Snowflake CEO Sridhar Ramaswamy on $1.49B quarter (+37% YoY), AI product breakouts, and M&A philosophy

Snowflake posted $1.49 billion in product revenue for the quarter, up 37% year-over-year, with CEO Sridhar Ramaswamy pointing to broad adoption of its two AI products, Cortex Analyst (Cocoa) and Cowork, as the primary drivers.

AI products

Cortex Analyst, Snowflake's AI builder product aimed at data engineers, has been a breakout despite running with a team that never exceeded five people for most of its existence. Ramaswamy describes that constraint as a feature, not a bug: small teams with room to experiment are how Snowflake finds "magic in a bottle." Cowork, the enterprise AI assistant product, is now being deployed to thousands of users within individual companies. Ramaswamy frames both products as operating at a meaningfully higher level than simple model routing or infrastructure plays.

“We had a pretty amazing quarter — $1,490,000,000, 37% year on year. Our AI products, Cocoa and Cowork, getting really broad adoption... We are thrilled that we bought Natoma because MCP is increasingly really important for Snowflake — it provides the real time context of everything that's happening in your life, whether it's Slack or email, right in the harness.”

Organizational model

The more consequential internal shift Ramaswamy describes is a redefinition of what scale means. Headcount is no longer the primary lever. Most functions, including engineering, are expected to extract significant leverage from agentic AI rather than grow through hiring. The exceptions are customer-facing roles: account executives tied to expanding customer spend are still expected to grow. Ramaswamy is explicit that telling people "organization size is not a barometer of growth" is a hard cultural lift, and he describes it as the central mindset challenge across the company.

On internal meetings, he is skeptical but candid that communication remains unsolved. Snowflake is building an internal "enterprise brain" to route the right information to the right person, and Ramaswamy's stated preference is that anyone with nothing to contribute to a meeting should read the notes instead of attending.

AI spend and model routing

Snowflake is actively experimenting with open-source models as a cost lever, using its own Cortex gateway to instrument token usage across teams. Ramaswamy singles out support and SRE teams processing tens of thousands of alerts daily as high-leverage targets for optimization. On the broader model routing market, he is measured: it is useful infrastructure, but not where Snowflake creates its highest value. The bigger prize, in his framing, is getting every data engineer on Cortex Analyst and large fractions of enterprise employees on Cowork.

M&A

The acquisition lens is straightforward: does a company accelerate Snowflake's position as the AI and data platform for the enterprise? The most recent deal, Natoma, fits that test. Ramaswamy describes MCP (Model Context Protocol) as increasingly critical for Snowflake and its customers, because it pipes real-time context from tools like Slack and email directly into the AI harness. Natoma was acquired specifically to deepen that capability.

On competition, Ramaswamy acknowledges the pressure is real and coming from two directions simultaneously: hyperscalers and frontier AI labs. The 37% growth rate is the clearest signal that Snowflake is holding ground, but he is not treating this as a stable moment.