VC Chamath: AI Costs Tripling Every 3 Months
Venture capitalist Chamath Palihapitiya warned that the costs associated with AI are tripling every three months, while revenues are failing to keep pace. His comments highlight growing concerns about the economic viability and potential bubble forming in the capital-intensive AI sector.
Palihapitiya's software startup, 8090, is reportedly on track to spend over $10 million a year on AI. He specified that these expenses, which have more than tripled since November 2025, are driven by "ginormous" inference costs from Amazon Web Services, as well as significant bills from AI coding tool Cursor and AI company Anthropic. He has been vocal about the need to find more cost-effective solutions, stating his company needs to migrate off of the AI coding tool Cursor, calling it "just too expensive" compared to alternatives like Anthropic's Claude Code. Palihapitiya also pointed to inefficient usage patterns, which he dubbed "Ralph Wiggum loops," where developers repeatedly prompt a model until it provides a solution, racking up large bills. The venture capitalist compared the current situation to the early days of ride-sharing, suggesting that venture capital is subsidizing "all-you-can-eat token consumption." He questions the long-term viability once this funding dries up, drawing a parallel to how Uber rides were initially cheap and then increased in price. This isn't Palihapitiya's only concern about the broader economic impact of artificial intelligence. He has previously warned that the immense power demands of AI data centers could cause electricity rates to double within the next five years, potentially leading to a public relations crisis for major tech companies. The discussion around AI expenses is growing, with other tech investors like Jason Calacanis and Mark Cuban echoing similar concerns. Calacanis noted paying $300 per day for a single AI agent operating at partial capacity, leading Palihapitiya to argue that for the cost to be justified, an AI model "need[s] to be at least two times as productive as another employee."
