In a recent statement on July 14, 2026, Chamath Palihapitiya warned that the unchecked expenditure on AI tokens could significantly harm corporate earnings. The founder of Social Capital and CEO of AI startup 8090 highlighted a growing trend of companies maximizing their AI token usage without properly evaluating the return on investment.

Palihapitiya introduced a term he calls “tokenmaxxing.” He predicts that firms could soon face earnings per share misses, potentially by just a few cents, as expenses associated with AI usage spiral out of control, particularly in areas finance teams may overlook.

AI tokens serve as consumption credits for large language models from providers like OpenAI and Anthropic. Unlike traditional SaaS subscriptions that come with set monthly fees, the cost of these tokens can fluctuate dramatically based on usage, making budgeting particularly difficult for companies.

Drawing from his own experience, Palihapitiya mentioned that his startup's AI usage credits are projected to exceed $10 million annually by March 2026. He is not alone in his concerns; Palantir CEO Alex Karp has echoed similar sentiments, questioning the viability of token-driven pricing models.

In June 2026, Palihapitiya's 8090 secured a $135 million Series A funding round led by Salesforce. As he navigates his AI venture while actively spending on tokens, he simultaneously alerts other businesses that the current spending framework may be flawed.

For investors, these warnings present a clear risk. Companies that flaunt aggressive AI adoption without implementing effective cost management strategies could face unexpected earnings shocks. Mid-cap tech firms and enterprises that hastily integrated AI into their processes without the necessary financial oversight are particularly vulnerable. In contrast, larger tech companies that possess their own foundational models, such as Google, Meta, and Microsoft, may have some protection due to their control over the underlying infrastructure.

This article is for informational purposes only and does not constitute financial advice.