Cursor just did something that should make finance leaders sit up: it formalized the CFO conversation around AI ROI, and then stacked the room with new CFOs. The AI coding company, being acquired by SpaceX in a $60 billion all-stock transaction, has launched a CFO Council that will meet quarterly in rotating cities, beginning Aug. 18 in San Francisco. On Wednesday, Cursor announced additional members joining the group.

The new names are Michael Brophy, CFO of Natera; Bea Ordonez, CFO of Payoneer; Dinesh Jain, CFO of Firstsource; Matthew Wajner, CFO of First American; Ed Grabscheid, CFO of JFrog; and Andrew Casey, CFO of Amplitude. They join members announced on July 6: Sonalee Parekh, CFO of SentinelOne; Madhur Deora, CFO of Paytm; and Aziz Megji, CFO of Asana. New members are being accepted on a rolling basis.

The practical question behind the council is simple and brutally familiar to CFOs: companies are spending on AI, but it is hard to measure tactical ROI. Cursor COO Jordan Topoleski, who helped develop the council, said the goal is to address that exact measurement gap, particularly for Fortune 500 customers using Cursor’s agentic platform at scale. Even as CIOs and CTOs typically lead AI implementation, finance is increasingly being asked to answer the tougher question that land on budgets and board decks: where does a dollar of AI spending reliably generate meaningful business value rather than a diluted, speculative promise.

That pressure is showing up in how AI is being treated inside organizations. Topoleski framed a shift from AI being an experimental line item to becoming a material operating expense. Once AI moves from “pilot” to “run,” the finance function inherits the burden of proof. And proof in finance is not vibes. It is whether the output can be tied to outcomes the company can defend: productivity, cost efficiency, and adoption that does not balloon into runaway spend.

The council is designed as a working forum with cross-industry representation, including finance leaders from both tech-forward and legacy companies, and across public and private organizations. Topoleski described how Cursor assembled the council through a mix of inbound interest and outreach to customers, rather than a formal application process. The council also has an expectation of concrete outputs, not just networking.

According to Topoleski, the council is expected to produce shared benchmarks for AI productivity. It will also work on a framework for measuring “return on intelligence,” plus guidance on model allocation and cost controls intended to encourage adoption instead of stifling it. That “encourage adoption” detail matters. In most enterprises, cost control conversations can quickly become adoption blockers, because model usage is easy to throttle and hard to balance. Guidance that reduces uncertainty about costs while keeping teams moving could be the difference between AI that scales and AI that gets frozen.

If you zoom out, this is part of a broader pattern: AI tooling is evolving from a single-use product into an agentic workflow that routes work to different models depending on the task. Cursor’s setup reflects that. Its agent framework routes work to the most appropriate models for specific tasks. Meanwhile, its “canvases” feature lets finance teams connect directly to live data and build AI-powered dashboards without adding extra software layers. Topoleski said that internally, two of Cursor’s 10 most active users are in the finance department, signaling that AI-driven analytical work is already happening outside engineering.

That matters for CFO peers because it changes what “AI adoption” means. It is no longer only about whether teams can use AI. It is about whether AI intelligence translates into durable economics CFOs can defend in the boardroom. A quarterly council with specific deliverables is basically Cursor trying to productize the CFO proof loop: measure ROI, benchmark productivity, and control costs in a way that supports continued investment.

Now add SpaceX. Cursor’s $60 billion all-stock acquisition positions it inside a larger capital-and-operations universe where the tolerance for “too hard to measure” can be especially low. Even if the council is focused on practical tools, the signal to the market is clear: finance leaders are demanding more than adoption metrics. They want ROI frameworks, model allocation logic, and guardrails that keep AI spend accountable.

For executives watching this, the second-order implication is straightforward. When multiple CFOs across different companies join the same ROI working group, measurement norms can start to converge. Benchmarks for AI productivity and frameworks for “return on intelligence” become shared language, which changes internal board dynamics. The next time AI budgets are debated, the discussion may shift from justification to comparison: not “should we spend,” but “are we spending smart, and are we getting what others are getting?”