I spent most of November doing what every CFO does in Q4: juggling year-end close prep, budget reviews, and the growing list of "can we talk before the holidays?" requests.

Meanwhile, the AI world delivered three shifts that finance leaders need to track: every major lab released meaningful upgrades, regulators made moves in both the US and the EU, and the infrastructure powering all of this hit a wall.

Not ideal timing, I know. But these aren't developments you can table until January. They're already affecting your costs, your compliance exposure, and your team's ability to keep up.

November was one of the biggest upgrade months of the year. OpenAI, Anthropic, Google, xAI, and leading open-source teams all released new versions or capabilities — raising the bar for reasoning, context handling, and reliability.

What launched this month:

General LLMs are becoming more capable across the board. A paid model subscription plus basic automation skills now unlocks significant efficiency gains for forecasting, audit prep, and financial modeling.

But the models are also becoming more complex. What worked reliably last quarter may behave differently now. Features get added, interfaces change, and performance characteristics shift with each release.

This creates a new imperative for finance teams: constant learning, an AI-first mindset, and empowering your people with tools, knowledge, and bandwidth to experiment are a must-have now. Teams that treat AI as a static tool will fall behind teams that build continuous learning into their workflow.

Open-source progress also accelerates private-cloud and data-sovereignty options for firms that need to keep sensitive data on-premises.

Regulators on both sides of the Atlantic moved aggressively in November. The U.S. sought to consolidate federal control over AI policy, while Europe proposed structural changes to privacy and AI laws.

United States: Attempt to preempt state AI laws stalls — for now

A draft Trump administration executive order aimed to override state AI regulations, threaten federal lawsuits, and withhold funding from states with "burdensome" AI laws.

Draft language directed agencies to identify state rules to challenge.

35 state attorneys general pushed back hard, urging Congress to block federal preemption and preserve states' authority to regulate AI.

The federal effort is paused but far from resolved. Congress may still revisit the idea in 2026.

European Union: Digital Omnibus proposes major relaxations of GDPR and AI Act

The European Commission introduced the Digital Omnibus package, a sweeping set of changes to GDPR, the AI Act, and related digital regulations.

Proposed changes include broader exceptions for AI developers to process special-category data, narrower definitions of sensitive data, potential reclassification of pseudonymized data, relaxed cookie and tracking rules, and delayed AI Act enforcement into 2027 in some areas.

Critics warned the revisions could undermine GDPR's foundation, calling it "death by a thousand cuts."

Multistate operations in the U.S. must prepare for regulatory fragmentation until a federal standard emerges. Multinational firms may face dual shifts: lighter rules in the EU but uncertainty on the U.S. side.

Data-classification policies, consent documentation, and model-usage logs will become more important than ever. The regulatory ground is shifting under your feet, and documentation becomes your proof of good-faith compliance.

Boards will expect quarterly updates on regulatory exposure tied to AI. If you're not tracking this, start now.

A wave of November reporting highlighted a new constraint on AI growth: infrastructure, not models, is now the bottleneck. This has direct implications for budgeting, cloud cost control, and long-term AI strategy.

What happened:

Big Tech announced major increases in AI-related capex. Microsoft reported accelerated spend driven specifically by AI infrastructure. Alphabet highlighted continued "significant investment in servers and AI compute."

GPU supply remains tight. Nvidia dictates the pace of the entire sector. Bloomberg reported long lead times and heavy pre-ordering of top-tier GPUs, constraining throughput for anyone scaling AI workloads.

Power and data-center capacity became real bottlenecks. Data-center energy constraints emerged in multiple U.S. and EU regions, delaying expansion and raising operating costs.

The "AI debt boom" began. AI infrastructure is increasingly funded through corporate debt issuances, raising systemic financial exposure across the tech sector.

Cloud price pressure is expected in 2026. Analysts warned that electricity costs and GPU scarcity will drive cloud price increases, especially for AI-heavy workloads.

AI expansion is now a capex story, not a software story. The infrastructure layer is getting expensive, constrained, and risky.

SaaS and ERP vendors will eventually pass their rising compute costs onto customers.

AI workloads will cause cloud-spend volatility during budgeting, audit, or forecasting cycles. The costs are less predictable than traditional SaaS.

Vendor risk increases. Companies relying on scarce GPU allocations may face degraded performance or feature delays. Your finance software's AI features might slow down or become unavailable during peak usage.

That's the “what” and the “so what” of November: model upgrades that demand continuous learning, regulatory chaos on both continents, and infrastructure constraints heading for your budget.

Here's the “what to do”.

Do these three things in December:

The subscriber-only section below walks through exactly how to do each one, with ready-to-use policy language, contract terms, board reporting templates, and a cost forecasting spreadsheet you can use in budget meetings.

That's November wrapped. Three major shifts happening simultaneously while we're all trying to close out the year.

The newsletter continues through December with practical editions on the topics you've been asking about: AI-powered forecasting workflows, year-end close automation, and setting up governance frameworks that actually work.

If you've been forwarding these editions to colleagues or have feedback on what would be most useful heading into 2026, hit reply. I read every response.

Thanks for reading, and I'll see you next Tuesday.

This newsletter is for you, and we want to make it as valuable as possible. Please reply to this email with your questions, comments, or topics you'd like to see covered in future issues. Your input shapes our content!

Want to dive deeper into balanced AI adoption for your finance team? Or do you want to hire an AI-powered CFO? Book a consultation!

Did you find this newsletter helpful? Forward it to a colleague who might benefit!

Until next Tuesday, keep balancing!

Anna Tiomina
AI-Powered CFO