OpenAI CFO Sarah Friar has published a four-question scorecard to help businesses determine if their AI investments are paying off. The framework emphasizes measuring 'useful intelligence per dollar' over traditional metrics like cost per token.
TL;DR
- OpenAI CFO Sarah Friar introduces a four-question scorecard to evaluate AI's economic value.
- The framework focuses on 'useful intelligence per dollar', assessing task completion, cost, reliability, and value growth.
What happened
OpenAI CFO Sarah Friar published a blog post outlining a four-question scorecard to measure the economic value of AI investments. The scorecard evaluates whether AI is completing meaningful work, the cost of successful tasks, the reliability of results, and the growth of value over time.
Friar argues that traditional software metrics like adoption and renewals are insufficient for AI. Instead, she proposes tracking the volume of high-quality AI-completed work, the total cost of that work, and the reliability of outputs.
OpenAI's Stargate initiative, announced in January 2025, plans to invest up to $500 billion over four years to build large-scale AI infrastructure in the U.S., with an initial phase targeting $100 billion and a goal of 10-gigawatt capacity by 2029.
Why it matters
For AI/ML developers, understanding and applying Friar's framework can help demonstrate the economic value of their work to stakeholders. It provides a concrete method to assess AI's impact beyond traditional metrics.
Startup founders can use this framework to evaluate their AI investments and communicate their value to investors. It offers a structured approach to measuring ROI, which is crucial for securing funding and justifying expenditures.
Tech investors can leverage this scorecard to assess the potential economic benefits of AI investments. It provides a clear set of criteria to evaluate the long-term viability and value of AI projects.
Key facts
- OpenAI CFO Sarah Friar introduced a four-question scorecard to measure AI's economic value.
- The scorecard focuses on 'useful intelligence per dollar', evaluating task completion, cost, reliability, and value growth.
- OpenAI's Stargate initiative aims to invest up to $500 billion over four years for AI infrastructure in the U.S.
- The initial phase of the Stargate initiative targets $100 billion and a 10-gigawatt capacity goal by 2029.
- OpenAI is valued at $852 billion and approaching the $1 trillion range.
- McKinsey's 24th annual Global CFO Forum revealed that about two-thirds of finance chiefs now have the strategy function reporting to them.
- Andy West, a senior partner at McKinsey, noted a shift in focus from experimenting with AI to enterprise-wide transformation.
Context
OpenAI, the company behind ChatGPT, is a leader in the AI industry. Its investments in AI infrastructure, such as the Stargate initiative, underscore the strategic importance of compute resources in driving AI advancements.
The increasing involvement of CFOs in strategic decision-making reflects the growing recognition of AI's transformative potential. Finance chiefs are now expected to play a pivotal role in determining long-term bets, including AI investments.
The shift in focus from experimentation to enterprise-wide transformation highlights the maturing of AI technologies. Companies are moving beyond initial trials to integrate AI deeply into their operations, aiming for significant economic benefits.
