Micro1, a rapidly ascending AI startup specializing in human-powered training data, has announced it has surpassed $100 million in annual recurring revenue (ARR). This significant milestone, revealed by founder and CEO Ali Ansari to TechCrunch, marks an explosive growth trajectory, more than doubling the revenue reported in September 2025 when the company secured a 35 million Series A funding round at a $500 million valuation. The three-year-old company, which began 2025 with approximately $7 million ARR, is quickly emerging as a formidable competitor to industry giants like Scale AI.

Rapid Expansion in the AI Data Market

Micro1's accelerated growth over the past two years places it among a select group of AI companies experiencing breakneck scaling. The startup's core business involves recruiting and managing human experts for AI labs, providing crucial training data that fuels the development of advanced artificial intelligence models.

The company's rise, alongside competitors such as Mercor and Surge, gained momentum after reports surfaced that OpenAI and Google DeepMind reportedly severed ties with Scale AI. This followed Meta's $14 billion investment in Scale AI and its subsequent decision to hire Scale AI's CEO.

Despite its impressive growth, Micro1's ARR has not yet reached the levels of its larger rivals. Mercor has reportedly achieved over $450 million in ARR, while Surge's 2024 figures were reported at $1.2 billion.

Fueling AI Innovation with Human Expertise

Ansari, 24, attributes Micro1's success to its efficient process for recruiting and evaluating domain experts. The company, which initially started as an AI recruiter named Zara—matching engineering talent with software roles—pivoted into the data-training market. Its proprietary tool now streamlines the interviewing and vetting of applicants seeking expert roles on its platform.

Micro1 currently partners with leading AI labs, including Microsoft, and Fortune 100 companies that are actively enhancing large language models through post-training and reinforcement learning. The demand for high-quality human data in this sector has created a rapidly expanding market, which Ansari projects will grow from an estimated $10-15 billion today to nearly $100 billion within the next two years.

Pioneering New Frontiers: Enterprise AI and Robotics

Beyond its current work with elite AI labs, Ansari identifies two emerging segments poised to significantly reshape the economics of human data:

Enterprise AI Agents

  • Non-AI-native Fortune 1000 enterprises are beginning to develop AI agents for internal workflows, support operations, finance, and industry-specific tasks.
  • This development necessitates systematic evaluation: testing frontier models, grading their output, selecting optimal solutions, fine-tuning them, and continuously validating performance in production.
  • Ansari emphasizes that this entire cycle heavily relies on human experts to evaluate AI behavior at scale.

Robotics Pre-training

  • This area demands high-quality, human-generated demonstrations of everyday physical tasks.
  • Micro1 is actively building what Ansari describes as the world's largest robotics pre-training dataset, collecting demonstrations from hundreds of generalists recording object interactions in their homes.
  • Robotics companies will require vast volumes of this data to enable their systems to reliably operate in diverse environments like homes and offices.

“We anticipate that a good portion of the product budgets at non-AI-native enterprises will go towards evals and human data, moving from 0% to at least 25% of product budgets,” said Ansari, who founded Micro1 while at UC Berkeley. “We’re also helping robotics labs create robotics data; these two areas will account for a massive share of that $100 billion-a-year market.”

While new markets emerge, Micro1's current growth remains primarily driven by elite AI labs and AI-heavy enterprises. The startup is intensifying its collaboration with these labs on reinforcement learning, a critical feedback loop for testing and improving model behavior.

Silicon Valley bets big on ‘environments’ to train AI agents

Micro1 aims to capture additional market share as the "data wars" intensify, leveraging its early entry into robotics data and enterprise agent development, alongside scaling its specialized reinforcement learning environments.

Responsible Growth and Human-Centric Approach

Ansari states that Micro1 is committed to scaling responsibly, ensuring experts are well-compensated, and keeping human involvement central to an industry built on training machines. The company currently manages thousands of experts across hundreds of domains, ranging from highly technical fields to less conventional, offline disciplines. Many of these experts earn close to $100 an hour.

“There are Harvard professors and Stanford PhDs spending half their week training AI through Micro1,” Ansari noted. “But the bigger shift is in the sheer volume and range of roles. It’s expanding into areas you wouldn’t expect to matter for language model training, including offline and less technical fields. We’re very optimistic about where this is heading.”