Is There an AI Bubble? What 2025 Taught CFOs for 2026

2025 has been another eventful year of AI advancements. Frontier AI models like GPT-5, Claude 4, and others delivered significant capability increases.

Specifically, a substantial increase in multimodal tasks (analysing voice, text, images, videos), a higher context window (in simple terms, can handle more prompts before running out), reduced hallucination, and extended thinking using web-search capabilities. Crucially, agentic AI breakthroughs in 2025 enabled these models to act much more autonomously. They could retrieve relevant information, make decisions, and iteratively refine outputs without constant human input. This could be a huge step towards bringing us closer to a practical autonomous finance.

However, progress in technology doesn’t equal progress in enterprise adoption. Some researchers argue that we might be close to hitting an AI-capability plateau, where the marginal capability increase does not justify the massive investments being made behind these frontier models. Could this potentially mean an AI bubble brewing?

There is no deterministic answer to this question. And as a business leader, there is little to be gained from this debate. Bubble or not, there is no denying how AI can transform operations and reinvent business models. Case in point, our focus needs to be on how we can unlock this potential as opposed to giving in to the hype and buzz.

What CFOs can learn for 2026

A sobering reality of 2025 was the MIT report I discussed in my last article. Many firms that invested heavily in AI saw limited returns. The initiatives are mostly derailed by fragmented data, skill shortages, and governance gaps. This gap between capability and impact raises a critical question: is AI a genuine transformation or another cycle of inflated expectations? The answer or solution, for CFOs, lies not in the technology itself but in three core lessons from 2025. Getting these right separates winners from those chasing hype.

Lesson 1: AI Capabilities Are Real but Complex

Despite hype, 2025 showed that frontier AI models have truly advanced in reasoning, contextual understanding, and instruction-following. However, using these capabilities in finance still requires deep process alignment, data maturity, and a robust business case.

Unlike a plug-and-play SAAS solution, AI requires a deeper wiring within the organisation. A standalone ‘use-case’ based approach can result in markedly disparate AI tools not talking to each other — creating further tech and data complexities.

Think of it like an orchestra. Each instrument can create a beautiful sound of its own, but only through conducting them in the right manner can a symphony be created. CFOs need to act as a conductor of the orchestra, where individual instruments are not just AI tools, but also your existing financial, operational, and reporting systems and processes. However, to build such an orchestra, we need to redesign the financial operating model. In my book “Reimagine Finance”, I posit three fundamental changes needed to build such an operating model: treating data as a strategic asset, building micro-digital innovation hubs, and reimagining finance talent strategies.

Lesson 2: Data Readiness – the (not so) hidden constraint

AI’s transformational power is only as strong as the data that feeds it. As Google’s Chief Scientist once shed light on Google’s dominance, “We don’t have better algorithms than anyone else; we just have more data.” If we can learn one thing from 2025, it is that fragmented, poor-quality data and siloed systems remain a primary blocker, causing many AI projects to stall or deliver underwhelming results. CFOs need to influence and co-create a data strategy with their CIOs and CTOs. A trusted, governable data foundation is a prerequisite to successful AI scaling. In the first quarter of 2026, every organization, if not done already, must carry out an AI-readiness assessment with a specific focus on its data capability.

As I briefly mentioned, treating data as a product (or strategic asset) will be a fundamental shift in both methodology and mindset. Think of data products as a complete, pre-packaged, reusable dataset that can be used across the organisation for multiple use cases, without any transformation. For example, a Customer 360 data product will have all the information regarding purchase history, demographics, preferences, payments, etc. This one data product will be used by marketing for campaign optimization and effectiveness measurement, sales for promotional planning and negotiations, finance for AR management and insights, and AI teams to build and deploy new AI/ML models.

Lesson 3: Talent and Governance Gaps Threaten ROI

Another lesson that can be drawn from this widening gap between AI investment and business impact is the persistent talent shortages and immature governance frameworks. According to a survey, only 54% of senior finance professionals with over a decade of experience feel equipped to use AI effectively. On the other hand, 89% of finance students claim sufficient AI experience, demonstrating a stark generational digital divide. While the new workforce will play a significant role in the future, current seasoned professionals who drive strategic decisions are underprepared.

Similarly, governance lags dangerously behind adoption. EY’s Responsible AI Pulse Survey found that 72% of executives claim their organizations have integrated and scaled AI, yet only one-third have proper governance protocols in place. This gap exposes finance functions to compliance, bias, and reputational risks as AI systems increasingly influence critical decisions.

For CFOs entering 2026, the path forward is clear. Start with your leadership team, evaluate AI literacy levels across finance, and identify skill gaps before scaling adoption. Build a structured upskilling program that combines hands-on AI exposure with domain expertise. Simultaneously, establish an AI governance council with finance, compliance, and technology leaders to define guardrails for responsible AI use, risk assessment, and auditability. This duality, when working in parallel, will differentiate AI laggers and leaders in 2026.

The future of business leadership belongs to CFOs who separate real AI progress from inflated expectations. 2025 taught us: technology capability alone doesn’t drive value. Start 2026 by assessing your AI and data readiness, building trustworthy data assets, investing in team capabilities, and embedding governance from day one. Bubble or no bubble, the opportunity to reimagine finance is very real — but only for those who dare to dream.

About the Author – Tariq Munir

Tariq is the Author of “Reimagine Finance” and advises businesses on unlocking the potential of AI, Data, and Digital. He is also an international keynote speaker, trainer, and monthly columnist at CFO Magazine A/NZ.

He can be reached at [email protected] or www.tariqmunir.me