Thank you @Louis_Frolio — I really appreciate you taking the time to read it so closely. That “tension as a feature” framing resonates a lot. Databricks doesn’t force teams to choose between speed and discipline; it reveals when it’s time to evolve from one mode to the other.

Looking ahead to 2026, if I had to prioritize one capability teams need to mature first, I’d say:

A strong Unity Catalog operating model.

Not because governance is the most exciting layer — but because it becomes the multiplier for everything else.

Without a clear UC model (ownership, domains, permissions, lineage as a first-class signal), data quality and observability struggle to scale beyond pockets of excellence. And GenAI governance or agent lifecycle management becomes reactive instead of intentional.

Once UC is treated as an operating system rather than a checkbox:

Observability shifts from “monitoring pipelines” to trusting data products

Data quality moves closer to the source, not bolted on later

GenAI guardrails and agent governance become design-time decisions, not afterthoughts

That’s when Builder speed and Strategist discipline stop competing — and start reinforcing each other.

Curious to hear your take as well — especially on how you’re seeing teams sequence this in practice.

Wishing you a joyful holiday season and a wonderful Christmas.