Anthropic said it will commit 5 percent of its equity to a public benefit fund focused on AI's impact on employment, with the stake potentially growing by an additional 2 percent for every 50 percent increase in valuation above an undisclosed baseline.
The company also reported third-quarter revenue of $16.8 billion, though it did not disclose profitability or unit economics. The fund announcement comes ahead of a reported November initial public offering.
The structure matters here more than the headline number. Rather than a traditional 501(c)(3), Anthropic is housing the fund in a limited liability company—a choice that grants flexibility in deployment but sidesteps the tax advantages and public accountability built into nonprofit governance. The equity will not be used for lobbying or political contributions.
The fund's potential uses range from large grants to established nonprofits to direct worker support, including retraining programs. Anthropic said it has already allocated $200 million to AI research, and the new fund signals the company views job disruption as requiring deeper commitment.
Public concern is substantial. An August Pew Research Center report found 71 percent of U.S. adults believe AI will reduce jobs over the next two years.
Lise Buyer, a partner at IPO advisory firm Class V Group, said the timing serves a dual purpose: fulfilling SEC disclosure requirements for going public while providing what she characterized as favorable optics during heightened scrutiny of AI development. She noted the market would view the fund more favorably if deployment centered on worker retraining and job transition support.
The real tension: a 5 percent equity stake tied to upside means the fund scales only if the business succeeds spectacularly. At current valuation, that is meaningful. If valuation stalls or compresses post-IPO, so does the fund's growth. The structure aligns the fund's interests with shareholder returns rather than decoupling them.
Elon Musk, via X, proposed a different model—direct equity distribution to the public—highlighting disagreement over profit-sharing mechanisms in AI.
