Gokul Rajaram Argues AI Data Businesses Should Operate as Cash-Dividend Companies

Summary

Gokul Rajaram argues that many AI data businesses should be structured as cash-generating, dividend-paying companies rather than as ventures built primarily for durable equity value. He says these businesses can become profitable quickly and distribute cash soon after reaching revenue. His proposed compensation model would give employees and contractors a share of the revenue they help generate instead of relying entirely on fixed salaries, with attribution handled through an appropriate internal method. He also recommends raising only a small angel round, below $1 million, because a modest raise could still support contracts with AI labs while preserving the flexibility to pay dividends. Rajaram says venture capital funds are generally not structured to receive dividends. As an illustrative calculation, he describes raising $1 million at a $10 million post-money valuation, reaching a $50 million gross revenue run rate and $35 million net revenue run rate by the end of the first year, and generating $10 million in profit after lending costs tied to working-capital needs. Under that example, distributing 80% of profit would return $8 million in dividends, giving the investor 80% of the original investment back in the first year. He presents the approach as a lesson partly informed by earlier search-engine-marketing businesses and notes that working capital is becoming an increasingly important issue.