Gross Profit per Token (tomtunguz.com)

Recent industry analysis suggests that gross profit per token may offer a more accurate indicator of AI companies' earnings potential than traditional revenue metrics. This comes as AI companies, known for lower gross margins compared to their software counterparts, increasingly adopt gross profit as a quota metric. A study comparing six AI inference companies revealed significant variations in gross profit margins and valuation multiples, highlighting how different monetization strategies affect perceived value. For example, Perplexity, primarily an application-focused AI, commands a staggering GP multiple of 222x, while DeepSeek and Together AI, which focus on reselling inference, trail with lower multiples. The implications of this shift are profound for investors and acquirers as they begin to prioritize the efficiency of token monetization over mere token volume. The study indicates a strong correlation (0.70) between gross profit per token and company valuation, suggesting that the market is already recognizing the importance of this metric, despite its relative obscurity in current discourse. This trend could reshape valuation strategies in the AI sector, encouraging companies to refine their monetization approaches and potentially leading to more informed investment decisions within the community.