Tech VCs: AI Agents Too Costly, Stablecoins May Become Agent Money

Prominent tech investors including Jason Calacanis, Chamath Palihapitiya and Mark Cuban say operating AI agents remains expensive and often exceeds the cost of human employees. Calacanis reported paying about $300/day for an Anthropic Claude agent (roughly $100k/year if run continuously) while utilization was only 10–20%. Palihapitiya argued models must be substantially more productive than humans to justify token/API and maintenance costs; Cuban highlighted ongoing token and upkeep expenses that can outstrip staff costs. These observations underscore that high AI agent operating costs and API/token billing are likely to slow large-scale, immediate job displacement and reduce near-term macro labour shocks. At the same time, crypto-industry leaders foresee AI agents using stablecoins for frequent microtransactions: Circle CEO Jeremy Allaire predicted billions of agents transacting in stablecoins within five years, and Binance’s CZ called crypto the native money for AI agents. The story notes increasing on-chain agent activity and experiments (examples in ecosystems like Base and Fetch.ai) and OpenAI’s new benchmark testing models on smart-contract security — a development with implications for both automated security testing and automated attack tools. For crypto traders: the near-term macro effect on markets may be muted because AI-driven layoffs are less likely while operating costs remain high, but medium-term trends could raise transaction volumes and new DeFi flows as agents and stablecoins gain traction. Key SEO keywords: AI agents, operating costs, stablecoins, token billing, smart-contract security. The term "AI agents" appears multiple times to improve search relevance.