Chamath Palihapitiya raised alarms on July 18 about the US government considering restrictions on open-source AI models. He warns such a move could put American companies at a steep disadvantage, forcing them to pay 50 times more for AI capabilities than competitors abroad.
Specifically, Palihapitiya estimates US firms might pay $26 to $56 per million tokens for proprietary AI services. In contrast, companies elsewhere relying on open-source alternatives would only pay about 50 cents to $1 for the same processing power. This cost gap isn’t trivial it could significantly inflate operational expenses and weigh on profit margins.
He pointed out that AI token costs at his own firm are already doubling every 45 days. If the US bans or tightly controls open-weight AI models, it won’t stop global access. Instead, non-American companies in China and Europe would continue advancing at lower costs while US businesses face a costly bottleneck.
Market Impact and Investor Signals
Higher AI expenses could force US companies to lower earnings expectations, pressuring tech valuations and possibly shaking crypto markets as well. Investors should watch which firms diversify their AI sources versus those relying heavily on proprietary providers.
Former Twitter CEO Jack Dorsey endorsed Palihapitiya's stance, emphasizing the importance of an open approach to AI. The economic disparity highlighted by these figures could push the market to demand either policy changes or a revaluation of companies exposed to AI costs.
