State of FinOps 2024: Reduce waste and use AI - Wishevoke

State of FinOps 2024: Reduce waste and use AI

The FinOps Foundation, which represents practitioners using spend observability to empower engineering teams, recently released its 2024 study State of FinOps Survey Report. In what the report describes as a reflection of “macroeconomic trends,” respondents gave waste and cost reduction a top priority. For the first time, the category “Empowering Engineers to Take Action” is no longer the top priority in the survey. The report also provides guidance on creating guardrails to ensure financial transparency of AI projects and highlights the value of FinOps data for sustainability initiatives.

Engineers remain the biggest beneficiaries of FinOps observability, even as engineering enablement has fallen to a lower spot in the report’s ranking of priorities examined. This suggests that engineers are best suited to respond to a sudden change in cost metrics. The report finds that the “engineering personality” reportedly derives the greatest benefit from both “FinOps training and self-service reporting.”

Mike Routier gave a FinOps X talk in 2023 titled “ Empower engineers to make cost-conscious decisions. Routier, senior cloud cost analyst at Choice Hotel International, described creating “cost discipline” in teams by providing spend intelligence that allows engineers to target cost avoidance or cost savings using a combination of AWS resource tagging, billing reports, anomaly detection, etc. uses monthly team check-ins. He shared:

Most of our engineers want to be cost-conscious but don’t know which decision is the right one. This is where I come in and say… these are the opportunities that I would take advantage of and that offer the most value.

While waste reduction is a common driver for all respondents; Segmenting the survey by cloud spending found that those with smaller budgets would tend to prioritize improvements in billing forecast accuracy. The report said these respondents faced the challenge of understanding “the evolution of spending” before it “spiraled out of control.” Most invested in low-effort solutions such as “manual adjustments” to generate forecast data. In contrast, those with larger budgets tended to prioritize optimization Loyalty-based discounts to benefit from economies of scale. This included proper sizing of “reserved instances, savings plans, committed usage discounts,” as well as specific negotiated discounts. Below is a breakdown of respondents’ priorities by cloud spending.

Respondents’ priorities by cloud spending.  Source: www.finops.org

Respondents’ priorities by cloud spending. (Source: FinOps Foundation)

Matt Saunders recently reported for InfoQ on AI and FinOps unite to drive innovation in cost observability in 2024. Citing people from Grafana, Saunders wrote that the interaction of AI and financial observability will “enable strategic decisions and avoid excessive cost cutting in revenue-generating areas.” The 2024 survey also highlights the use of machine learning to combat waste, highlighting two areas:

  • “AI for FinOps” – Use of ML for great transparency, optimization and forecasting of costs.
  • FinOps for AI – Gain early cost visibility as teams scale AI experiments, projects, and spend.

The survey describes this current phase of AI as “a throwback to when the cloud itself first came onto the scene.” It notes that cloud migrations initially had “little consideration of cost or governance,” and remained so until “budget thresholds” were exceeded at an ever-increasing rate. Warning that cloud experimentation has suddenly “slowed or stopped”; The report advises practitioners to implement early observability into AI and ML infrastructure:

For any new cloud technology an organization plans to adopt, we recommend that reporting, forecasting, and basic guardrails be implemented early on to ensure that experimentation is possible within limits and major rework is avoided.

The survey looks at examples of areas where there was waste and cost reduction, and shows that the greatest optimization efforts went toward enabling teams to manage compute spend, which for most is also the “top spend ” is. Such optimizations have not yet penetrated into areas “related to newer technologies,” with data indicating that optimizations to “AI/ML” spend remain under-invested. While smaller companies are making smaller AI investments, the survey found that larger companies are investing in AI and ML are a “rapidly growing source of variable costs that need to be managed.”

Optimizations and waste reduction by cost category.  Source: www.finops.org

Optimization and waste reduction by cost category. (Source: FinOps Foundation)

To help practitioners reduce waste, the 2023 survey led to the FinOps Foundation creating a Library of waste reduction scenarios. Scenarios are organized by cost and cloud provider and range from optimizing EBS volume utilization to right-sizing capacity commitments. Each scenario comes with a short case study, reference material and examples of relevant infrastructure-as-code.

In a recent outlook for 2024, Forbes published an article about it Impact of AI on Cloud ERP (Enterprise Resource Planning) in 2024, indicating that the resulting insights will enable improved tracking of “sustainability goals” and environmental impacts through a “green ledger.” The Green Ledger is a concept that existed discussed at Cop28 This correlates financial investments with the observability of carbon emissions. The State of FinOps Report states that “less than 20% of FinOps teams are currently collaborating with sustainability teams (or taking responsibility for cloud sustainability themselves).” Half of respondents say that “collaborating with sustainability teams” will increase in the future.

Respondents to the State of FinOps survey provide insight into the adoption of FinOps practices across a range of organizations. 23% of respondents came from companies with fewer than 1,000 employees and 15% came from companies with over 100,000 employees. Respondents reported an average annual cloud spend of $44 million and reported major spend of over $1 billion. The FinOps Foundation has made its data queryable https://data.finops.org.

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top