"APIs led us to data silos. MCP could lead us to context silos, and that's why I'm hoping people will be a bit more restrictive about where they hook up MCP." — Luke
Luke Ambrosetti has spent a decade arguing that the data warehouse should sit at the center of every enterprise martech stack. Now, as Databricks plants its CDP flag and Snowflake expands into AI decisioning, the architecture he championed is either winning or getting absorbed into something bigger. In this episode, Luke, principal industry architect for marketing and advertising at Snowflake, joins Jacqueline to unpack whether warehouse-native has become a genuine foundation or just another vendor checkbox.
They cover the full arc: why schema ownership beats security as the real argument for warehouse-native, how SaaS vendors have quietly marked up compute and resold it back to brands, and what the explosive growth of MCP servers without governance infrastructure actually means for enterprise martech teams.
This is a conversation about structural power in the stack, who controls the serving layer AI operates from, and whether innovative challengers or deep-pocketed incumbents walk away with the next decade.
Timestamps
03:39 — Resist Buying Anything First: When building a martech stack from scratch, Luke's instinct is to resist purchasing any tool for as long as possible and start with a business conversation instead, a direct challenge to the tool-first culture of the industry.
06:30 — The Warehouse as a Serving Layer for AI: Every major platform is racing to become the governed, AI-ready data layer because whoever owns that foundation wins the next generation of martech. Luke explains why that race matters more than the CDP wars ever did.
09:29 — Warehouse-Native as a Checkbox Feature: Luke names the single most common architectural mistake he sees: brands going warehouse-native in name only, with vendors attaching the label without committing to the underlying architecture.
11:55 — Vendors Are Gouging You on Compute: Martech vendors have historically marked up cloud compute and storage and resold it to brands as SaaS fees, and warehouse-native flips that dynamic by forcing vendors to compete on actual differentiated functionality.
14:44 — Schema Ownership as the Underrated Benefit: Luke argues the most overlooked advantage of warehouse-native is not security or cost savings but owning your own definition of what a customer looks like, so you are never force-fit into a vendor's data model.
17:40 — Brands Failing Publicly With AI: Luke argues that brands that experiment with AI and fail publicly deserve more credit than those that stay too conservative; visible failure is evidence of a culture willing to learn.
26:40 — The Self-Referential Nature of Big AI Deals: Luke calls out the circular dynamic of major platforms paying each other billions in AI arrangements, while still affirming that real brand-level results keep him from writing it off entirely.
28:30 — Who Wins, Innovators or Incumbents: Luke gives his personal read on where the real competitive advantage sits as every tool races toward becoming every other tool.
Sponsor
Brought to you by Hightouch, the leading composable CDP and decisioning platform trusted by brands like Domino's, Chime, and Aritzia. 90% of customers have a real use case live within their first week, delivering world-class personalization at scale. Learn more at hightouch.com/msom.
Connect & Subscribe
Subscribe to Making Sense of Martech wherever you get your podcasts. Follow us on TikTok, LinkedIn, and don't forget to like and subscribe on YouTube.
