I want to start this week’s newsletter by wishing all the Fathers out there a Happy (belated) Father’s Day. I know I’m a couple of days late, so I hope everyone got a chance to spend time with family and friends celebrating.

What did I do over the weekend … Well, I read a number of articles associated with FinOps, AI, and the markets. And I came away with a few thoughts and feedback.

I want to start with a post from last week: Pew published its 2026 read on how Americans feel about AI. About half of us now use AI chatbots at least sometimes. Daily use is already normal for a meaningful share of adults. Yet, only 16% think AI will make society better over the next twenty years. 40% think it will make things worse. (Wow!)

That research tells you something about where we actually are. We are moving fast, and we do not yet trust what we are moving toward. That gap between adoption and confidence is not really a story about the technology. It is a story about value nobody has fully proven, and a foundation nobody has finished building.

The money is doing what money does in a gold rush. SpaceX went public on June 12, then moved into a reported $60 billion deal for Cursor’s parent. OpenAI’s path toward public-market financing keeps getting louder. GitHub Copilot moved to token billing on June 1, and power users watched projected bills jump dramatically almost overnight, with some reporting costs far above their old subscriptions in Business Insider and Tom’s Hardware.

Every cycle works the same way. Cloud was the gold rush, then SaaS, now AI. Each time, the pitch is identical: the new thing changes everything, so buy in before you are left behind. And each time, we sprint toward the new layer before we have gotten any real control of the last one.

That is the part worth sitting with. While everyone races to govern AI spend, most companies still cannot tell you what they are paying for the software they bought three years ago, or who is using it.

Look at SaaS, the gold rush right before this one. The average enterprise now runs hundreds of SaaS applications, with large enterprises far beyond that and shadow IT adding more on top that nobody officially tracks. Roughly half of purchased licenses go unused. Only about half of users log in during a given month. The average enterprise wastes millions a year on software that delivers nothing at all. Repeat for Cloud.

That is not a frontier-technology problem. That is the basics, unmanaged, years after the hype that sold them moved on to the next thing.

I see this every time I talk or work with an organization. The AI conversation is the loudest one in the building. Meanwhile no one can produce a clean list of what the company actually spends across its tools, who owns each contract, or what half of it is even for. The enthusiasm is all pointed at the newest layer. The mess is sitting underneath it, untouched.

And now we are stacking AI on top of that. If a company cannot tell you who owns a SaaS license or a cloud resource, and whether anyone logs into it or is using the resource, it has no chance of governing an AI agent that spins up spend across cloud, tools, and data on its own. You do not get to skip the fundamentals because the new thing is exciting. The new thing makes the missing fundamentals more expensive, and it does it faster. But FinOps teams are being required to manage.

Here is the uncomfortable part, and it matches what I keep hearing from others across the industry. FinOps is still early, too.

The State of FinOps 2026 shows the scope exploding. Nine in ten practitioners are now asked to manage SaaS, up from 65% a year ago. Licensing jumped to 64%, private cloud to 57%, data centers to 48%. AI sits at 98%. That looks like maturity. It is not, not yet. Being asked to manage something is not the same as controlling it, and the survey is honest about the gap: even organizations that have cloud handled are starting over at square one in every new domain they take on.

So the real picture is a young discipline being handed five new problems at once, while the hype demands it solve the newest and shiniest one first. That is backwards. You do not stabilize an organization by chasing its flashiest line item. You stabilize it by getting the fundamentals in across everything it spends, then taking on the new domains from a base that actually holds.

The fix is not another tool, and it is not an AI-specific cost dashboard. It is governance and planning, applied to all of it, while the basics get implemented underneath.

The FinOps Foundation landed on the same conclusion this year. Its read is that the next wave of impact comes from governing and shaping spend before it happens, not chasing waste after the fact. Automated policy. Guardrails before deployment. Tagging that actually holds. Ownership that is named and accountable. None of it is exciting, and all of it is what makes every later decision cheaper, AI included.

The order of operations matters more than the tooling. Inventory what you actually spend, across cloud, SaaS, licensing, and AI, not just the part that is trending. Give every recurring cost a named owner, and treat no owner as a reason not to renew. Reclaim what nobody uses before you buy anything new. Put a governance policy in front of spend instead of a report behind it. Do that, and when AI arrives as the next domain, it slots into a system that already works instead of landing on a pile that does not. And this is not just hyperbole. I’m implementing this same plan concept and framework with clients I work with.

AI is not a special case. It is the newest layer on a foundation most companies have not poured. So, pour the foundation.

The gold rush will keep moving, and the pressure to chase it will keep coming from every vendor pitch and every headline. The FinOps job is to build the fundamentals while everyone else sprints ahead. Governance and planning, applied evenly, before the spend, across the whole estate, not just the part that is in fashion this quarter.

Get the basics in and AI becomes manageable. Skip them and AI just makes the mess bigger and faster. Slow down enough to pour the foundation. That is the unglamorous work that decides whether any of this ever pays off.

Written with the help of AI. All the ideas expressed are mine and mine alone.

If you would like your company included in the Spotlight, contact the CloudXray AI Team

Company: CloudXray AI

Category: Managed Services & Consulting

What They Do: FinOps Consulting & Advisory Services; Owners & maintainers of the single largest FinOps company directory (finops.cloudxray.ai)

Why It Matters: Companies still need guidance on implementing FinOps and understanding the landscape of companies that exist