Reflections from Big Data LDN 2025: What Will Shape the Next Wave of AI & Analytics? 

Posted by | Anurag Kapoor

Big Data LDN 2025 was buzzing with energy. Over two days, thousands of practitioners, technologists, and business leaders came together to explore what’s next in Data, Analytics, AI, and the rapidly emerging world of Agentic AI. 

The event wasn’t just about showcasing the latest technologies. What stood out for me was the candor in the discussions: real-world challenges, risks, and lessons on what it actually takes to unlock value from AI. The blend of optimism and pragmatism made this year’s event particularly inspiring. 

Key Insights from Big Data LDN 

1. Agentic AI runs on strong foundations 
Voice-driven and agent-based AgenticAI may feel like science fiction, but they only succeed when the underlying DataPlatforms are fast, reliable, and well-architected. Without strong plumbing, the “intelligence” collapses. 

2. Context is the hardest nut to crack 
Enterprise AI rarely fails because of algorithms. It fails because solutions lack context. The most successful initiatives are those tightly aligned with the business environment, commercial strategy, and competitive realities — not generic “data strategies.” 

3. Commercial alignment is critical 
You don’t need a standalone “data strategy.” Data and AI strategies must flow directly from the business and commercial strategy. The closer the alignment, the greater the value delivered. 

4. Data readiness remains the #1 barrier 
It’s estimated that 95% of AI initiatives fail within 12 months. The culprit? Data that isn’t production-ready. Poor DataQuality, weak observability, and fragile pipelines continue to be the Achilles’ heel of enterprise AI. 

5. Governance must be platform-agnostic 
Effective DataGovernance cannot be tied to any one platform or tool. Governance must be independent, trusted, and embedded across the organisation, ensuring that oversight and accountability work regardless of technology stack. 

6. Culture beats tools 
Fancy dashboards don’t guarantee adoption. A shared DataCulture, a common language, and deep collaboration with business teams are far more important for lasting success. Tools help — but culture wins. 

7. Sustainability in AI is moving centre stage 
One GenAI query consumes the same energy as a lightbulb running for ten seconds. Multiply that by billions, and the environmental impact is enormous. Responsible organizations are beginning to track carbon costs alongside accuracy (for example, using tools like CodeCarbon) and are redefining what “success” means in AI. 

8. Scaling AI is still a major challenge 
Even when prototypes succeed, many client-facing AI solutions stumble at scale. Common issues include: 

• Pulling the wrong data 

• AI-generated code introducing hidden bugs 

• Hallucinations slipping through without human oversight 

• Vector databases lagging on updates 

• Small configuration changes causing large swings in outcomes 

This challenge goes beyond hallucinations — it’s about shifting inputs, evolving pipelines, and the non-deterministic nature of modern AI. 

9. Humans remain central in the AI world 
AI can process data 8X faster and make insights accessible to non-technical users. But ResponsibleAI demands human involvement — to provide oversight, apply context, and ensure systems remain grounded in reality. 

10. A Cosmic Perspective
The highlight of the event for me was the closing keynote by Brian Cox, who drew parallels between AI, big data, and the cosmos. His reflections were a timely reminder: 

• Will AI take over human intelligence? No. 

• Will it discover new theories like Einstein’s relativity? No — not for many years to come. 

AI is a powerful tool — but it is here to augment human intelligence, not replace it. The real opportunity is to use it wisely: to make businesses sharper, decisions faster, and lives easier.

Final Reflection 
The overarching lesson from Big Data LDN 2025 was clear: to win with AI, data must become a competitive advantage. 

Technology alone will not deliver results. Success requires: 

• Strong and resilient data foundations 

• Commercial alignment, not siloed strategies 

• Platform-independent governance 

• A culture where business and data teams pull together 

At Fuld, we are helping clients address these very challenges — turning data into a driver of competitive advantage, and AI into a source of tangible value.