Most people are focused on AI. The smarter play right now is watching what's happening underneath it.

Over the last 18 months, the data infrastructure market has been quietly consolidating. The hyperscalers are building out native analytics capabilities — not to beat Snowflake or Databricks, but to keep workloads on-platform. The independent vendors are being forced to differentiate on depth, not breadth.

Three dynamics worth tracking:

1. The warehouse wars are over — it's now about the ecosystem

Snowflake and Databricks have effectively won the platform battle. The next competition is over which ecosystem developers build on top of. Snowflake's app marketplace and Databricks' Mosaic AI are both bets on the same thesis: whoever controls the compute layer controls the monetization.

2. M&A is picking up in the semantic layer

Tools like dbt, Cube, and AtScale are being evaluated for acquisition by larger BI and cloud vendors. The semantic layer is the missing piece between raw data and business decision-making — any major platform that doesn't own it is at risk of disintermediation.

3. The next funding wave is going to governance and observability

With AI data pipelines multiplying, data quality, lineage, and compliance are becoming boardroom issues. Monte Carlo, Soda, and similar players are going to see serious institutional interest over the next 12–18 months.

That's the macro thesis. Breaking it down deal-by-deal at The Data Stack.