A decade ago, cloud infrastructure was the invisible scaffolding that allowed the internet’s biggest companies to scale. Today, training data is playing a similar role for artificial intelligence. The providers who can reliably supply it — much like AWS, Azure, and Google Cloud did for the internet — are quickly becoming the indispensable backbone of the industry.
Into this arena steps Micro1, a three-year-old startup that has just raised a $35 million Series A round at a $500 million valuation, betting it can be to AI data what cloud was to software.
The round was led by 01 Advisors, the venture capital firm founded by former Twitter executives Dick Costolo and Adam Bain. As part of the deal, Bain is also joining Micro1’s board alongside Joshua Browder, founder of AI legal assistant DoNotPay.
Register for Tekedia Mini-MBA edition 19 (Feb 9 – May 2, 2026): big discounts for early bird.
Tekedia AI in Business Masterclass opens registrations.
Join Tekedia Capital Syndicate and co-invest in great global startups.
Register for Tekedia AI Lab: From Technical Design to Deployment (next edition begins Jan 24 2026).
A Market Shaken by Scale AI’s Fallout
The data labeling market has been roiled since Meta invested $14 billion in Scale AI and hired its CEO. Concerned about their research pipelines being too closely tied to Meta, companies like OpenAI and Google began cutting ties with Scale AI. The company insists it does not share client data with Meta, but the perception of risk left AI labs searching for alternatives.
That opened the door for competitors such as Mercor, Surge, and now Micro1. And unlike Scale AI, which pioneered paying relatively low wages for global contractor labor, Micro1 says it’s focused on the next frontier: high-quality domain expertise.
Growth Against Bigger Rivals
Despite being only three years old, Micro1 is scaling quickly. Its 24-year-old CEO, Ali Ansari, told TechCrunch that the startup’s annual recurring revenue has surged from $7 million at the start of 2025 to $50 million now. Its clients include Microsoft and several Fortune 100 companies.
Still, the numbers pale compared to its larger rivals. Mercor generates $450 million in ARR, while Surge earned $1.2 billion in 2024. Yet the trajectory suggests Micro1 is carving out a spot in a market that may ultimately resemble the cloud sector — a handful of giants supporting countless smaller but essential providers.
Ansari argues that AI labs no longer just want bulk labeling at the cheapest possible cost. To make models smarter, they need training data from senior engineers, doctors, professional writers, and even university professors. To meet that demand, Micro1 built Zara, an AI-powered recruiter that interviews and vets candidates. The system has already onboarded thousands of experts, including faculty from Stanford and Harvard, and Micro1 says it plans to add hundreds more every week.
“Really the only way models are now learning is through net new human data. Micro1 is at the core of providing that data to all frontier labs, while moving at speeds I’ve never seen before,” Bain said.
Micro1 isn’t stopping at labeling. The startup is moving into AI training environments — simulated workspaces where AI agents can practice real-world tasks. Analysts say this could become as important as cloud servers were for the early internet, because it allows companies to scale model training without being constrained by static datasets.
Just as no single cloud provider controls the entire internet, no single data provider can meet all of an AI lab’s needs. OpenAI, Anthropic, Meta, and Google all work with multiple partners, and likely will continue to do so. That fragmentation, combined with the sheer demand for data, means there’s room for startups like Micro1 to grow even as bigger competitors dominate.
A Cloud-Like Trajectory
The parallel to cloud computing is striking. A decade ago, AWS, Azure, and Google Cloud turned server capacity into an essential service, giving software startups the foundation to grow without building infrastructure from scratch. Now, companies like Micro1, Surge, and Mercor are doing the same with training data, offering AI labs a way to scale without building vast labeling operations in-house.
The difference is that while cloud became highly consolidated, the training data space remains messy and fragmented. Analysts suggest it could eventually consolidate in a similar fashion — a few giants handling the bulk of the market, while smaller firms survive by specializing in niche domains.
For now, though, the prize is still wide open. With a fresh $35 million in its pocket and a board stacked with high-profile backers, Micro1 is betting it can turn its early momentum into long-term staying power — just as AWS once did in a very different but equally foundational market.