Behind Q3's Record AI Megadeals, Investors Are Also Funding the Plumbing
On October 5, a Singapore-based startup called OneByZero closed a modest but telling round: $20 million in Series A funding, its first outside capital, led by Jungle Ventures. The company doesn't build a frontier model or sell GPU capacity. It does something narrower and, for enterprise buyers, more immediately useful: it places engineers inside large financial-services, telecom and retail companies to deploy and govern AI "agents" that work alongside human staff. CEO and co-founder Niket Vaidya framed the thesis plainly, saying the value of AI "comes when it runs inside the systems that run the business, under the company's own rules, and gets better every week."
A Record Quarter for Megadeals
OneByZero's raise landed the same week Crunchbase published its tally of global venture activity for the third quarter of 2026, and the contrast is instructive. Crunchbase counted 27 companies that closed billion-dollar-plus rounds during the quarter — a record, up from 16 in Q2 and 14 in Q1 — with those mega-rounds alone accounting for roughly a third of the $159 billion in global venture funding tracked for the period. AI-focused startups broadly captured $102 billion of that total, or 64% of every venture dollar deployed worldwide, even as the share cooled slightly from earlier in the year. U.S.-based companies pulled in $91 billion of it, 57% of the global total, with the San Francisco Bay Area alone accounting for nearly a quarter of worldwide venture investment. Put simply: a shrinking number of companies are absorbing a historically large share of AI's investment dollars.
But Crunchbase's stage-level breakdown tells a second story. Early-stage funding reached $40.6 billion for the quarter, up 25% year-over-year, and seed funding totaled $13 billion, with $2.6 billion of that specifically going to "mega-seed" rounds above $100 million. Nearly 90% of late-stage dollars went into checks of $100 million or more — evidence that growth investors are increasingly willing to write only large checks — yet smaller, more targeted bets at the earliest stages kept finding capital. That's the segment where OneByZero, and a handful of other small AI raises this week, actually sit.
Funding the Plumbing, Not Just the Models
OneByZero's "NEO" platform is built to let enterprises define exactly what an AI agent is allowed to do and where a human has to sign off — a governance layer that Jungle Ventures managing partner Yash Sankrityayan pointed to as the reason the firm backed the company.
OneByZero stood out for having spent three years deploying AI in large, complex organizations with measurable results. — Yash Sankrityayan, Managing Partner, Jungle Ventures
The company says it has doubled revenue annually for three years and now automates more than 90% of some customers' interactions; it plans to use the new capital to expand engineering teams across nine markets and open an office in Japan. It wasn't the only narrow bet to get funded that day. Copenhagen-based Pandektes raised a €13.5 million Series A led by Alstin Capital to build legal data infrastructure specifically for AI research use, and Paris-based Fleuret AI raised a €4 million pre-seed led by RAISE Ventures to apply agentic AI to penetration testing — automating a cybersecurity function that used to require a human red team. Neither will show up near the top of any quarterly funding leaderboard. Both are bets on a specific, narrow job AI can now do reliably, rather than on another general-purpose model.
Crunchbase's report points to the same pattern one layer down, outside software entirely: aerospace, robotics, data centers, semiconductors and energy each separately cleared $10 billion in the quarter, as capital spreads into the physical infrastructure AI depends on, not just the models themselves. It's a reminder that "AI funding" is no longer one trade; it's several, moving at different speeds and funded by different kinds of checks.
What It Means for Investors
The resulting market looks increasingly barbell-shaped. At one end, a small cohort of frontier labs and infrastructure players are pulling in historic sums — enough to make 27 the new record for billion-dollar rounds in a single quarter. At the other, specialist startups solving one deployment, governance or vertical problem are still raising smaller rounds at a steady, even accelerating, clip. For investors trying to read where AI spending is actually headed next, both ends of that barbell matter: the top signals where frontier-model economics are being bet on; the bottom — in enterprise deployment infrastructure, data governance and vertical automation — signals where the technology is actually being put to work inside businesses today.