He Deleted $7M in Revenue. Now His AI Data Company Is Worth $4B | Ali Ansari, micro1

Key Takeaways

• Markets beat products. Three companies in, Ali landed on his central lesson: "You can't take all the credit — the market is just incredible... You've got to be in the right market when you're building."
• Only founders can inject risk. Deleting a $7M-run-rate business wasn't a pivot — Ali rejects the word. It was a deliberate, founder-only decision: "You must inject risk into the company for these sorts of outcomes to be possible."
• Solo by default, team by design. Ali never had a co-founder because he didn't know the concept existed. What he built instead: a founding team so strong he says, "I don't even consider myself solo at micro1."
• Velocity is the variable you control. micro1 runs remote and async with a hard rule set: open a thread, propose solutions, minimize messages, get to "final actions." Never suggest circling back next week.
• Just start. His advice to anyone stuck in co-founder matchmaking: "Just start. Get somebody to pay you for something, and learn."

Listen: Apple Podcasts | Spotify

Introduction

There's a moment about half an hour into this conversation where Ali Ansari describes the day he recorded a Loom for his team titled "micro1 is going all in on data." The video meant shutting down the first version of his company — an AI recruiting business doing seven or eight million dollars in run rate. His team's reaction, by his own telling: "Are you sure about this? We're literally giving up."

He was sure. A year and a half into the company, one anomalous customer had come through micro1's AI recruiter and hired engineers at a scale micro1 had never seen — hundreds of hires where typical accounts made one or two. Ali got curious about what they were building, and the answer rewired his understanding of the market: the data industry was shifting from generalist labor to expert talent for training AI models, and his accidental product — an AI recruiter named Zara, originally an internal tool at his dev agency — happened to be the perfect machine for recruiting experts at scale.

What happened next is the subject of this episode: how a solo founder decides to delete a working business, drag a team through the fear that follows, and come out the other side running one of the fastest-growing AI data companies in the world. Ali has never had a co-founder — for the first few companies, he didn't know co-founders existed — and in this conversation with Julian he makes the most operator-flavored case for solo founding the show has recorded yet.

The Education of a Garage-Sale Arbitrageur

Ali's founding story starts at 12, when his mom looked at a kid glued to his phone and suggested he make some money with it. He found eBay, listed the first thing he could find — his dad's shoes, without asking — and watched them sell instantly because he'd priced them under market. "I just made like 70 bucks. Why is making money supposed to be so hard?" His dad had a different question: "Why are you selling my shoes?"

The house-inventory problem led to garage sales; the backpack-on-a-bike problem led to textbooks. From there, a straight line that only looks straight in retrospect: CashBooksNow, a textbook buyback site running on Amazon's API; a peer-to-peer tutoring company that hit six figures while he was still in high school; a dev agency at Berkeley; and inside that agency, the hiring bottleneck that became micro1.

What makes the timeline instructive isn't the hustle — it's the exits. Ali can name the specific catalyst that ended each business: Google Ads economics that wouldn't scale, a mentor bluntly asking "why are you working on this?", a side tool obviously bigger than the company that built it. All three taught him the same thing.

"You can't take all the credit — the market is just incredible... You've got to be in the right market when you're building."

Only Founders Can Inject Risk

The all-in decision wasn't clean. Months before, micro1 had lost about 60% of its revenue overnight when a major account — a future competitor — saw them edging into the data space and terminated. "We were very close to not surviving that," Ali says. The near-death made the decision personal: we must win in this space now.

This is the thesis of the episode, and Ali states it like a law of physics:

"Only the founding team can inject risk into the company in these ways — and you must inject risk into the company for these sorts of outcomes to be possible."

He's made the move more than once. Two years ago, micro1 still ran 15-minute human screening interviews downstream of Zara's reports. Ali decided overnight that the product had to carry the full weight: an emergency all-hands, every recruiter's calls cancelled that week, zero human screens from that day forward. A few people nearly quit. Then the expert pool started growing at a rate that, on micro1's internal charts, tracks revenue growth almost exactly.

