej Karpathy, Scale AI CEO Alexandr Wang, Palo Alto Networks CEO Nikesh Arora, and Android founder Andy Rubin.
The size of the round is notable because /dev/agents is still early. CapitalG partner Jill Chase said it is exceedingly rare for the growth-stage fund to invest in a pre-product company. In this case, she pointed to the size of the market opportunity and the strength of the team.
Index Ventures partner Nina Achadjian also framed the company’s ambition around difficulty and scale. She said the problem is technically hard and the idea is large, while arguing that a team with experience building Android and Stripe from the ground up is unusually well suited to attempt it.
The founders bring Android, Stripe, Meta, Chrome, Dropbox, and Figma experience
Singleton was previously CTO of Stripe and, before that, led Android Wear at Google. The company’s founding group also includes Hugo Barra as CPO, Ficus Kirkpatrick as CTO, and Nicholas Jitkoff.
Barra was previously VP of Android’s product management at Google and led Meta’s Oculus VR division. Kirkpatrick was an early Android engineer and later a former VP of AR/VR at Meta. Jitkoff worked as a principal designer on Google Chrome and held senior roles at Dropbox and Figma.
That background matters because /dev/agents is not presenting the AI agent problem as a single-app challenge. It is presenting it as an ecosystem challenge, where developers, interfaces, devices, and commerce may all need to work through a common layer.
- Product focus: a cloud-based operating system for AI agents.
- Developer problem: a lack of standard tools and systems for building agents.
- User experience: personalized interfaces generated with generative AI.
- Market belief: fully functional AI agents are coming, but the infrastructure is not ready yet.
How the business could work
/dev/agents expects the first version of its product to be available by early-to-mid next year. The company has not yet reached the stage where its model is fully proven, but Singleton said its business model likely will not be too different from how Android monetizes its operating system now.
He suggested that commerce could become a major part of the platform. In that scenario, /dev/agents could take a cut of sales or charge users for subscriptions.
That business idea follows from the kinds of tasks AI agents are expected to handle. If agents are able to manage bookings, payments, and other transactions across services, then the platform beneath them could become a place where meaningful commerce happens.
The wager behind /dev/agents is that AI agents will need more than powerful models. They will need a shared operating environment, developer tools, and interfaces that can adapt across devices. Investors are betting that this platform layer could become one of the key pieces of AI infrastructure.