Micro1 Seeks $1 Billion in Business Data for AI

Micro1 plans to invest $1 billion to acquire internal company data intended for AI developers. The announced offers range from $100,000 to $2 million per company, utilizing methods that combine data transformation, testing agents, and simulated environments. Executives from VanHack and MyPortal Marketing have already sold various work archives. The market, which is experiencing significant growth according to Gartner, raises questions about the role of employees.

Sales Without Employee Share: A Contested Framework

The internal datasets sold by companies rarely benefit the teams that produced them: current and former employees are unlikely to be compensated or consulted regarding the use of their work. Ifeoma Ajunwa, a law professor at Emory University, describes these sales as ethically gray areas. On the ground, some executives claim to inform their teams and report no complaints, like Grant Gooley at MyPortal Marketing, while others note a positive internal effect related to a new revenue source, such as Ilya Brotzky at VanHack. Meanwhile, Micro1 and its competitors are rapidly advancing in this niche.

Micro1 Licenses Complete Corpora and Offers Three Usage Paths

Micro1 structures its offering around three modalities for AI model developers. It can transform a company's raw data to protect privacy and then license it to a lab. It also offers a developing AI agent to companies to gather usage feedback. Finally, it freezes data at a given moment to provide training simulation environments. The collection covers sources like Slack, Canva, or Google Docs, and Micro1 generally seeks to obtain complete corpora to approximate the experience of an employee. According to William Almeida, the connectivity, diversity, and traceability of data points are crucial, and large, complex, or highly specific datasets are better valued.

From $100,000 to $2 million per Company and $1 Billion Announced Over One Year

Micro1 indicates it pays between $100,000 and $2 million to acquire data from a medium-sized company. The company has also announced plans to invest $1 billion next year to support this program. Executives who have signed agreements do not disclose the exact amounts received, but the market already values terabytes of professional interactions at several hundred thousand dollars.

Concrete Cases: VanHack and MyPortal Sell Their Internal Archives

At VanHack, Ilya Brotzky received a substantial offer after filling out a questionnaire and, reassured by the privacy procedures, agreed to share all requested data, including from Google Suite, Confluence, Slack, and the CRM. He mentions receiving a proposal from another broker, Mercor, which was ultimately dismissed in favor of Micro1. On the MyPortal Marketing side, CEO Grant Gooley views the agreement as an innovation and has sold files from Figma and Canva as well as checklists from Monday.com. He explains that he did not anticipate this use of his processes initially and finds it rewarding, even presenting the initiative to his teams as a contribution to the future growth of marketing agencies.

Booming Simulations and a Rush for Work Data

Access to work data is part of a trajectory where AI companies have moved from the web and books to human assessments, and then to complex simulations grounded in real data. Gartner estimates that spending on AI-related data could reach nearly $6.5 billion by 2027, up from less than $1 billion last year. At Micro1, William Almeida describes a booming sector and emphasizes to prospects the first-mover advantage. The company positions itself in a new category of brokers that condition internal data for the benefit of labs, in a context where developers are seeking real-world data and executives are discovering a new revenue source.