Why data infrastructure delivers value regardless of market direction

The AI narrative appears to be shifting.

Bloomberg has reported that businesses are struggling to demonstrate return on AI investments, whilst Sam Altman, CEO of OpenAI, has cautioned investors: “When bubbles happen, smart people get overexcited about a kernel of truth.”

To add fuel to the fire, Yann LeCun, Meta’s Chief AI Scientist, has announced he plans to leave Meta, having publicly stated that “LLMs are basically an off-ramp, a distraction, a dead end”. In Gartner Hype Cycle terms, we may well be approaching the trough of disillusionment.

So how do you position your business to deliver growth, leveraging AI, regardless of how these markets evolve?

If you are to believe the rhetoric, all your competitors are jumping on the AI bandwagon. At the same time, concerning signals are emerging regarding economic sustainability and (in some cases) a lack of genuine business impact. If the company you have staked your strategy on goes away, what happens to your business? The answer is reassuringly straightforward and (unsurprisingly!) mirrors any investment strategy in uncertain markets… Focus on the fundamentals that create lasting value.

—

The infrastructure gap

According to Gartner, 85% of AI projects fail to move beyond a proof-of-concept and into production. The primary cause of this isn’t the sophistication of the AI, it is the absence of data maturity in the business.

Organisations are still facing the same challenges:

• Data scattered across digital estates, sequestered in spreadsheets and private folders
• Inconsistent, unlabeled data that cannot be reliably cited
• Legacy systems that cannot support modern integration patterns
• Lack of governance making it impossible to expose data to new technologies safely
• No technical expertise to own or evaluate what actually works

Our message at Waracle has remained consistent throughout the AI hype cycle. Invest in data infrastructure and everything else will start to fall into place. This is the foundation necessary for modern businesses to adapt, experiment with and validate any new technology that hits the market.

Move towards action

Organisations with solid data foundations can build automated reporting, train AI models, or implement LLM applications in weeks rather than months. Those without, spend their time debating which numbers are correct, working off anecdotal evidence, and getting bogged down in manual processes.

At the core, you need a cloud-based data warehouse or lakehouse that consolidates data from across your estate into a reliable single source of truth. This platform is fed by robust integration pipelines that move and transform data from source to system, with version control and automated testing to guarantee reliability.

On top of this, sits solid data governance. Data cataloging so teams understand what they’re looking at, data quality monitoring that catches issues before they cascade, and clear data ownership with access controls. Add orchestration for workflow management, observability so you know immediately when things break, and documentation that prevents knowledge from being trapped in individuals’ heads. And you have an industry-leading data capability.

Developing these new data muscles takes time. But can happen gradually, starting with small slices of your infrastructure. Done properly, it will fundamentally change how you operate: you can trust your data, teams can find what they need, you understand lineage and dependencies, and you can move quickly when requirements change. You transition from qualitative anecdotes to quantitative insights.

From AI hype to AI value

Businesses with these foundations unlock the ability to rapidly deploy high ROI AI solutions with decades of proven market value:

• Customer purchase prediction and churn prevention that directly impact revenue
• Predictive maintenance preventing costly downtime across operations
• Fraud detection protecting millions in revenue
• Computer vision improving diagnostic accuracy and quality control

These traditional machine learning (ML) applications rest entirely on solid data foundations. They cost a fraction of LLM implementations to develop. And they’re interpretable, auditable, and trainable on your specific use case, timeframe, and scale. Most importantly, they deliver measurable value regardless of what happens in the generative AI market.

Whether LLMs become the dominant paradigm or not, the infrastructure that enables reliable predictive analytics remains your competitive advantage.

The strategic reality

The failure rates for AI projects are sobering. Gartner predicts that through 2026, 60% of AI projects will be abandoned due to insufficient data infrastructure, with 63% of organisations either unsure of or lacking in appropriate data management capabilities for AI.

The economics are equally compelling. Over 75% of organisations state that AI-ready data remains one of their top five investment areas in the next two to three years. The question isn’t whether to invest in data infrastructure, it’s whether to do it proactively as a strategic choice, or reactively after expensive AI failures.

Are LLMs disappearing? 100% not, although the market trajectory is uncertain. There are legitimate concerns about the economics of the current boom, yet we’ve also never seen this level of capital commitment to a technology. What we can state with confidence: the value of solid data infrastructure transcends any single technology cycle.

Businesses that have built proper data foundations are positioned to exploit whatever emerges victorious from the AI landscape, be that the next breakthrough in language models, multimodal systems, or technologies not yet conceived. Those who chase flashy demos and headlines over investing in the data fundamentals will remain stuck playing with basic concepts while their infrastructure-ready competitors scale meaningful solutions. This work isn’t glamorous, but getting the fundamentals right is what separates businesses that thrive on technology cycles from those that get torn apart by them.

Contact Waracle to schedule your Data Maturity Assessment. We’ll provide a clear-eyed evaluation of where you stand and what it will take to deliver the foundations that make AI, and everything after AI, actually work.