2 Sources
[1]
micro1 commits $1bn to company data for AI agents, financed by Citi
Companies can earn more than $1m for highly unique, proprietary data. micro1, which bid for Spirit Airlines' records, uses such data to build training environments for AI agents. AI training company micro1 will spend $1bn over the next 12 months buying and licensing operational data from
[2]
Micro1's $1 Billion Bet Isn't on Better Models. It's on Your Company's Data
Micro1, an AI company, is making a 10-figure commitment to data acquisition. On Friday, the company announced that it will spend $1 billion over the next 12 months to license enterprise operational data through its Company Data Partnerships program. The company landed the No. 37 spot on the 2025
Share
Copy Link
San Francisco-based micro1 announced a $1bn initiative to acquire operational data from companies over the next 12 months. Financed by Citi and Hercules Capital, the Company Data Partnerships programme will use de-identified company data to build reinforcement learning environments where AI agents learn real-world business workflows, decision-making and complex tasks.
San Francisco-based AI training company micro1 announced on Friday a $1bn commitment to purchase and license operational data from companies over the next 12 months. The $1bn initiative, financed by Citi and Hercules Capital, aims to expand the range of industries and workflows covered in micro1's training systems for AI agents.
1
The company, which ranked No. 37 on the 2025 Inc. 5000, will deploy this capital through its Company Data Partnerships programme to source proprietary operational data from businesses.2

Source: Inc.
micro1 uses de-identified company data to construct reinforcement learning environments that mirror actual business operations. In these environments, AI agents learn to navigate workflows, make decisions, and handle complex tasks that reflect real-world conditions. "Real business operations involve incomplete information, competing priorities and exceptions that require judgment," micro1 stated in its announcement.
1
The company emphasized that training AI agents with new data makes complex cases easier to handle, but requires environments that preserve the context behind how businesses actually work.2
The Company Data Partnerships programme targets specific types of operational information. micro1 seeks standard operating procedures, knowledge bases, internal documents, CRM data, project histories, and quality assurance processes. The company also wants decision-making records and human feedback on AI outputs.
1
Payment depends on the quality, uniqueness and value of the data provided. A qualified partnership earns more than $100,000, while large datasets or ongoing work across several teams can earn more than $500,000. Companies with highly unique, proprietary data can earn more than $1m.1
The initiative creates a new revenue stream for businesses to monetize the operational intelligence they have built over years.2
Related Stories
Companies retain ownership of their underlying data under the partnership terms. micro1 removes personal information "when relevant" and may rewrite some datasets synthetically to loosen their link to the original records. Companies can review samples before micro1 uses the data in its training systems.
1
The programme targets companies with more than 30 employees and established documentation, giving priority to US companies, then other Western markets. All material must be in English.1

Source: The Next Web
The company positions itself as a data lab that helps AI research companies and enterprises train frontier models and evaluate AI agents. It operates three in-house labs, including one focused on robotics.
1
In September, micro1 made a competing bid for Spirit Airlines' customer records during the airline's bankruptcy proceedings. The court-appointed privacy official has since backed Google's $10m deal for that data. Elon Musk's SpaceX has also explored buying data from startups to train its AI models, signaling broader industry interest in acquiring real-world datasets.1
This pattern suggests that access to proprietary operational data is becoming a competitive advantage in developing more capable AI agents that can handle real-world business workflows.Summarized by
Navi
21 Aug 2026•Business and Economy

29 Jul 2025•Business and Economy

05 Sept 2026•Business and Economy

1
Technology

2
Business and Economy

3
Technology
