Blog | Memree

4 June 2026

Tokens are the new cloud units and FinOps is not ready.

The Tokenomics Foundation launches at FinOps X this week, a Linux Foundation project dedicated to defining the primitives of the token economy. Its framing is precise and useful: token economics is the practice of managing the production, consumption, and monetisation of token-based AI to generate business value. Those three competencies are not the same as cloud cost management. Tokens are simultaneously the unit of cognition the model produces, the unit of compute the data centre serves, the unit of price the lab charges, and the unit of value the enterprise extracts and those four things do not move in lockstep. For hiring teams, this creates a familiar problem in an unfamiliar context: companies will soon need to hire for a role that does not yet have a shared definition. The practitioners who engage seriously with this shift now will be asking the right questions earlier than most.

29 May 2026

Why FinOps job descriptions are written backwards.

There is a pattern that appears with remarkable consistency across FinOps job postings. It begins with a list of certifications, continues with a stack of tools, and ends with a paragraph about “collaborating with stakeholders” that could describe any role in any function anywhere. What is almost never present is a description of the actual problem the hire is expected to solve. A job description describes inputs—experience, certifications, tool familiarity. What it rarely describes is what the role will be asked to produce. In FinOps this gap is especially costly, because the function is still being defined at most organisations and there is no job architecture to default to. The most useful question to ask before writing a FinOps job description is not “what do we need this person to know?” It is “what does this person need to change?” Answer that first, and the description, the search criteria, and the interview all follow from it.

20 May 2026

The “AI cost owner” is quietly becoming a real role.

Six months ago, AI spend was someone’s side responsibility—usually a platform engineer or a finance analyst who inherited it. Now we are starting to see dedicated mandates: someone whose actual job is to own how AI workloads are governed, attributed, and optimised. The problem is that most of these roles are being scoped by taking a cloud FinOps job description and pasting “AI” in front of it. The skills overlap, but they are not the same. Owning AI cost means understanding inference economics, model-selection trade-offs, and how agentic workflows scale with business volume—alongside the usual governance and stakeholder work. If you are about to open one of these roles, define it from the workload up, not from the old template down.

10 May 2026

Cost and carbon are starting to be the same conversation.

For years, cloud cost and cloud sustainability lived in separate meetings—one owned by finance, the other by an ESG team that rarely touched infrastructure. That is changing. The same levers that reduce spend—right-sizing, scheduling, region choice, efficient model selection—also reduce emissions, and boards are increasingly asking about both in the same breath. For FinOps, this means GreenOps is less a separate discipline than a second lens on the same data. For hiring, it means the profile is broadening: the strongest candidates can speak to cost and carbon without treating them as competing priorities. We think that convergence is permanent, and worth designing roles around now rather than later.

7 May 2026

FinOps X 2026 in San Diego: three things worth watching.

FinOps X returns to San Diego in June. Beyond the usual rate-and-discount sessions, three storylines deserve attention this year. First, how mature practitioners are absorbing AI workloads into their FinOps practice rather than handling them on a parallel track. Second, the slow shift from cost-reporting roles toward decision-support ones, which is quietly rewriting the job description. Third, what “FinOps Certified for AI” is doing to the senior talent profile. We’ll be watching, with hiring notes to follow.

1 May 2026

The bill for AI is starting to look like a strategy problem.

Token costs that were rounding errors twelve months ago are now showing up in P&L reviews. Inference scales with usage; training scales with ambition. Both are starting to need governance, not just dashboards. The question for hiring teams is whether your current FinOps lead has the mandate—and the technical depth—to own AI workloads, or whether you need a different shape of person at the table. The role profile is broadening, and many existing mandates were written before that was true.

24 April 2026

When the FinOps search isn’t working, the brief is the problem.

We hear this often: “we’ve been searching for six months and can’t find the right person.” Most of the time, that is not a candidate-pool problem. The brief was written before anyone clarified what the role actually needs to deliver, and now every shortlist looks plausible without any of them being a fit. The fix is not a wider search; it is a sharper mandate. The companies that get the right hire on the second attempt usually go back and rewrite the role first.

17 April 2026

Who actually owns FinOps in your organisation?

Cloud spend is a finance issue and an engineering issue and an executive issue, which is exactly why it ends up being nobody’s issue. The most durable FinOps practices we see have one thing in common: a single role with explicit authority to make calls across both sides. Whether that sits under the CFO, the CTO, or a Cloud Centre of Excellence matters less than whether the role has been defined with real decision rights—before the first hire is made.

Notes from the field.