Christian Pickett and Bera Sogut had little more than a working prototype when they caught the attention of Pantera Capital last winter. Eight months later—a blink in startup time—they'd closed a $4.3 million seed round for Orthogonal, a San Francisco company building what it calls the payment infrastructure for autonomous AI agents.
The timeline is almost comically compressed. November 2025: first code. January 2026: Y Combinator acceptance. March: Demo Day in front of 190-196 other startups. June 25: funding announcement, with more than 35 APIs already live on the platform.
Whether that velocity signals genuine market pull or just another frothy moment in the AI hype cycle is harder to say. But Pantera—a firm not known for whimsy—clearly saw something worth backing. So did Y Combinator, Pioneer Fund, Decasonic, Blast, Outbound, Surreal, and a cluster of strategic investors who joined the round.
The Bet on Agents Who Pay
Orthogonal's pitch rests on a simple premise: AI agents, as they grow more autonomous, will need to transact with APIs and third-party services on their own. Current payment rails—built for humans clicking checkout buttons—aren't designed for that world.
The company's solution is an SDK, MCP server, and REST API that let agents discover, orchestrate, and pay for verified services through a single integration. Think of it as Stripe for bots, though the analogy isn't perfect. Orthogonal prices on a pay-per-call model using proprietary credits, with costs starting at $0.01 per transaction. The platform also supports x402 on-chain micropayments and multi-party payment splitting, a nod to its crypto-native DNA.
That Pantera led the round is no accident. The venture firm has spent years funding blockchain infrastructure, and Orthogonal's on-chain capabilities likely mattered as much as—maybe more than—its orchestration layer. It's a bet that autonomous agents and crypto payments will converge, perhaps sooner than the broader market expects.
Pedigrees and Prototypes
Pickett and Sogut met at McGill University, though their résumés suggest they were destined to cross paths eventually. Pickett handled payments at Coinbase and billing systems at Vercel—unglamorous but essential work. Sogut's background skews more technical: reCAPTCHA Mobile and Maps APIs at Google, robotics at Amazon. He's also a two-time ACM ICPC World Finalist, which is the sort of credential that impresses investors even if they're not entirely sure what it means.
According to a LinkedIn post from someone who encountered the pair early on, they'd "just started building Orthogonal" in November 2025. Two months later, they published their Launch YC post. The Winter 2026 batch wrapped Demo Day on March 24, and by then, the founders were already in conversations that would lead to the Pantera round.
Traction, or the Appearance of It
The Y Combinator network provided Orthogonal's early believers, though it's unclear if they were paying customers or testers. The Launch YC post name-checked Precip (W24), Riveter (F24), Andi (W22), Shofo (W26), Fiber AI (S23), and Sixtyfour (X25), among others. Whether these companies are paying customers or just friendly early testers isn't entirely clear from public disclosures.
What's more concrete: the platform now lists APIs across search, web scraping, datasets, go-to-market tools, data enrichment, identity verification, KYC, and AI inference. Publicly available providers include Apollo, People Data Labs, Coresignal, Crustdata, ContactOut, Tomba, Serper, and ScrapeGraphAI. It's a respectable roster, though hardly exhaustive.
In the weeks after the funding news, Orthogonal published partner spotlights for You.com (June 19), Icypeas (June 26), and Findymail (July 6). The cadence suggests momentum, or at least a well-oiled marketing operation.
Infrastructure Before Demand
There's a particular kind of risk in building infrastructure for a future that hasn't quite arrived yet. Orthogonal is banking on the agent economy materializing at scale—and soon. That's the same thesis driving dozens of other startups, many of them also funded by crypto-adjacent investors convinced that autonomous systems will need native payment layers.
The counterargument is obvious: what if the agent economy takes longer to mature than expected? What if APIs and services decide to build their own payment integrations rather than relying on a third-party layer? Or what if—and this is perhaps the more uncomfortable question—AI agents don't actually need the kind of transactional complexity Orthogonal is building?
Pickett and Sogut seem unbothered by such doubts. Their approach has been to ship early, integrate fast, and worry about product-market fit as it reveals itself. It's a strategy that works until it doesn't.
For now, though, the bet is paying off. A $4.3 million seed round buys runway, credibility, and the luxury of time—at least by startup standards. Whether seven months was enough to build something lasting, or just enough to build something fundable, is a question that will take longer to answer.
