The concept that data is too large and costly to move is an illusion created by legacy pipelines that repeatedly copy entire datasets. Fivetran's CEO asserts that modern change-data-capture techniques make data movement small-scale and inexpensive.

Many current agentic AI products are built by connecting AI to technologies, like databases, that were never designed for it. Mykhailo Marynenko calls this 'gluing shit and sticks together' and argues it's a fundamentally flawed approach. Truly innovative AI products require rebuilding the underlying infrastructure from first principles.

Contrary to conventional wisdom, MongoDB's CEO reveals enterprise leaders have a surprising appetite for full system replacement. An AI-native company that can replace an entire legacy system of record—making it cheaper, faster, and better—will get a leader's attention far more effectively than one offering an incremental feature layer on top of an existing platform.

To build a multi-billion dollar database company, you need two things: a new, widespread workload (like AI needing data) and a fundamentally new storage architecture that incumbents can't easily adopt. This framework helps identify truly disruptive infrastructure opportunities.

AI coding's true enterprise value is limited because models struggle with legacy systems. Companies run on trillions of lines of mediocre code in old languages like COBOL—a problem that requires human intervention over decades, not a simple AI solution, which limits immediate, real-world impact.

Top AI labs like OpenAI and Anthropic build internal data platforms with conventional tools like Fivetran and Snowflake. This indicates a modern data stack is perfectly sufficient for providing AI context, and companies don't need to build bespoke, exotic infrastructure.

When developers use AI to code, the AI agent itself selects the underlying infrastructure like databases. This shifts the purchasing decision from human developers and central IT teams to the AI, fundamentally disrupting how the multi-trillion dollar enterprise infrastructure market operates.

Truly massive database companies only emerge every ~15 years when three conditions are met: a new ubiquitous workload (like AI), a new underlying storage architecture that predecessors can't adopt (like NVMe SSDs and S3), and a long-term roadmap to handle all possible data queries.

Legacy business software like Excel are "IDEs for analysts" and are doomed. The core abstraction layer is shifting from graphical interfaces with complex, hard-to-discover functions to direct, natural language interaction with agents like Claude Code, which is a fundamentally superior workflow.