AI agents are about to get their own credit cards. Sapiom, a startup founded by former Shopify payments director Ilan Zerbib, just closed a $15 million seed round led by Accel to build the financial infrastructure that lets AI agents autonomously purchase software, APIs, and compute resources. The timing couldn't be better - as vibe-coding tools like Lovable turn plain-language prompts into working apps, there's no easy way for those AI-built applications to handle the authentication and micro-payments required every time they ping an external service.
Sapiom is tackling one of the messiest problems in the AI agent revolution - how do autonomous systems actually pay for things? The answer matters more than you'd think. Every time an AI agent needs to send a text via Twilio, spin up a server on AWS, or pull data from an API, there's a payment lurking underneath. Right now, humans have to manually set up accounts, enter credit card details, and copy API keys. It's a workflow that completely breaks the promise of autonomous AI.
That's the infrastructure gap Zerbib spotted after spending five years running payments engineering at Shopify. He launched Sapiom last summer with a focused thesis - build the financial layer that enterprises actually need, not the flashy consumer wallet that VCs keep pushing. The bet paid off. Accel just led a $15 million seed round, joined by Okta Ventures, Array Ventures, Menlo Ventures, Anthropic, and Coinbase Ventures, according to TechCrunch.
"If you really think about it, every API call is a payment," Amit Kumar, partner at Accel, told TechCrunch. "Every time you send a text message, it's a payment. Every time you spin up a server for AWS, it's a payment." Kumar screened dozens of AI payment startups before backing Sapiom, convinced that Zerbib's enterprise-first approach is what makes autonomous agents viable at scale.
The immediate use case is vibe-coding platforms - tools that let non-developers describe an app in plain English and watch AI spit out working code. Anyone who's tried building with Lovable or Bolt knows the pain point. You can prototype a beautiful SMS-enabled app in minutes, but actually connecting it to Twilio requires signing up for an account, adding payment details, and wrestling with API documentation. Sapiom erases that friction entirely. The AI agent handles everything in the background, and users get billed as a pass-through fee by their vibe-coding platform.
It's the same pattern playing out across enterprise software. AI agents need to autonomously access services, but the payment and authentication layer has been an afterthought. Zerbib's building the pipes that make it work - think Stripe for AI agents, but handling both the financial transaction and the identity verification required to access external tools securely.
The market timing is sharp. Non-technical creators are discovering they can build custom apps without writing code, but they're hitting infrastructure walls the moment they try to scale beyond prototypes. Sapiom is positioning itself as the invisible layer that makes production-ready AI apps possible, targeting vibe-coding companies and enterprises deploying AI agents for internal workflows.
Kumar's conviction comes from watching the space evolve. "In the future, apps are going to consume services which require payments. Right now, there's no easy way for agents to actually access all of that," he said in the TechCrunch interview. The insight that separated Sapiom from the pack - focusing on B2B infrastructure instead of chasing consumer AI wallet dreams.
Zerbib isn't ruling out consumer applications eventually. Personal AI agents could theoretically handle transactions like ordering Uber rides or making Amazon purchases. But he's skeptical that AI will suddenly make people spend more. "AI won't magically make people buy more things," he believes, which is why the near-term focus stays locked on enterprise infrastructure where the pain point is operational efficiency, not consumer behavior change.
The $15 million seed is notably large, reflecting both the technical complexity and the massive addressable market. Building secure payment infrastructure that handles authentication, compliance, and micro-transactions across thousands of APIs isn't trivial. The investor lineup adds strategic weight - Anthropic brings AI agent expertise, Coinbase Ventures understands digital payment rails, and Okta Ventures knows enterprise identity management.
For now, Sapiom remains in early stages, but the thesis is clear. As AI agents proliferate across enterprises and vibe-coding platforms democratize app development, someone needs to build the financial plumbing. Zerbib's placing a bet that the winner won't be whoever builds the flashiest consumer product, but whoever solves the boring, critical infrastructure problem that makes autonomous AI transactions actually work.
The AI agent economy has a payments problem, and Sapiom is building the solution. With $15 million in fresh capital and a former Shopify payments veteran at the helm, the startup is positioning itself as the critical financial layer between autonomous AI systems and the software services they need to access. The question isn't whether AI agents will need payment infrastructure - they already do. The question is whether Zerbib's enterprise-first approach captures the market before someone else figures it out. For vibe-coding platforms and enterprises deploying autonomous agents, that infrastructure can't come fast enough.
