Micro1’s rapid growth over the past two years has pushed it into the ranks of artificial intelligence corporations that are growing at breakneck speed. The three-year-old startup, which helps artificial intelligence labs recruit and manage training data experts, began the 12 months with about $7 million in annual recurring revenue (ARR).
Today, it claims to have surpassed $100 million in ARR, founder and CEO Ali Ansari told TechCrunch. The figure is also greater than double the revenue Micro1 reported in September, when it announced a $35 million Series A at a $500 million valuation.
Ansari, 24, said at the time that Micro1 works with leading artificial intelligence labs, including Microsoft, in addition to Fortune 100 corporations that are racing to enhance large language models through post-training and reinforcement learning. Their demand for top-notch human data is driving a rapidly growing market that Ansari expects to grow from $10 billion to $15 billion today to just about $100 billion inside two years.
The growth of Micro1 and larger competitors like Mercor and Surge accelerated after OpenAI and Google DeepMind has reportedly severed ties with Scale AI following Meta’s $14 billion investment in the vendor and the decision to rent Scale’s CEO.
While Micro1 is growing rapidly, in keeping with ARR’s founder, it has yet to catch up with its rivals: Mercor is value greater than $450 million, sources tell TechCrunch, and Surge reported $1.2 billion in 2024.
Ansari attributes Micro1’s growth to its ability to quickly recruit and evaluate domain experts. Like Mercor, Micro1 began as an AI recruiter called Zara, combining engineering talent with software roles before entering the data training market. The tool is currently interviewing and vetting candidates looking for expert positions on the platform.
In addition to providing expert-level data to leading AI labs, Ansari says two recent segments, barely visible today, are on their option to changing the economics of human data.
Techcrunch event
San Francisco
|
October 13-15, 2026
The first is for non-AI Fortune 1000 corporations to start building AI agents for internal workflows, support operations, finance, and industry-specific tasks.
Developing these agents requires systematic evaluation: testing pioneering models, evaluating their performance, choosing winners, tuning them, and repeatedly checking performance in a production environment. Ansari argues that this cycle depends largely on human experts assessing AI behavior at scale.
The second is introductory robotics training, which requires high-quality demonstrations of human-performed on a regular basis physical tasks. Micro1 is already building what Ansari calls the world’s largest robotics pre-training dataset, collecting demonstrations from lots of of execs recording object interactions in their homes. Robotics corporations will need huge amounts of this data before their systems can operate reliably in homes and offices, he said.
“We anticipate that a significant portion of product budgets in non-AI enterprises will be devoted to human judgment and data, moving from 0% to at least 25% of product budgets,” said the CEO, who founded Micro1 while at the University of California, Berkeley. “We also help robotics labs create robotics data; these two areas will be a huge share of the $100 billion-a-year market.”
Even as recent markets emerge, Micro1’s current growth continues to primarily come from elite AI labs and AI-heavy enterprises. The startup is expanding its work by using labs dedicated to reinforcement learning, which is a feedback loop for testing and improving model behavior.
Micro1 hopes that an early move into robotics and enterprise agent development, in addition to scaling specialized RL environments, will help it gain additional market share as the data wars intensify.
Ansari says the company’s current focus is on scaling responsibly, paying experts well and keeping people at the center of an industry built on training machines.
The company currently manages 1000’s of experts in lots of of fields, from highly technical fields to surprisingly offline disciplines. According to Ansari, many earn near $100 an hour.
“There are Harvard professors and Stanford graduate students spending half the week training artificial intelligence with Micro1,” Ansari said. “But the bigger change is in the sheer number and scope of roles. It’s expanding into areas you wouldn’t expect to be important in language model training, including offline and less technical areas. We’re very optimistic about where this is headed.”