Inside Anthropic’s IPO + How AI Agents Spy on Your Messages

Anthropic’s recent S-1 filing reveals a company scaling at unprecedented rates, with 900% net revenue retention and positive operating cash flow, effectively countering skepticism regarding the profitability of frontier AI labs. To maintain control over safety-critical decisions, the company employs a founder-controlled voting structure and public benefit designation, insulating it from short-term shareholder pressure. Meanwhile, Signal Foundation president Meredith Whittaker highlights the severe privacy risks posed by AI agents, which often capture screen data and bypass encryption protocols, threatening the integrity of secure communications. Whittaker argues that the industry’s branding of these systems as "AI" obscures the material reality of deep learning architectures, which prioritize data extraction and power consolidation. These developments signal a critical juncture where the convenience of agentic systems increasingly conflicts with the fundamental necessity of private, secure communication in a digital-first economy.