OpenAI's CFO Sarah Friar has published a strategic overview of how the company is pursuing "abundant intelligence"—AI that is simultaneously more capable, faster, and cheaper. Friar explains how OpenAI's approach ties together custom chips (like Jalapeño), computational infrastructure, model architecture, and product design into a unified system.

The thesis is that breakthroughs in any single layer compound across the stack. Better chips reduce inference costs, which enables larger-scale deployments, which generate more data for model training, which improves capabilities, which drives more adoption. This interconnected strategy contrasts with competitors who may excel in one area but lag in others.

What This Means for Your Business

This document articulates OpenAI's multi-year competitive strategy. For enterprises evaluating long-term AI partnerships, understanding OpenAI's vertical integration (chips + compute + models + products) helps assess whether the company can sustain cost advantages. The strategy suggests OpenAI believes it can reduce the price of AI capabilities faster than competitors, which has direct implications for procurement decisions and contract negotiations. Companies locked into multi-year vendor agreements should assess whether their terms allow renegotiation as commodity costs inevitably fall.