Snowflake’s Chief Executive Officer Sridhar Ramaswamy emphasized the primacy of data over artificial intelligence models during a keynote at The Six Five Summit: AI Unleashed 2026. Speaking on Tuesday, Ramaswamy stated that trusted data, governance and contextual understanding form the foundation of sustainable AI advantage in enterprise settings.
The executive highlighted the company’s internal AI deployment, codenamed Snowhouse, which integrates data from Salesforce, Workday and Snowflake systems to power AI agents across sales, finance and engineering. Ramaswamy noted that Snowflake’s approach has reduced costs below the level of traditional dashboarding software while maintaining operational efficiency. The company’s Cortex AI offering has achieved 30% penetration across its customer base, with 46% of customers planning to increase spending on the platform in the near term.
Snowflake’s stock traded at $317.86 on Tuesday, reflecting a 66% return over the past 12 months and a market capitalization of $110.25 billion. The company’s sales force comprises 4,000 employees, who utilize internal AI tools to enhance productivity. Ramaswamy described the firm’s cost discipline as "very cutthroat," rejecting tools that deliver marginal improvements at disproportionate expense.
Ramaswamy cautioned that poor data quality cannot be remedied by advanced models, asserting that even the most sophisticated AI systems fail when presented with incorrect or misinterpreted datasets. He also stressed the importance of semantic clarity and governance controls to distinguish between operational metrics such as metered consumption and financial reporting standards like GAAP revenue.
Addressing the rapid pace of AI adoption, Ramaswamy invoked the concept of "AI dog years" to describe the accelerated timelines now possible for modernization projects, which previously required three to four years but can now be completed in weeks or months using tools like Cortex. The company employs a multi-model strategy, distributing traffic across OpenAI, Anthropic and open-source alternatives to manage costs and avoid vendor lock-in. A $30 monthly per-user budget threshold was cited as a practical limit for internal AI tool deployment.
