The Idea:

At 10, Ali Ansari swapped Tehran for a third-grade classroom in California without a word of English. For five years, he didn't fit in. So he hustled. By the end of middle school he was cycling to garage sales, buying textbooks for pennies and flipping them on Amazon at 30 to 40% margins. In high school he built a peer tutoring marketplace and sold it before he graduated.

By 2021 he was a computer science student at UC Berkeley, running a software agency on the side and hiring engineers from around the world for client work. The problem was the hiring. Interviewing candidates was eating 30 to 40 hours of his week. So he built an AI interviewer to do it for him. The first version was rough, stitched together from early neural networks. Then GPT got good enough to power it properly, and Ali watched it run interviews he'd otherwise have sat through himself.

He realised the tool wasn't a shortcut. It was the company.

This was the beginning of micro1.

The Execution:

• Late 2022: In his final year at Berkeley, Ansari spun micro1 out as a product built around Zara, the AI recruiter he'd built as an undergrad. Within months he added a marketplace of pre-vetted engineers, and the marketplace grew faster than the software.
• August 2023: VCs wanted nothing to do with another recruiting startup, but Jason Calacanis backed a $3.3M pre-seed. By August 2024, micro1 had raised $6.6M across its early rounds.
• End of 2024: The bet-the-company moment. A data vendor asked micro1 to hire 700 engineers, not to write code but to train AI models. It became micro1's biggest customer almost overnight. Ansari killed both working business lines, the SaaS tool and the engineering marketplace, to sell straight to AI labs, with no guarantee they'd buy.
• Early 2025: micro1 started the year as a recruiting business doing $7M in ARR.
• Mid 2025: Meta put $14B into Scale AI and hired its CEO. OpenAI and Google said they'd cut ties, leaving a hole in the data market that micro1 was now built to fill.
• September 2025: A $35M Series A at a $500M valuation, led by Dick Costolo and Adam Bain's 01 Advisors. ARR had jumped from $7M to $50M in nine months. Ansari was 24.
• December 2025: micro1 crossed $100M ARR.
• March 2026: $250M+ in annual revenue with 80 employees, 35x growth in a year. Employee number 80 was Andrew Maas, the Stanford professor who ignored Ansari's first job pitch a year earlier.
• August 2026: Gross run rate hit $500M, up from $100M eight months before. After paying its experts, micro1 keeps $150M to $200M of that.
• Today: micro1 has raised $100M+ at a $4B valuation, an 8x jump in a year. It's now chasing robotics data and even lodged a $12.5M bid for bankrupt Spirit Airlines' operational data, trying to outbid Google.

The lesson? micro1 didn't find its market. Its biggest customer walked in and showed it. A request for 700 engineers to train AI looked like a strange one-off order. Ansari read it as a map to where the money was, then torched two working businesses to follow it. Most founders protect what's already working. The best ones notice when a customer is using their product for something bigger than they built it for. Your weirdest order might be your whole company.