Synopsis: In this Techstrong.ai Leadership Insights interview, Jay Litkey, senior vice president for Flexera, explains why organizations should be crafting FinOps playbooks for optimizing the running of artificial intelligence (AI) models to maximize their return on investment (ROI).
The cost of AI is hitting hard, and most organizations don’t have a plan for how to manage it. Sound familiar? It should—it’s the same movie we saw with cloud. Everyone jumped in, spent freely, and only later realized they needed a framework. That’s where FinOps was born. Now, Jay Litkey from Flexera argues it’s time to bring those same principles to AI.
FinOps isn’t just about cutting costs. At its best, it’s about aligning finance, engineering, and business leaders around outcomes. With AI, that means agreeing up front on why you’re doing a project, what success looks like, and how you’ll measure it. The return on investment is often delayed, with benefits showing up as better decisions, efficiency gains, or customer satisfaction rather than quick revenue. If you don’t define that value at the start, the finance team is only going to see ballooning GPU bills.
Litkey stresses that AI costs can’t be viewed in isolation. Workloads vary—some belong on cutting-edge GPUs, others can run fine on cheaper or older infrastructure. Hybrid strategies, distributed computing, and understanding what models should run where are all part of the emerging playbook. And just like cloud, AI won’t be a one-size-fits-all environment.
Cross-functional alignment is critical. Finance doesn’t yet understand AI deeply, but they need a seat at the table early—before projects scale, not after. The same goes for diverse stakeholders across IT, engineering, and business.
We’re still early in the hype cycle, where expectations outpace ROI and pilots often fizzle. But if cloud taught us anything, it’s that FinOps maturity eventually turns chaos into discipline. Done right, spending more on AI isn’t a failure—it’s a sign you’re delivering measurable business value.
