In partnership with
AI
What I Noticed Reading the re:Invent FinOps Keynote Slides
I was not at re:Invent in 2025, but I spent time going through the AWS FinOps keynote slides afterward.
What stood out was not the AI announcements themselves. It was how familiar the themes felt if you already spend time around cloud cost, forecasting, or procurement.
If you read between the lines, the keynote did not introduce a new direction. It reinforced where FinOps and AI are already heading into 2026.
AI spend is a legit problem, but still scattered
One stat showed up clearly: most organizations are still experimenting. Very few are running AI workloads at meaningful scale.
That tracks. AI spend today often lives in pilots, proofs of concept, or small teams testing ideas. From a finance lens, that makes it hard to budget. From a FinOps lens, it makes cost data reactive and incomplete.
The signal for 2026 is not aggressive optimization… yet. It is getting visibility before the spend solidifies.
Data hygiene is still doing the heavy lifting
The keynote kept circling back to familiar problems: missing tags, unclear ownership, shared services, and poor allocation.
This is not an AI maturity gap. It is a cloud maturity gap that AI is amplifying.
Finance cannot talk about ROI if costs are floating. Procurement cannot plan commitments if usage is fragmented. FinOps ends up reconciling instead of analyzing.
In 2026, work like cost allocation models, shared cost strategies, and standards like FOCUS 1.3 will matter more than whatever the next AI feature is.
GenAI is NOT replacing forecasting
One thing the slides were fairly clear about: traditional machine learning still handles forecasting, anomaly detection, and trend analysis better than generative AI.
GenAI is leveraged after the math. It helps explain drivers, summarize changes, and answer questions in plain language.
For finance, this keeps the numbers defensible. For engineering, it lowers the barrier to understanding cost. For FinOps, it reinforces the role of translation instead of control.
AI ROI struggles for predictable reasons
The keynote did not pretend AI ROI is easy; costs are incomplete, benefits are indirect, timelines move, and baselines are missing.
This does not mean AI is failing. Right now, it is a mismatch between how finance expects ROI to work and the reality of how AI value shows up in our current environment.
In 2026, AI ROI will likely be treated more like a portfolio. Directional confidence over time, not precision on day one. FinOps helps by making tradeoffs visible, not by forcing false certainty.
GPU cost management is becoming a thing
One case study focused on GPU usage with extremely granular tracking and accountability.
This feels like the early days of cloud compute all over again. GPUs are expensive, shared, and easy to waste if no one is watching.
By 2026, GPU FinOps will likely look less experimental and more structured, with its own metrics, planning cycles, and governance.
The bigger takeaway
If you strip away the stage language, the keynote did not introduce a new FinOps playbook. It reinforced the existing one.
Measure, allocate, optimize, govern. AI just increases the number of decisions happening before finance ever sees the bill.
That is why FinOps, finance, leadership, and procurement conversations are tightening right now. Not because AI is magical, but because uncertainty forces coordination.
Quick Note: Starting off 2026, the takeaway is not that AI changed everything. It clarified where our foundations are strong and where they are not. As we move into 2026, FinOps continues to matter not because of any single technology, but because it helps organizations make sense of uncertainty, together.
The Future of Shopping? AI + Actual Humans.
AI has changed how consumers shop by speeding up research. But one thing hasn’t changed: shoppers still trust people more than AI.
Levanta’s new Affiliate 3.0 Consumer Report reveals a major shift in how shoppers blend AI tools with human influence. Consumers use AI to explore options, but when it comes time to buy, they still turn to creators, communities, and real experiences to validate their decisions.
The data shows:
Only 10% of shoppers buy through AI-recommended links
87% discover products through creators, blogs, or communities they trust
Human sources like reviews and creators rank higher in trust than AI recommendations
The most effective brands are combining AI discovery with authentic human influence to drive measurable conversions.
Affiliate marketing isn’t being replaced by AI, it’s being amplified by it.
RESOURCES
The Burn-Down Bulletin: More Things to Know
Capital One is rethinking AWS for AI as GPU costs climb
A reported internal memo shows Capital One questioning how sustainable large-scale AI workloads are on traditional hyperscaler pricing. This is a real-world example of AI spend moving from experimentation into executive-level cost strategy conversations, especially around GPUs.CIOs will underestimate AI infrastructure costs by 30%, IDC says
IDC highlights a growing gap between expected and actual AI infrastructure costs, driven largely by GPU usage and supporting services. This reinforces why finance and FinOps teams are being pulled deeper into AI planning earlier than expected.AI cost governance report shows margins under pressure
This report found that most organizations are seeing AI costs directly impact margins, with forecasting accuracy still low. It is a useful reminder that AI ROI struggles are structural, not execution failures.Tangoe wins InfoWorld Technology of the Year for cloud cost management
Recognition like this signals where enterprise attention is shifting: from basic visibility to disciplined, scalable cost control. As AI workloads grow, cloud cost management tooling is becoming foundational, not optional.
A Short Letter to My Subscribers + FinOps Community:
This newsletter started in the second half of 2025, and I did not expect the level of engagement and curiosity that has grown around it in such a short amount of time.
When I began my professional teaching career, I thought I would stay in that field for most of my life. Funny how we make plans and the universe laughs. I knew fairly early it was not the path for me, but I had no clear sense of where my skills or curiosity would land.
FinOps ended up being that place. Not because I had it all figured out, but because it gave me a way to keep learning while helping others make sense of complicated systems. This newsletter grew out of that same instinct: to slow things down, ask better questions, and think through cloud finance together without pretending there are easy answers.
I am grateful for everyone who reads, shares, challenges, and builds alongside me here. This space works because it is curious, honest, and grounded in genuine work.
As we head into 2026, my goal stays simple. Keep learning out loud. Keep connecting finance, engineering, procurement, and leadership in ways that actually help. And keep making FinOps feel approachable for people who are still finding their footing.
Thank you for being here.
That’s all for this week. See you next Tuesday!