The pattern he wants founders to internalize: outlier outcomes come from decisions your own team initially reads as crazy. Make none of them and you converge to a normal company.

Solo Because Nobody Told Him Otherwise

When Julian asks whether he's always been a solo founder, Ali's full answer is two words: "Yeah. Always." He didn't choose solo founding over the alternative — he didn't know there was an alternative. "I just thought: you start a company and you're running it. I didn't know what a co-founder was."

The orthodoxy only became visible when he applied to YC and sensed the unwritten requirement. He tried to reason his way to it and couldn't: the best explanation he found was investor risk management — if one founder quits, there's a spare. That's a reason for them, not for you.

His counter-model is the founding team. Early teammates and even his first-check investor, Joshua Browder, function as effective co-founders — minus the cap-table and governance weight.

"I don't even consider myself solo at micro1, to be honest — the founding team is that good."

What he refuses is the gate: "I'm not going to start the company until I find a co-founder — the matchmaking and all. It's nonsense. Just start. Get somebody to pay you for something, and learn."

The Operating System: Threads, Final Actions, No Circling Back

micro1 is remote-by-default and async-by-default in an industry that fetishizes badge-swipe hours. Ali's joke is that you can run "996 in stealth" — which he finds cheesy — or you can drop the arbitrary constraints and let someone ship from their couch at 2 a.m.

The async culture only works because it's ruthlessly specific. Problems open as threads. The thread-opener proposes solutions and owns the final actions. Participation is limited to people who add value; opinions that don't are withheld on purpose. The target is always the same: minimize messages until you reach a final action, because a discussion 10x longer usually converges on the same decisions. "You should never suggest circling back next week," Ali says. "These are phrases we avoid."

Progress, in his simplified model, is the quality of actions times the volume of actions — and velocity is the variable a team can actually control.

The Data Wars, From the Inside

What micro1 actually builds now is environments — seed data, tooling, tasks, and verifiers that frontier labs use to train models. Realism is the scaling law, and realism requires real operational data, which is why micro1 made headlines bidding $12.5 million for bankrupt Spirit Airlines' internal records after Google won the auction at $10 million. (The court outcome was still pending when this episode was recorded.)

It's also why the episode's most charged moment is about who shouldn't get this data. Ali, who came to the US from Iran when his family won the green card lottery, walks through why micro1 works only with the US and its allies — at a real cost to revenue:

"If you package [US company data] into environments and sell it to adversarial nations, you're exporting American intelligence. It's absurd to me that companies are able to do that."

Agree or not, it's the clearest articulation you'll hear of the stakes in the AI data supply chain — from someone building the supply chain.

The Case For (and Against) Going Solo

Julian closes with the customary two questions. Ali's bear case is thin by his own admission: it's lonelier early, and genuinely hard-mode companies (deep tech, heavy hardware) may warrant complementary co-founders — though he frames even that as a later-career move. "Those are not strong arguments, to be honest."

The bull case comes easier: faster decisions with one person clearly leading; board control, which he ties directly back to the ability to make outlier decisions; a founding team you can attract faster once something real exists; and — with a grin — "you get to own more of the company. That's a selfish one, but it's just the reality."

About Ali Ansari

Ali Ansari is the founder and CEO of micro1, the AI training-data company building environments for frontier AI labs. Born in Iran, he moved to the US after his family won the green card lottery, and started selling on eBay at 12. Before college he'd built a textbook-buyback site and sold a six-figure tutoring company; at Berkeley he ran a dev agency whose internal hiring tool became Zara, micro1's AI recruiter. In 2025 he deleted the profitable recruiting business to go all-in on AI training data — a bet that made micro1 one of the fastest-growing companies in AI. He has been a solo founder for every company he's started.

Subscribe to the Solo Founders Podcast on YouTube, Apple Podcasts, or Spotify.
Join our newsletter for weekly insights from solo founders: solofounders.beehiiv.com