Brief IA

Micro1: $500M Gross and High Margins on Data Reselling

🤖 Models & LLM·Tom Levy·

Micro1: $500M Gross and High Margins on Data Reselling

Micro1: $500M Gross and High Margins on Data Reselling
Key Takeaways
1Micro1 reaches $500 million in annual gross revenue, up from $100 million eight months ago (according to a source close to the company)
2Margins of 80 to 90% are reported on ready-to-use resold data (according to a source close to the finances)
3Ali Ansari claims not to sell to Chinese manufacturers and denounces rivals who do
💡Why it mattersResearchers estimate that spending on data could rival that of computing, while the demand for unique data boosts the revenues of annotation specialists.
Le brief IA que lisent les pros

Le brief IA que les pros lisent chaque soir

Les 7 actus IA du jour, décryptées en 5 min. Gratuit.

Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.

Choisis ton rythme

Gratuit · Pas de spam · Désabonnement en 1 clic

📄
Full Analysis

Micro1 claims gross margins of 80 to 90% on reusable datasets and asserts that it does not sell to Chinese manufacturers, despite the resale fueling controversy. The four-year-old startup has seen its gross revenue soar from $100 million to $500 million in eight months and reports an increase in contracts, while betting more on synthetic generation.

Resale of Data and Positioning Regarding China

Part of the data produced by Micro1 is sold to multiple clients. According to a source close to the startup's finances, these so-called "ready-to-use" datasets yield gross margins between 80% and 90%. However, the commercialization of the same sets to several buyers has sparked controversy, with critics arguing that the distribution of ready-to-use data to Chinese AI developers helps bring their models closer to the best American ones. Founder Ali Ansari stated last month on X that Micro1 does not sell its data to Chinese model manufacturers, unlike some competitors. He claims that human data companies work with foreign adversaries, citing Kimi K3 as an example, and finds it shameful to claim dominance in American AI while selling millions of dollars worth of data to hostile competing countries. Meanwhile, Micro1 says it is generating more and more synthetic data without human intervention, for example through automated descriptions of video content.

Significant Revenue Increase and Larger Contracts

According to a source close to the company, Micro1's annual gross revenue has risen from $100 million to $500 million in eight months. The company reportedly retains about 60% to 70% of this, placing its annual net between $150 million and $200 million. Micro1 reports seeing the size of its contracts grow at an accelerated pace and anticipates an improvement in its margins over time. On the funding side, the company raised its Series A at a $500 million valuation last September and has recently completed a new round at a significantly higher valuation.

A Sector Driven by Demand and Billionaire Rivals

The demand for unique AI training data from large labs and companies fuels the market for annotation specialists. In this landscape, competitors like Mercor and Handshake reached $2 billion this summer and $1 billion earlier this year, leaving Micro1 still behind them in gross revenue levels. Researchers also estimate that, in the future, spending on data could rival that of computing.

From AI Recruitment to 'RL Gyms' and Robotics

Micro1 started as an AI recruitment startup, similar to Mercor. After observing that some of its annotation clients were using its platform to evaluate and hire engineers, Ali Ansari chose to pivot the business towards data annotation. The company submits model results for evaluation by experts, a method called reinforcement learning gyms, and develops a pre-training dataset in robotics by enlisting several hundred generalists to record common object manipulations at home. Micro1 did not respond to a request for comment.

Brief IA — L'actualité IA en français

L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.